Here’s the thing: I believe talent exists. You aren’t born identical to the kid beside you, you won’t adapt identically, and you probably won’t share the same ceiling. What I don’t believe is that I can look at you at twelve, measure today’s performance, and reliably discover how far you’ll eventually go.

If you’re coaching tomorrow, you’ll still have to choose a team, offer one athlete extra support, or tell another athlete, “not here, not now.” You can’t avoid every decision, and I won’t pretend that you can. But you shouldn’t pretend that your decision has revealed somebody’s destiny.

Talent exists. The clean prediction doesn’t

What exactly do I mean by talent?

You might mean current performance: who’s running fastest, scoring most, or dominating today. Or you might mean rate of improvement: who’s learning or adapting fastest after exposure to training. And you might mean potential ceiling: how far you’d reach under some favourable but unknowable combination of coaching, health, motivation, opportunity, and time (Baker, Wattie, and Schorer 2019; Pickering and Kiely 2017).

Those aren’t the same thing. Current performance is visible, but it’s already mixing biology with maturation, training history, family support, confidence, access, and luck. You can’t estimate improvement rate from one test, and you won’t know the ceiling until somebody gets close—if they ever receive the conditions needed to approach it.

I don’t deny innate abilities, constraints, or ceilings. But I can’t inspect you at thirteen and separate temperament from upbringing, trainability from training history, or persistence from years of encouragement (Côté 1999; Tucker and Collins 2012). If you want to call hard work “character,” fair enough, but I still can’t tell you where that character came from.

Honestly, most of the time, I don’t know. And neither do you.

That’s not an argument against talent. It’s an argument against pretending that an uncertain construct has become visible because you’ve given a child a score.

The bullshit isn’t talent. The bullshit is converting an early observation into certainty about a distant future.

Because the decision won’t wait for certainty, I’d use a barbell. It lets me gamble without pretending that I can see the cards.

When you don’t know, use the barbell

I’ve borrowed the Barbell Strategy from Nassim Nicholas Taleb’s writing on uncertainty: The Black Swan is the original source I’m using here, and Antifragile develops the logic further (Taleb 2007, 2012). I’m making the sporting translation. Picture an actual bar with most of its weight at two distant ends and almost nothing in the middle (and no, Taleb didn’t mean a gym programme).

In Taleb’s financial example, you’d keep most resources protected from ruin while putting a much smaller portion into speculative bets with limited losses and potentially large gains. You’d avoid the apparently sensible middle, where “moderate” risk can hide serious exposure without much upside. It isn’t a literal 50/50 split. It’s a way to gamble without allowing one wrong bet to destroy you.

Children aren’t financial assets, so I don’t want sport copying a portfolio literally. At first, the comparison feels cold, but it makes sense once you ask what each error costs.

But the decision logic transfers. At one end, I’d protect broad participation, reduce dropout, and avoid ruin. At the other, I’d make smaller, reversible bets on athletes who may benefit from extra coaching, a stronger competition, or a temporary place in a performance hub. The fragile middle is prediction theatre: narrow pathways, expensive certainty, and labels that can’t easily be undone.

Look, it’s like building a bridge while refusing to anchor either bank (that analogy just occurred to me, but it fits). If the participation base isn’t protected and the performance bets aren’t reversible, the polished middle won’t save the system.

On the protective end, I’d use Via Negativa, Latin for “the negative way.” You don’t begin by adding another test, academy, or dashboard; you improve the system by subtracting known obstacles and harms. You’re removing fees, transport failures, locked courts, one annual trial, and permanent deselection—not practice, responsibility, frustration, or demanding competition.

The barbell doesn’t reveal who has talent. It tells me how to act while I don’t know. Protect the base, make bounded bets, and keep enough doors open to correct the mistakes.

I’ve mapped that sporting version in Figure 1. You’ll see that the two ends do different jobs: one protects the floor, while the other exposes athletes to informative upside. And you’ll see why I don’t trust the crowded middle.

Help talent by not destroying it

I’m using Via Negativa in a very practical way: before you ask how to manufacture talent, you stop destroying possible talent through fees, transport, humiliation, premature specialisation, one annual trial, and permanent deselection. You won’t know which removed barrier mattered most, and that’s fine. Basically, you’re trying to keep more people inside enough good environments for long enough. Not a prophecy.

But easier access isn’t the same as easier development. I don’t need to remove a hard practice when I remove a bus fare, and you don’t need to promise playing time when you open a court. You can protect a child from abuse without protecting that child from losing, waiting, practising, failing, being corrected, or meeting somebody better.

At Aspire Academy, I’d see boys with school, food, transport, medical care, therapists, and excellent facilities in one place. I don’t know whether that abundance blunted anyone, and I won’t invent a causal fairy tale. But I kept asking where autonomy, improvisation, frustration tolerance, and responsibility would enter if every rough edge disappeared.

In my 2016 article, I also asked whether “talent needs trauma” (Jovanović 2016). Actually, let me rephrase that: trauma isn’t a training prescription. The relevant studies suggest that how athletes engage with challenge—through coping, reflection, responsibility, and learning—may distinguish developmental paths (Collins and MacNamara 2012; Collins, MacNamara, and McCarthy 2016).

Remove pointless friction. Preserve productive struggle.

That’s the distinction. I want to raise the floor without lowering the ceiling, but I won’t wrap the athlete in bubble wrap.

Selection changes the thing you’re measuring

Suppose you’re watching two thirteen-year-old rugby players. One’s matured early, trained for four years in a strong club, and always has transport; the other’s still maturing, recently changed sports, and sometimes misses the bus. The first is stronger and more influential today, but that isn’t a clean reading of the future.

Relative age isn’t mysterious; it’s simply how old a child is compared with others in the same annual age group. Today’s performance doesn’t come from one hidden substance, because maturation, access, training history, confidence, position, and playing time are all shaping it (Cobley et al. 2009; Vaeyens et al. 2008; Johnston et al. 2018). Once you’ve selected the first player, you’ll also change access to coaching, stronger peers, facilities, competition, and encouragement.

That’s the feedback loop I’ve shown in Figure 2. You’ll see that selection doesn’t merely record development; it intervenes in development, and later improvement can then appear to validate the original choice. Without a re-entry arrow, the system can’t easily discover its own false negatives.

A measurement can’t choose the purpose

Selection isn’t only a measurement problem. It’s a value problem. “Player A is faster today” doesn’t tell me whether I should select Player A for Saturday, invest in Player B for five years, keep both in sport, or redesign the pathway. The choice depends on what the system is trying to achieve.

Does that make selection useless? No, it doesn’t. A coach can’t give everybody equal playing time, and a federation can’t fund everybody at the highest level. But a bounded observation should support a bounded claim.

“Better today” is defensible, while “more talented” is shakier and “future international” is mostly storytelling. That’s why you shouldn’t let one useful observation carry three different meanings.

The evidence doesn’t make that prediction problem disappear. A meta-analysis of 13,392 athletes across Olympic sports found only a weak junior-to-senior association, (r = 0.148), and it was weaker in younger categories (Barth et al. 2024). Other reviews don’t support a simple “earlier is better” story either: earlier programme entry aligned more with junior success, while better senior athletes tended to enter later (Barth et al. 2022; Güllich and Barth 2024).

Genetics doesn’t rescue the verdict. In Craig Pickering’s case study, genotype scores couldn’t separate five elite track-and-field athletes from 503 controls, and 68 controls scored higher for speed-power than the elite power athletes (Pickering 2020; Pickering and Kiely 2021). Biology matters, but I can’t turn that truth into a reliable early selection algorithm (Webborn et al. 2015).

So I don’t stop selecting, funding, or exposing some athletes to stronger coaching and harder competition. I can’t avoid those choices. But I’ll treat them as bounded bets, not discoveries of destiny.

If a positive bet fails, I’ve spent some time and resources. If a negative judgment becomes permanent, I might lose an athlete and never learn that I was wrong. That’s the error I don’t want your system to hide.

Talent can move sideways

But the thing is, development doesn’t move only upward through one sport’s pyramid. Sometimes you’ll need to help an athlete move sideways, and that’s where donor sports and talent transfer become useful.

In the Athletic Skills Model, a donor sport isn’t a lesser sport feeding a superior one; it’s a sport or activity that develops action capabilities that may be useful in a target sport (Wormhoudt et al. 2018). You’d look for overlap in movement, perception, timing, balance, spatial orientation, tactical problems, emotional regulation, or learning demands.

You’re seeing talent transfer when a trained athlete moves from a donor sport into a recipient sport. A recent review doesn’t treat that move as automatic, because transfer still depends on the athlete, task, environment, support, and an actual opportunity to progress (Green et al. 2024). Similarity isn’t a guarantee, but it’s a sensible clue.

So I’d build cross-sport trial days, short bridge blocks, and referral agreements between clubs. An athlete who isn’t retained shouldn’t have to start from zero or disappear. The receiving coach may get someone who already knows how to practise, compete, recover, and take feedback.

I’ve drawn that wider ecosystem in Figure 3. You’ll see entries for children, late starters, transferring athletes, and adult beginners, plus routes back when the first judgment doesn’t hold. And it doesn’t pretend that every door must lead to elite sport.

Re-entry isn’t an afterthought

The routes back in Figure 3 matter as much as the routes forward. If you haven’t entered football early, joining a serious team at fourteen can be almost impossible. Sessions assume years of shared knowledge, and nobody’s responsible for bringing a late starter up to speed. That doesn’t prove you can’t learn; it proves the beginner door has disappeared.

Adults aren’t treated much better. If you’re forty and want to learn basketball, you’ll find clubs or informal groups whose members already play, but few actual schools for adult beginners. An adult beginner school probably won’t produce an Olympian, but that’s not its only return.

Those adults can become active parents, role models, volunteers, teammates, and coaches; they’ll help children see sport as something people do rather than something selected children own. Parental support and modelling are associated with children’s physical activity, although the relationship isn’t deterministic (Yao and Rhodes 2015). That’s why I’d build beginner groups, seasonal entry windows, community leagues, cross-sport trials, and return-to-sport blocks. One route can stay demanding without every other door being locked.

Talent is an oblique outcome

The ecosystem in Figure 3 is more than a kinder selection diagram. It’s a different strategy for developing sport. John Kay calls the underlying idea obliquity: in complex systems that we don’t fully understand, important goals are often achieved more reliably through indirect routes than through a frontal attack (Kay 2010).

Obliquity isn’t passive. I can directly improve the conditions: I can help you land, throw, swim, cooperate, and practise; I can improve your coaching, facilities, competition, and access. What I can’t do is manufacture an elite athlete to order. It’s sitting far downstream of biology, family, culture, opportunity, learning, injury, persistence, luck, and time.

So if I were designing a national sport system, I wouldn’t make “produce more champions” its main operating target. That’s too distant and too easy to game. I’d choose nearer goals that you can actually influence: help more people move, let more children sample activities, reduce dropout, build movement competence, improve health and longevity, and keep people involved long enough to discover what they might become.

This is closer to public health than a medal factory, and I’m comfortable with that. About four in five adolescents worldwide don’t meet recommended physical-activity levels (Guthold et al. 2020). Improving that number isn’t only a sporting project; it’s already a worthwhile national project, even if the medal table doesn’t move.

I don’t need to romanticise childhood twenty or thirty years ago, when access and safety were hardly equal, to notice what many children have lost: walking, cycling, climbing trees, roaming parks, inventing games, and accumulating movement without calling it training. That activity won’t simply reappear because adults miss it.

We’ll sometimes have to organise it, but “organise” should mean arranging access, time, space, and safety—not scripting every movement. I’d open school grounds, build safe parks and routes, make clubs affordable, and provide loose equipment with playful challenges. And I’d ask adults to notice when they’re needed and when they should get out of the way (Farley et al. 2007; Lafleur et al. 2013; Brussoni et al. 2015). If it looks like another compulsory lesson with queues, cones, and lectures, don’t be surprised when children avoid it.

And I’d put qualified physical-education specialists into kindergarten and the first four grades, leading or co-leading PE rather than leaving it to whichever classroom teacher has the period. That isn’t an insult to classroom teachers; movement education is its own expertise, and school is the one institution that reaches children whose families can’t buy sport. UNESCO’s Quality Physical Education guidance places qualified personnel and early, inclusive provision at the centre of policy, while the SPARK trial found that specialist-led elementary PE produced better activity-related outcomes than the usual programme and generally outperformed trained classroom teachers (McLennan and Thompson 2015; Sallis et al. 1997).

Access won’t ever be perfectly even. Maybe I could’ve become a great skier, but without snow, a slope, equipment, money, or a skiing culture, that possibility would’ve stayed mostly theoretical (there’s no spreadsheet that fixes geography). No national system can provide every sport in every village. Regional hubs, school partnerships, short camps, transport support, and cross-sport referrals can widen the menu, but they won’t remove geography or culture.

The national goal should be a healthier, more active population with more opportunities and more optionality. Elite talent can emerge, but it won’t be the only return.

Will talent simply take care of itself?

Not quite. It’ll still need coaching, challenge, facilities, competition, support, and years of hard work. But it doesn’t need me to predict it correctly at twelve. It needs a system that won’t destroy it before it becomes visible.

Where I land

I believe talent exists, and I don’t think you’re born as a blank slate. I believe innate differences matter. But I also believe current performance, improvement rate, and eventual ceiling aren’t the same thing, and we can’t separate them cleanly when you’re young.

I’m still left with a gamble. That’s where the barbell helps me: I’ll protect broad participation, remove barriers that can cause premature ruin, preserve the hardship required for development, and place smaller reversible bets where extra opportunity may reveal something useful.

I’m using obliquity to aim the larger system. You shouldn’t chase champions directly, because you can’t control that distant outcome. You’ll do more good by raising population activity and health, making movement attractive and ordinary, supporting active parents, improving early physical education, creating sensible hubs, and keeping doors open across ages and sports.

Talent’s real. We just can’t see it clearly in advance. So build the conditions, stop destroying possibilities, and let talent emerge.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

References

Baker, Joseph, Nick Wattie, and Jörg Schorer. 2019. ‘A Proposed Conceptualization of Talent in Sport: The First Step in a Long and Winding Road’. Psychology of Sport and Exercise 43: 27–33. https://doi.org/10.1016/j.psychsport.2018.12.016.

Barth, Michael, Arne Güllich, Brooke N. Macnamara, and David Z. Hambrick. 2022. ‘Predictors of Junior Versus Senior Elite Performance Are Opposite: A Systematic Review and Meta-Analysis of Participation Patterns’. Sports Medicine 52 (6): 1399–1416. https://doi.org/10.1007/s40279-021-01625-4.

———. 2024. ‘Quantifying the Extent to Which Junior Performance Predicts Senior Performance in Olympic Sports: A Systematic Review and Meta-Analysis’. Sports Medicine 54 (1): 95–104. https://doi.org/10.1007/s40279-023-01906-0.

Brussoni, Mariana, Rebecca Gibbons, Casey Gray, Takuro Ishikawa, Ellen Beate Hansen Sandseter, Adam Bienenstock, Guylaine Chabot, et al. 2015. ‘What Is the Relationship Between Risky Outdoor Play and Health in Children? A Systematic Review’. International Journal of Environmental Research and Public Health 12 (6): 6423–54. https://doi.org/10.3390/ijerph120606423.

Cobley, Stephen, Joseph Baker, Nick Wattie, and Jim McKenna. 2009. ‘Annual Age-Grouping and Athlete Development: A Meta-Analytical Review of Relative Age Effects in Sport’. Sports Medicine 39 (3): 235–56. https://doi.org/10.2165/00007256-200939030-00005.

Collins, Dave, and Áine MacNamara. 2012. ‘The Rocky Road to the Top: Why Talent Needs Trauma’. Sports Medicine 42 (11): 907–14. https://doi.org/10.2165/11635140-000000000-00000.

Collins, Dave, Áine MacNamara, and Neil McCarthy. 2016. ‘Super Champions, Champions, and Almosts: Important Differences and Commonalities on the Rocky Road’. Frontiers in Psychology 6: 2009. https://doi.org/10.3389/fpsyg.2015.02009.

Côté, Jean. 1999. ‘The Influence of the Family in the Development of Talent in Sport’. The Sport Psychologist 13 (4): 395–417. https://doi.org/10.1123/tsp.13.4.395.

Farley, Thomas A., Rebecca A. Meriwether, Erin T. Baker, Liza T. Watkins, Carolyn C. Johnson, and Larry S. Webber. 2007. ‘Safe Play Spaces to Promote Physical Activity in Inner-City Children: Results from a Pilot Study of an Environmental Intervention’. American Journal of Public Health 97 (9): 1625–31. https://doi.org/10.2105/AJPH.2006.092692.

Green, Adeline, David Fleischman, Rory Mulcahy, and Bridie Kean. 2024. ‘Game Changer—Talent Transfer Pathways in Sport: A Systematic Review’. Frontiers in Sports and Active Living 6: 1401409. https://doi.org/10.3389/fspor.2024.1401409.

Güllich, Arne, and Michael Barth. 2024. ‘Effects of Early Talent Promotion on Junior and Senior Performance: A Systematic Review and Meta-Analysis’. Sports Medicine 54 (3): 697–710. https://doi.org/10.1007/s40279-023-01957-3.

Guthold, Regina, Gretchen A. Stevens, Leanne M. Riley, and Fiona C. Bull. 2020. ‘Global Trends in Insufficient Physical Activity Among Adolescents: A Pooled Analysis of 298 Population-Based Surveys with 1.6 Million Participants’. The Lancet Child & Adolescent Health 4 (1): 23–35. https://doi.org/10.1016/S2352-4642(19)30323-2.

Johnston, Kathryn, Nick Wattie, Jörg Schorer, and Joseph Baker. 2018. ‘Talent Identification in Sport: A Systematic Review’. Sports Medicine 48 (1): 97–109. https://doi.org/10.1007/s40279-017-0803-2.

Jovanović, Mladen. 2016. ‘Is Talent Identification BS?’ Complementary Training. 15 June 2016. https://complementarytraining.com/is-talent-identification-bs/.

Kay, John. 2010. Obliquity: Why Our Goals Are Best Achieved Indirectly. London: Profile Books.

Lafleur, Mariah, Eloisa Gonzalez, Liz Schwarte, Rajni Banthia, Tony Kuo, Joanie Verderber, and Paul Simon. 2013. ‘Increasing Physical Activity in Under-Resourced Communities Through School-Based, Joint-Use Agreements, Los Angeles County, 2010–2012’. Preventing Chronic Disease 10: E89. https://doi.org/10.5888/pcd10.120270.

McLennan, Nancy, and Jannine Thompson. 2015. ‘Quality Physical Education (QPE): Guidelines for Policy-Makers’. Paris: United Nations Educational, Scientific and Cultural Organization. https://unesdoc.unesco.org/ark:/48223/pf0000231101.

Pickering, Craig. 2020. ‘Is There Utility to Genetic Information in Elite Sport?’ Doctoral thesis, University of Central Lancashire. https://knowledge.lancashire.ac.uk/34497/.

Pickering, Craig, and John Kiely. 2017. ‘Can the Ability to Adapt to Exercise Be Considered a Talent—and If so, Can We Test for It?’ Sports Medicine - Open 3 (1): 43. https://doi.org/10.1186/s40798-017-0110-4.

———. 2021. ‘Can Genetic Testing Predict Talent? A Case Study of Five Elite Athletes’. International Journal of Sports Physiology and Performance 16 (3): 429–34. https://doi.org/10.1123/ijspp.2019-0543.

Sallis, James F., Thomas L. McKenzie, John E. Alcaraz, Bohdan Kolody, Nancy Faucette, and Melbourne F. Hovell. 1997. ‘The Effects of a 2-Year Physical Education Program (SPARK) on Physical Activity and Fitness in Elementary School Students’. American Journal of Public Health 87 (8): 1328–34. https://doi.org/10.2105/AJPH.87.8.1328.

Taleb, Nassim Nicholas. 2007. The Black Swan: The Impact of the Highly Improbable. New York: Random House.

———. 2012. Antifragile: Things That Gain from Disorder. New York: Random House.

Tucker, Ross, and Malcolm Collins. 2012. ‘What Makes Champions? A Review of the Relative Contribution of Genes and Training to Sporting Success’. British Journal of Sports Medicine 46 (8): 555–61. https://doi.org/10.1136/bjsports-2011-090548.

Vaeyens, Roel, Matthieu Lenoir, A. Mark Williams, and Renaat M. Philippaerts. 2008. ‘Talent Identification and Development Programmes in Sport: Current Models and Future Directions’. Sports Medicine 38 (9): 703–14. https://doi.org/10.2165/00007256-200838090-00001.

Webborn, Nick, Alun G. Williams, Mike McNamee, Claude Bouchard, Yannis Pitsiladis, Ildus Ahmetov, Euan A. Ashley, et al. 2015. ‘Direct-to-Consumer Genetic Testing for Predicting Sports Performance and Talent Identification: Consensus Statement’. British Journal of Sports Medicine 49 (23): 1486–91. https://doi.org/10.1136/bjsports-2015-095343.

Wormhoudt, René, Geert J. P. Savelsbergh, Jan Willem Teunissen, and Keith Davids. 2018. The Athletic Skills Model: Optimizing Talent Development Through Movement Education. London: Routledge. https://doi.org/10.4324/9781315201474.

Yao, Christopher A., and Ryan E. Rhodes. 2015. ‘Parental Correlates in Child and Adolescent Physical Activity: A Meta-Analysis’. International Journal of Behavioral Nutrition and Physical Activity 12: 10. https://doi.org/10.1186/s12966-015-0163-y.

Hard sparring. Hard pads. Hard rounds. Hard bag work. Hard conditioning.

The problem is that hard can point in at least two directions.

A task can be hard because the fighter works at a high pace, produces force, accumulates fatigue, absorbs contact, takes risk, or pays a recovery bill. Another task can be hard because the fighter does not know what is coming, must read the opponent, distinguish the real attack from the feint, choose among several viable weapons, and adapt when the first action changes the situation.

One task can be physically brutal and informationally simple. Another can be physically light and technically-tactically difficult.

Those are not the same kind of hard.

Once we collapse them into one word, practice design becomes muddy. We call a drill “intense” without knowing whether we mean effort, speed, contact, uncertainty, danger, or social pressure. We call something “complex” because it contains a twelve-step combination, although every step is known before the athlete starts. We call something “specific” because it looks like fighting, while removing the opponent-generated information that gives fighting its meaning.

Then we argue about drill names.

Is heavy-bag work useful? Is pad work specific? Are chain drills bad? Is playful sparring too easy? Is hard sparring necessary?

These questions start one level too late. A drill name is a container. The actual practice task depends on its rules, information, partner behaviour, pace, consequences, athlete, and purpose.

What we need is not another ranking of good and bad drills. We need a map that separates the information the task contains from the price the fighter pays for engaging with it.

I call it the Combat Quadrant.

Classify the version, not the label

Imagine two coaches saying they did pad work.

In the first version, the fighter repeats a known combination against a cooperative pad holder. The pace is high. The work is dense. The goal is to produce, rehearse, and maintain mechanics under fatigue. The informational demand is low because the sequence is known.

In the second version, the pad holder presents meaningful cues. Several actions are viable. Timing and position change. The fighter must read, select, and continue. Physical demand is deliberately controlled so that errors remain affordable.

Same label. Different task. Different dominant intention. Different location on the map.

The same applies to sparring, chain drills, shadowboxing, wall work, Dutch-style exchanges, bag work, positional grappling, and almost everything else we name as if the noun had already specified the practice.

This is why the first rule is simple:

Classify the task as performed, not by its traditional name.

In my earlier work on Problemming, I argued that coaches should start with the recurring performance problem rather than their favourite drill, method, or capacity (Jovanović 2026a, 2026f). The Combat Quadrant continues that argument. Once the problem is provisionally framed, we still need to decide how much of it to preserve, how much to simplify, and which cost is justified.

The problem is the anchor. The dose is the regulator.

Two nominal directions of hard

The horizontal axis of Figure 1 is competition-specific task complexity.

Moving right adds viable weapons, uncertainty, opponent freedom, representative timing and distance, deception, adaptive continuation, and the need to use information to organise action. This is not generic complication. A twenty-movement chain can be difficult to remember while remaining completely predictable. Two live options can be more competition-specific if the fighter must detect which option is unfolding.

The vertical axis is physical intensity and consequence.

Moving upward can add pace, force, effort, work density, fatigue, contact, mechanical exposure, psychological threat, error cost, risk, and recovery demand. These components often travel together, but they can separate. A slow selection round may carry enormous psychological consequence. A maximal bag round may be exhausting without containing an opponent decision. A playful task may contain rich information while remaining physically cheap.

So the vertical axis is not a hidden physiological scalar. It is an expected-cost profile. When the distinction matters, unpack it.

The axes are deliberately simplified. They are separable for thinking, not independent in reality. Increasing opponent freedom may increase fatigue and threat. Increasing contact can alter what information the athlete attends to. A useful map compresses. A dangerous map forgets that it compressed.

Four regions, four dominant intentions

The two axes create four broad practice regions. They are dominant intentions, not exclusive effects. Learning can occur everywhere. Capacity may change during sparring. Confidence may improve during maintenance. A transfer test may reveal the next learning problem.

The useful question is not, “What could this drill possibly develop?” If we stare long enough, every task produces a Christmas tree of possible adaptations.

Ask instead:

What is the main job of this version today?

The quadrant names are practical shorthand for common dominant uses, not intentions mechanically generated by coordinates. A high-information, low-cost task may maintain an expert’s skill; a constrained task may teach a novice something about their own coordination. The map proposes a likely job. The coach still has to state the actual one.

Q1 — Capacity Building

Capacity Building occupies the upper-left: relatively constrained information with relatively high physical demand.

Here we simplify the problem so the fighter can produce, repeat, tolerate, or organise substantial work. The target may be output, mechanics under load, force transmission, tissue tolerance, robustness, repeatability, or work capacity. Heavy-bag work is a useful extreme point. Pre-arranged pressure and predictable partner or pad exchanges occupy a wider region.

This gives Substance a legitimate home. Capacity work builds potential; it does not prove that the potential will be organised against an adaptive opponent (Jovanović 2026e). That is not a criticism. A screwdriver is not a bad hammer. It has another job.

The question is: Which resource are we loading, and where will we later check whether it becomes available in representative action?

Q2 — Practice to Transfer

Practice to Transfer occupies the upper-right: high competition-specific complexity combined with meaningful physical or consequential pressure.

This is where we ask what survives. Can the fighter still perceive, decide, coordinate, and act when pace, fatigue, threat, opposition, and consequence converge?

The match sits inside a wider match-demand envelope. A match is not one stable dose, nor is it maximal on every component. A practice task may exceed match pace, repetition, force, or option density on one selected dimension. That is component overload, not proof that the whole task has become “more specific than competition.”

Transfer practice is expensive. It can narrow exploration, reduce the number of quality attempts, create recovery cost, and punish unstable solutions. That does not make it bad. It makes it scarce.

The question is: What must be present for this to count as a meaningful transfer test, and what is the minimum justified cost of obtaining that information?

Q3 — Practice to Maintain

Practice to Maintain occupies the lower-left: familiar, predictable work with relatively low physical and consequential cost.

Its job is to rehearse, refresh, stabilise, prepare, or restore confidence. Warm-up is an extreme point, but it is not the whole quadrant. Familiar chain drills, controlled technical work, low-cost shadowboxing, and simple partner exchanges may live here.

Maintenance is not failed learning. Not every session needs novelty, destabilisation, and a nervous system press conference. Sometimes a solution must remain available. Sometimes the fighter needs successful repetitions without an additional recovery bill. Sometimes this practice prepares the athlete for what follows.

The question is: What should remain available, and how can we preserve it without paying an unnecessary price?

Q4 — Practice to Learn

Practice to Learn occupies the lower-right: high competition-specific complexity with bounded physical and consequential cost.

The fighter still reads, chooses, times, adapts, and solves. We reduce the price of being wrong so that errors remain usable rather than merely painful.

Lower consequence can create more attempts, wider exploration, more interpretable errors, and less defensive protection of status. Playful practice is one valuable region here. Play is not the absence of structure or intent; it is a motivational frame that can permit invention without turning every attempt into selection.

But learning practice is not transfer practice. Behaviour can change when force, fatigue, threat, scoring, and stakes rise. That is precisely why the two regions remain distinct.

The question is: Have we reduced the cost without removing the information that makes this the same problem?

The map has four layers, not only two axes

The Combat Quadrant becomes clearer when we separate four questions, shown in Figure 2.

First, the two axes locate the nominal task: what the coach designed, independent of who performs it.

Second, representativeness asks about task–target fit. Does the task preserve the information–action–consequence relations that matter for the target performance problem?

Third, functional difficulty asks about task–athlete fit. How challenging is this task for this fighter, with this partner, in this state, today?

Fourth, practice intention asks why the coach is using it: learn, transfer, maintain, or build capacity?

This separation prevents several category errors. A task can be highly complex but poorly representative because the options are irrelevant to the competition problem. A task can be representative of a narrow problem without being complicated. A task can be nominally modest and functionally punishing for a novice. A task can be highly specific yet deliberately easy because the goal is maintenance.

Similar to Challenge Point, but not the same model

The closest theoretical relative is the original Challenge Point Framework (Guadagnoli and Lee 2004) and the Extended Challenge-Based Framework developed for sports coaching (Hodges and Lohse 2022).

The original framework distinguishes nominal task difficulty from functional task difficulty. The same nominal task can provide a very different challenge to a novice and an expert. Learning is expected to benefit from usable information—not from difficulty for its own sake—and too little or too much functional difficulty can reduce learning potential. The nominal–functional distinction has also been examined experimentally in transfer tasks (Sanli and Lee 2015). Even so, the framework should not be treated as a machine that calculates one universal optimal difficulty.

Hodges and Lohse add motivation, competition specificity, and three non-exclusive practice priorities: practice to learn, practice to transfer, and practice to maintain. Their central two-dimensional figure places specificity to competition on the horizontal axis and functional task difficulty on the vertical axis. Figure 3 is a simplified redrawing of that relationship, not a reproduction of their published figure.

The family resemblance is obvious. Both models care about information, challenge, specificity, motivation, learning, transfer, and maintenance. Both reject the idea that one drill has one fixed difficulty.

The difference is the level of decomposition.

The Extended framework maps specificity against the athlete-relative difficulty of the task. The Combat Quadrant first decomposes the nominal practice environment into two directions from which challenge may emerge: informational complexity and physical intensity/consequence. Functional difficulty then becomes an athlete-specific overlay. The Combat Quadrant also adds Capacity Building because combat practice includes legitimate high-output, decision-light work that does not need to pretend it is skill-transfer practice.

In compact form:

where functional difficulty depends on competition-specific complexity , physical intensity/consequence , athlete capability , current state , partner behaviour , and context .

This is not an equation to calculate. It is a reminder not to confuse the task’s coordinates with the athlete’s experience.

Rotated axes? Useful analogy, not identity

In our group discussion, I compared the relationship to principal component analysis: perhaps the Extended model and the Combat Quadrant describe similar territory with axes rotated in another direction.

I still find that analogy useful, but only as a conceptual sketch. Nobody has collected a task-rating dataset and run PCA here. We do not know that the variables are statistically correlated, much less that the axes are orthogonal components.

The safer claim is that the constructs overlap and may covary.

Competition specificity can depend on opponent information, timing, options, speed, force, fatigue, contact, psychological pressure, and the target problem. Functional difficulty can rise because options overwhelm the athlete, because pace removes time, because fatigue changes action capability, or because being hit makes the error expensive. Neither construct is identical to either Combat Quadrant axis.

Figure 4 shows the relationship as a factor-loading sketch rather than an empirical result.

This is why intensity belongs inside the story of functional difficulty without becoming its synonym. Being hit can make a task functionally harder. It can also make a stupid task more costly without making it more informative.

Representativeness is an overlay, not a compulsory Z-axis

Representative Learning Design asks whether practice samples relevant informational variables and preserves the perception–action relations required by the performance environment (Pinder et al. 2011). That identifies a real limitation in any two-axis model: a task can be complicated and brutal while preserving the wrong relationship.

This point is especially important in combat sports. Ecological-dynamics accounts of MMA emphasise live opponent information and the athlete’s continuing coupling of perception, decision, and action (Yearby et al. 2024; Myszka, Yearby, and Davids 2023). Yet this should not become another tribal slogan in which opposed practice is good and unopposed practice is bad. Unopposed tasks can have a bounded role when the coach is honest about the problem, the information removed, and the intended job (Parry et al. 2025).

Yet I would not add representativeness as a third spatial axis to the working map. A 3D model may improve descriptive precision while becoming harder to use on the mat. More importantly, representativeness is a relationship between the task and the target problem, not simply another quantity sitting inside the drill.

I therefore treat it as an overlay, contour, or entry threshold.

For Practice to Learn and Practice to Transfer, the task should be representative enough for the selected problem. That does not mean reproducing the whole match. A narrow wall-wrestling task can be highly representative of one problem while excluding most of MMA. A choreographed chain may contain many movements and high fatigue while preserving almost no opponent-coupled information.

For Capacity Building and some Maintenance work, falling below that representativeness threshold can be completely legitimate. Heavy bags, fixed chains, and shadowboxing do not need to lie about their job to earn a place in training.

Drills have paths, not addresses

The map becomes useful when we follow how a task moves.

A chain drill at low cost usually supports maintenance. Increase pace, force, density, or fatigue while keeping the sequence known, and it moves upward toward capacity. Add genuine branches and opponent-generated cues, and it moves right toward learning. Add adaptive continuation, scoring, and representative consequence, and it can move toward transfer—although at that point it is no longer a purely pre-arranged chain.

Shadowboxing can serve as warm-up, familiar rehearsal, rhythm work, or capacity development. Strong movement constraints may help a fighter explore self-coordination. But shadowboxing lacks an opponent supplying live information and adapting in return. I would not place pure shadowboxing deep inside opponent-coupled learning or transfer. It may help the athlete learn something; it does not independently demonstrate that the athlete can solve a fighting problem.

Pad work may occupy several regions. The holder can deliver choreography, meaningful cues, or a confusing hybrid in which the fighter guesses the coach’s private script. Ask whether the fighter is reading, remembering, producing, or solving.

Sparring is not one address either. Playful sparring, constrained learning sparring, technical sparring, hard performance rehearsal, selection rounds, and competition form a broad landscape. The noun tells us almost nothing about the actual information and cost.

Figure 5 shows several legitimate paths. Progress is not always up and to the right.

Moving down may restore confidence or exploration. Moving left may isolate a bottleneck. Returning to maintenance may stabilise a fragile solution. Capacity work may be correct when the solution exists but disappears under physical demand.

The athlete does not unlock Match Mode after collecting enough drill points.

Three problems, several legitimate practices

One recurring problem can generate several useful tasks. The quadrant changes because the job and constraints change—not because the coach has betrayed a methodology.

Problem 1 — Escaping the fence

Suppose a fighter repeatedly accepts a square stance, loses inside position, and remains trapped on the fence.

A low-cost sequence of framing, pummelling, turning, and resetting can support maintenance once the solution is already available. Keep the partner’s response predictable, raise the pressure, work density, and duration, and the task moves upward toward capacity: can the fighter keep posture and produce the escape against a stable but demanding load?

For learning, let the pinning partner choose between two or three representative controls while contact consequence remains bounded. The escaping fighter must read the pressure and discover when to frame, turn, change level, or reverse direction. For transfer, restore continuation, scoring, fatigue, strikes where appropriate, and the possibility of being re-pinned. Now we ask what survives.

The problem remained recognisable. The information and price changed.

Problem 2 — Entering behind the jab without living in the pocket

Shadowboxing the rhythm can prepare or maintain the movement. A fixed bag or pad sequence can build output and repeatability. Neither proves that the fighter can read distance or leave before the reply.

A lower-cost learning game can give the opponent two responses: retreat or counter after the jab. The attacker scores only by entering behind the jab and exiting at a useful angle. The defender’s genuine choice moves the task right; controlled contact keeps errors affordable. Later, broader weapons, continuation, fatigue, and meaningful scoring move the task toward transfer.

Problem 3 — Recovering from bottom position

A familiar frame–hip-escape–guard chain may maintain a solution. A partner providing predictable but substantial pressure can load the shapes, breathing, and repeated effort required to keep working: capacity. A constrained live task in which the top player may choose between passing and pinning, while the bottom player explores several escapes, belongs closer to learning. A hard positional round beginning from the compromised match position asks whether the solution transfers under representative cost.

These are not four compulsory stages. They are four possible jobs. The coach may move directly from a transfer failure to capacity, return from high consequence to learning, or use maintenance between expensive exposures. Problemming supplies the direction; the quadrant helps clarify the version.

Problemming comes before quadrant selection

There is an obvious danger here. Coaches can turn a problem-led map into another drill-storage cabinet.

“Heavy bag goes here. Playful sparring goes there. Good. Framework completed.”

No. The purpose is not to organise the exercise library more beautifully.

Start with the recurring performance problem. What repeatedly breaks, for whom, and under which conditions? What would meaningful improvement look like? Which information must remain for this still to be the same problem?

Then choose the dominant intention. Does the fighter need to learn, transfer, maintain, or build capacity?

Only then locate the nominal task. How much competition-specific information is present? What physical and consequential price is expected? Is the task representative enough for the target problem?

Now scale the functional difficulty. Who is the athlete today? Who is the partner? What is the motivational climate? Does the athlete experience this as play, learning, examination, or a threat to status?

Finally, regulate the dose and inspect what happened. Did the behaviour change? Did it survive when cost rose? What did the task cost? What should change next?

The full return path is shown in Figure 6. It is a loop because transfer and recovery do not close the case; they produce the next, better-framed problem.

This connects directly with the distinction between the Dose Coach and the Problemming Coach (Jovanović 2026d). The Dose Coach protects us from vague practice theatre. The Problemming Coach protects us from sterile programming theatre. The useful synthesis is problem-led and dose-regulated.

It also connects with Substance~Form (Jovanović 2026e). Isolate to load. Reintegrate to express. If the fighter produces more output against a stable target, the capacity change may be real. Transfer remains a hypothesis until the fighter can organise it against representative information. Conversely, if an intelligent solution collapses whenever fatigue rises, the athlete may not need another explanation. They may need enough physical substance for the form to survive.

What the map deliberately hides

Every useful map compresses. The danger begins when we forget what disappeared.

• The axes are bundles. Competition specificity and complexity can diverge. Effort, force, contact, risk, threat, and recovery cost can also diverge.

• The borders are porous. Most practice lives in clouds, transitions, and mixed intentions—not clean boxes.

• The athlete changes the task. Skill, fatigue, confidence, injury history, partner quality, and current state alter functional difficulty.

• The motivational frame matters. A coach can call something playful while athletes experience hidden selection.

• Representativeness is target-specific. A task can preserve one important relation without recreating the whole fight.

• The model cannot calculate acceptable risk. Contact exposure and safety require qualified coaching, medical judgment, rules, athlete maturity, and local governance.

• This is not a validated taxonomy. It is an evidence-informed coaching model open to inter-coach testing, criticism, and revision.

These are not reasons to discard it. A map does not fail because it leaves out the smell of the forest. It fails when the missing feature matters for the turn you are about to take.

What changes on Monday?

Take one practice task you already use. Do not invent a new drill yet.

Ask:

1. What recurring problem should the athlete solve?

2. What information must remain for this to be the same problem?

3. Where does this version sit on complexity and physical consequence?

4. Is it representative enough for the target problem—or is representativeness not its job?

5. What is the dominant intention: learn, transfer, maintain, or build capacity?

6. How functionally difficult is it for this athlete, with this partner, today?

7. What physical, psychological, recovery, and opportunity cost is justified?

8. What observation would make you keep, move, or abandon the task?

Then change one major dimension.

Preserve the information and reduce the cost. Preserve the cost and add a genuine decision. Simplify the information to load a physical resource. Return an isolated capacity to a representative problem. If useful, collect both Rating of Perceived Exertion and Rating of Perceived Challenge rather than forcing physical and technical-tactical difficulty into one number (Hendricks et al. 2019).

Do not call the new version better until you state its job.

A task is not improved because it became harder, more chaotic, more exhausting, or more realistic. It improves when its design serves the current problem at a justified cost.

That is the purpose of the Combat Quadrant.

Not to tell us where drills live.

To help us decide where practice should go next.

A working map, open to attack

The Combat Quadrant is similar to the Extended Challenge-Based Framework, but it is not the same model and certainly not a validated extension. It borrows compatible distinctions, decomposes the nominal practice task in another way, adds an explicit combat-sport cost dimension, and gives capacity work a place without asking it to cosplay as representative learning.

That makes it useful. It also makes it provisional.

I am interested in three questions:

• Which task changes position most dramatically when its rules or intention change?

• Which part of the intensity-and-consequence bundle needs to be separated?

• Where does representativeness work better as an overlay, and where would it genuinely deserve its own axis?

Try the map on a real week. Classify the task as performed. Follow one task as it moves. Then tell me where the map stops helping.

That is where the next version begins.

For the wider argument, continue with my articles on Problemming, Substance~Form, combat-practice design, and what MMA cannot reliably deliver (Jovanović 2026f, 2026e, 2026c, 2026b).

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

References

Guadagnoli, Mark A., and Timothy D. Lee. 2004. “Challenge Point: A Framework for Conceptualizing the Effects of Various Practice Conditions in Motor Learning.” Journal of Motor Behavior 36 (2): 212–24. https://doi.org/10.3200/JMBR.36.2.212-224.

Hendricks, Sharief, Kevin Till, Jon L. Oliver, Rich D. Johnston, Matthew J. Attwood, James Craig Brown, David Drake, Simon MacLeod, Stephen D. Mellalieu, and Ben Jones. 2019. “Rating of Perceived Challenge as a Measure of Internal Load for Technical Skill Performance.” British Journal of Sports Medicine 53 (10): 611–13. https://doi.org/10.1136/bjsports-2018-099871.

Hodges, Nicola J., and Keith R. Lohse. 2022. “An Extended Challenge-Based Framework for Practice Design in Sports Coaching.” Journal of Sports Sciences 40 (7): 754–68. https://doi.org/10.1080/02640414.2021.2015917.

Jovanović, Mladen. 2026a. “The Problemming Coach: Training as Problem-Setting.” Agile Periodization. March 19, 2026. https://agileperiodization.substack.com/p/the-problemming-coach-training-as.

———. 2026b. “Conditioning for MMA: What the Sport Cannot Reliably Deliver.” Agile Periodization. May 17, 2026. https://agileperiodization.substack.com/p/conditioning-for-mma-what-the-sport.

———. 2026c. “[Serbian] Veština, Haos i Agilna Praksa Borbe.” Agile Periodization. May 25, 2026. https://agileperiodization.substack.com/p/serbian-vestina-haos-i-agilna-praksa.

———. 2026d. “The Dose Coach Vs. The Problemming Coach.” Agile Periodization. May 30, 2026. https://agileperiodization.substack.com/p/the-dose-coach-vs-the-problemming.

———. 2026e. “Substance Form in the Gym: Specialists, Generalists, Bottlenecks, Barbell Strategy, and Why the Same Squat Is Not the Same Squat.” Agile Periodization. June 8, 2026. https://agileperiodization.substack.com/p/substance-form-in-the-gym.

———. 2026f. “The Ten Commandments of Problemming: A Short Manifesto for Problem-Led Planning and Prescription.” Agile Periodization. June 11, 2026. https://agileperiodization.substack.com/p/the-ten-commandments-of-problemming.

Myszka, Shawn, Tyler Yearby, and Keith Davids. 2023. “(Re)conceptualizing Movement Behavior in Sport as a Problem-Solving Activity.” Frontiers in Sports and Active Living 5: 1130131. https://doi.org/10.3389/fspor.2023.1130131.

Parry, Tom E., Shawn Myszka, Tyler Yearby, Mark O’Sullivan, and Fabian Otte. 2025. “The Value of Opposed and Unopposed Practice: An Ecological Dynamics Rationale for Skill Development.” Quest 77 (1): 80–100. https://doi.org/10.1080/00336297.2024.2420759.

Pinder, Ross A., Keith Davids, Ian Renshaw, and Duarte Araújo. 2011. “Representative Learning Design and Functionality of Research and Practice in Sport.” Journal of Sport and Exercise Psychology 33 (1): 146–55. https://doi.org/10.1123/jsep.33.1.146.

Sanli, Elizabeth A., and Timothy D. Lee. 2015. “Nominal and Functional Task Difficulty in Skill Acquisition: Effects on Performance in Two Tests of Transfer.” Human Movement Science 41: 218–29. https://doi.org/10.1016/j.humov.2015.03.006.

Yearby, Tyler, Shawn Myszka, Andy Grahn, Scott Sievewright, Adam Singer, and Keith Davids. 2024. “Applying an Ecological Dynamics Framework to Mixed Martial Arts Training.” Sports Coaching Review. https://doi.org/10.1080/21640629.2024.2325822.

The squat must increase force. The jump must improve reactive strength. The easy run must build the aerobic base. The mobility drill must restore range. The walk must produce steps. The day off must improve readiness for tomorrow.

Nothing is allowed to simply be worth doing.

This is the default grammar of training:

The solid arrows in Figure 1 are not wrong. If you coach athletes, they are part of the job. The problem begins when we pretend they are the whole job.

In Archetypes in the Gym, I argued that the same gym means different things to different people. In Domains in the Gym, I described performance, robustness, looks, and health/longevity as clusters of aspirations. I also threw fun, belonging, structure, ritual, community, and therapy-like grounding into the miscellaneous drawer called “other.”

I knew that drawer was doing too much work.

Oliver Burkeman’s Four Thousand Weeks gave me a useful label for what was hiding inside it. He describes a friend who can make running meaningful only by turning it into training for a future 10K. She cannot merely go for a run. He contrasts this with walking in the countryside, listening to music, or talking with friends—activities whose value is not postponed until some later result arrives.

Burkeman is joining a much older argument about leisure. Aristotle reversed our modern hierarchy and treated worthwhile leisure as an end for which work was a means. In The Decline of Pleasure, Walter Kerr described a culture that could barely read, socialize, or spend a weekend without making the activity useful. John Gray later asked what happens to play when everything must lead somewhere else. Different periods, different politics, same unpleasant suspicion: we have made instrumentality the admission price for using our own time.

The word is atelic.

Atelic comes from the Greek telos: an end, aim, or terminal point. This distinction matters because atelic training is not training without ideas, structure, or intelligence. It is activity that does not need a finish line to justify the time spent doing it.

This essay opens that “other” drawer. But it also makes the map slightly more annoying—which usually means more useful. The atelic cluster can be treated as another practical domain, yet it is not quite the same kind of domain as performance or health. Those domains mostly point toward a future state. Atelic value describes what the activity is while it is happening.

It is less another destination and more another layer of the map.

Telic, atelic, and the finish-line problem

The philosopher Kieran Setiya, whom Burkeman draws on, distinguishes between telic and atelic activities.

A telic activity aims at a terminal state:

• run a 10K under 40 minutes;

• add 20 kilograms to a squat;

• regain a specified range of motion;

• make the starting team;

• complete a training block;

• learn a particular movement solution.

The activity succeeds by helping to bring about its own ending. Once the target has been reached, that project is finished. Another one may immediately replace it—which is partly the problem.

An atelic activity has no terminal state that exhausts its value:

• running;

• lifting;

• playing football;

• hiking;

• moving with friends;

• practicing a craft;

• being an athlete.

You can stop doing these things. You cannot finish them in the same sense that you can finish a 10K plan or a rehabilitation milestone.

This does not make the activity aimless. A hike has a route. A game has rules and a score. A lifting session has exercises, sets, and rest periods. An atelic activity can be difficult, effortful, competitive, and highly organized. The distinction concerns where its value comes from, not whether it has local structure.

Nor must we choose one category forever. The same run can be telic and atelic at different levels:

• Training for Sunday’s 10K is telic.

• Being a runner is atelic.

• Today’s interval session may serve the first while still expressing the second.

Setiya’s argument is not that telic projects are bad. Writing a book, preparing for a competition, and helping an athlete return to sport all require them. His criticism is directed at excessive investment in telic activity—at letting projects dominate the relationship with everything we value. The mistake is not having finish lines. The mistake is living at one.

Sport makes this especially easy. There is always another competition, selection, benchmark, season, contract, or testing day. The future keeps offering the present a perfectly respectable reason to disappear.

Four distinctions we should stop mixing

Atelic is close to several familiar motivation concepts, but it is not a fancy synonym for any of them.

Figure 2 separates four questions that are usually mashed together:

1. Telic versus atelic describes the temporal structure of an activity: does it aim at a state that would complete it?

2. Controlled versus autonomous regulation describes why the person is acting: pressure, guilt, endorsement, interest, or enjoyment.

3. Delayed versus immediate reward describes when some value is experienced.

4. Training domain describes the aspiration being served: performance, robustness, looks, health, or experience in the present.

A telic goal can be autonomously chosen. An athlete may genuinely love the project of qualifying for a championship. An atelic activity can be controlled: someone may take a daily walk only because a watch, coach, or guilty conscience demands it. A session may have a delayed performance payoff and also be enjoyable now.

These constructs overlap. They are not interchangeable.

This distinction prevents two common bits of conceptual bullshit.

The first is treating atelic activity as lazy, easy, or unserious. Play can be brutally demanding. Hiking can be uncomfortable. Lifting can be enjoyable precisely because the load asks a clear question and the person answers it.

The second is assuming that enjoyment magically removes instrumental value. A run done for its own sake still produces a physiological dose. A playful small-sided game still changes perception, coordination, and conditioning. The effect may be real. It is simply not the activity’s admission ticket.

“Evidence-based” is not the enemy. Tunnel vision is.

The phrase evidence-based training sometimes behaves as if every exercise must present a passport at the border:

What variable do you target? What dose do you deliver? What adaptation do you expect? How does this transfer?

These are good questions. I ask them. You should probably ask them too.

But they privilege what can be connected to a future outcome. A workout that helps an athlete feel capable, autonomous, connected, or simply alive is harder to defend in a spreadsheet. So these values are demoted to soft benefits, adherence tricks, or pleasant noise around the “real” intervention.

This is not a necessary feature of science. Motivation research has been telling us for decades that people participate in physical activity for multiple reasons. The Motives for Physical Activity Measure–Revised, for example, distinguishes fitness, appearance, competence/challenge, social connection, and enjoyment. Self-determination theory further distinguishes controlled from more autonomous forms of motivation and emphasizes the experiences of autonomy, competence, and relatedness1.

So the evidence is not missing. The reduction often happens when evidence enters applied training culture. We retain the variable that looks most like a dose-response mechanism and drop the lived experience because it seems vague.

That is scientific tunnel vision, not science.

There is also good evidence that immediate rewards matter. In exercise settings, enjoyment and competence motives have been associated with better attendance and adherence. Across several goal-pursuit studies, Kaitlin Woolley and Ayelet Fishbach found that immediate rewards predicted persistence more strongly than delayed ones, and that making goal-directed activities more enjoyable could increase persistence.2

Useful. But watch the conceptual sleight of hand:

Enjoyment matters because it improves adherence.

We have just defended present value by converting it into a future training effect.

That may be strategically true. It is still incomplete. Defending music only because it makes people exercise longer is like defending friendship because it expands your professional network. The side effect may exist. It is not the whole value of the thing.

Enjoyment can improve adherence.

It does not need adherence’s permission.

The atelic cluster of aspirations

In my earlier domain map, performance, robustness, looks, and health/longevity mostly answered this question:

What do I want training to produce?

The atelic cluster asks a different one:

What makes this worth doing now?

The cluster in Figure 3 contains at least six recurring aspirations:

• Play and curiosity. I want to explore, improvise, and see what happens.

• Capability and mastery. I want to experience myself as someone who can do difficult things—not merely own a future result.

• Agency and choice. I want some authorship over how I move and train.

• Belonging and connection. I want to share effort, ritual, jokes, and time with other people.

• Ritual and grounding. I want a stable practice that gives shape to the week and brings attention back to the body.

• Aliveness and identity. I want to feel like an athlete, lifter, runner, player, or simply a capable human being now.

These are not necessarily one psychological construct. I am not proposing an Atelic Factor that we should validate, norm, and place in an app by Friday. This is a practitioner’s cluster: a bundle of related aspirations that are often experienced together and routinely ignored together.

And because they are cross-cutting, the same session can serve one of the familiar domains while carrying very different present value.

Two people can perform the same exercise with the same external load. One experiences surveillance, compliance, and countdown. The other experiences competence, rhythm, ownership, and connection. On the spreadsheet, the prescription is identical. In human terms, it is not the same intervention.

Alchemy in the gym

This is where Rory Sutherland’s Alchemy becomes useful. Sutherland’s recurring point is that value is not exhausted by an engineering description of the product. Context, perception, expectation, and meaning can change the value people experience even when the physical object barely changes. His TED talk on perceived value makes the point more directly: changing perception can create value rather than merely fake it.3

Training people occasionally treat this as heresy. Steel is steel. Load is load. Work is work.

Yes. And a restaurant meal is macronutrients.

The mechanical description is real. It is not complete.

A coach can change the experienced value of a session through:

• meaningful choice between roughly equivalent options;

• music, rhythm, and the social arrangement of the room;

• visible progress without turning every action into a test;

• a session arc that feels coherent rather than randomly assembled;

• rituals that mark entry into and exit from training;

• opportunities to help, watch, teach, and be taught;

• language that emphasizes capability rather than permanent deficiency.

None of this violates physiology. It changes the product the person is actually consuming.

It can also change the physical intervention indirectly. Experience affects attention, effort, persistence, communication, and willingness to return. The prescribed dose and delivered dose are not always the same. But even if the physical dose remained identical, the experiential value would still count.

This does not give us permission to rescue a stupid program with neon lights and a playlist. Psychological value is not frosting applied to poor training. Nor should “fun” become a covert manipulation for squeezing more work from people. If enjoyment matters only because it helps us capture another dose, we have repeated the same instrumental logic with better marketing.

The harder claim is simpler:

The lived experience of training is part of its value, not merely a delivery system for adaptation.

A two-ledger view of training value

Instead of forcing every session onto one scale, keep two ledgers:

1. Future payoff: What might this activity change later?

2. Present value: What makes the activity worthwhile while it is happening?

Figure 4 is not a moral ranking.

Instrumental grind is sometimes necessary. Rehabilitation has tedious parts. Technical change can require constrained repetition. Peaking for a competition involves work that would be absurd as a permanent lifestyle. The problem is not visiting this quadrant. The problem is renting an apartment there.

Craft combines future direction with present absorption, interest, or meaning. This is often the sweet spot: the athlete works toward something and finds value in the work itself.

Play and leisure may produce useful second-order effects, but the future payoff is not the reason they must be defended. This quadrant is legitimate even if the measured adaptation is small, uncertain, or zero.

Empty compliance has neither compelling present value nor credible future payoff. We retain a surprising amount of it because it looks serious.

The matrix also exposes a coaching mistake: assuming everything must sit in the top-right corner. Not every hard session can be delightful. Not every playful activity needs to transfer. Trying to optimize every experience for every outcome is how we turn life back into project management.

Bus benches, park benches, and two gears

It turns out I have already been writing around Burkeman’s distinction. I just used barbells and benches rather than philosophy vocabulary.

Dan John popularized the Bus Bench~Park Bench distinction in training.4

On a bus bench, you expect a particular bus, going to a particular destination, at a particular time. Applied to training, this means a clearly bounded phase with a priority, a timetable, and an expected result. You are trying to improve something specific, and you have accepted that the work will be judged partly by whether it does so.

On a park bench, there is no scheduled arrival. You sit, pay attention, and let the experience unfold. Applied to training, this does not mean doing nothing, avoiding effort, or wandering around the gym while updating Instagram. It means showing up and training without demanding that a particular adaptation arrive on the 14:35 bus.

In Strength Training Manual, I folded Dan John’s two benches into a broader Push~Pull model. The book uses Push the Ceiling~Pull the Floor. Push tries to force adaptation through a more aggressive progression. Pull adjusts progression to the athlete’s observed rate of adaptation. Push goes deeper into the unknown and accepts more risk, cost, and recovery. Pull expands the known territory more slowly while keeping a route back home.

I later used Raise the Floor~Push the Ceiling, and more recently Cruising~Blasting, as plainer shorthand for the same planning tension.

Cruising is the default gear. It slow-cooks qualities, accumulates repeatable exposures, preserves options, and keeps the athlete in the game. The dose is conservative enough to recover from and robust enough to survive imperfect weeks. Progression responds to feedback instead of treating the spreadsheet as a court order. You do not need to prove fitness at every session. You raise the floor beneath what the athlete can reliably do.

Blasting is a temporary campaign. You narrow the problem, choose a priority, accept more strain, and push for adaptation or realization. A pre-competition block for a strength specialist, a concentrated phase, or a short attack on a clear limitation can all be blasting. It has a reason, a timetable, a recovery bill, and—most importantly—an exit.

The manual already points toward the present argument: Push fixates more on the destination, while Pull gives more weight to process and the enjoyment of the journey. Burkeman makes me tighten that claim. A Pull/Cruise plan can still be defended only as the more sustainable route to later performance. That is clever programming, but it is still instrumental. The stronger claim is that the journey itself is one of the things the program should protect.

Most of the time, cruise. Occasionally, blast.

Do not blast every quality simultaneously unless your periodization model is based on panic.

The alignment in Figure 5 is useful, but it is not an identity.

Bus Bench, blasting, and pushing the ceiling are strongly telic. They organize training around a future state and make the route more direct, concentrated, and time-sensitive.

Park Bench, cruising, and pulling or raising the floor are more continuous, process-oriented, and tolerant of uncertainty. They create better conditions for present value because the session does not have to keep auditioning for its place in the program.

But they can still be thoroughly instrumental. If I cruise only to accumulate work capacity, reduce variance, and prepare for the next blast, cruising remains a means to an end. If I raise the floor only because it eventually permits a higher ceiling, the floor is still working for the ceiling.

The atelic question cuts across both gears:

If the expected adaptation failed to appear, would this session still have contained anything worth doing?

A hard blast can have atelic value when the athlete loves the challenge, the craft, or the temporary experience of going all in. An easy cruise can be empty compliance performed by someone counting the minutes until release. Intensity does not decide the category. The athlete’s relationship to the activity does.

This also prevents a common category error in periodization. In the manual, I argued that most work for strength-generalists—team-sport athletes who use strength rather than compete in it—should live on the Park Bench/Pull side. Strength specialists have clearer reasons to use short Bus Bench/Push phases before competition. Both groups need both gears in different proportions. The problem begins when every week becomes a bus bench and every bus is supposedly the last one before greatness.

Park Bench work gives the training process somewhere to live between campaigns. Cruising builds continuity. Raising the floor makes performance less dependent on heroic days. Then, when there is a defensible reason, we blast one or two priorities and see what the athlete can realize.

The floor supports the ceiling. It should not spend its entire life staring upward.

Obliquity: when the result arrives sideways

John Kay’s Obliquity adds another distinction. In complex systems, important goals are often reached more effectively indirectly than by marching straight at the metric.

Kay separates high-level objectives, intermediate goals, and basic actions. In a simple and well-understood problem, we may be able to define the objective, specify the means, perform the action, and collect the result. Nice work if you can get it.

In the problems that matter most, the hierarchy is rarely so polite. High-level objectives are loose and multidimensional. Goals conflict. Other people react to what we do and why they think we are doing it. Our model of the system is incomplete. Most importantly, the means help us discover and revise the ends.

Obliquity is therefore a disciplined process of experiment, discovery, and successive adjustment. It is not aimless improvisation with an academic-sounding alibi.

I explicitly imported this distinction into Strength Training Manual. Table 2.1 (page 47 in Strength Training Manual) contrasts direct and oblique decision-making, and I linked the direct route to top-down planning and the oblique route to bottom-up planning. Agile Periodization still provides direction through release and phase planning, but the working program remains a provisional vehicle for learning. Act. Observe. Update. Do not confuse a plan with advance knowledge of the result.

Direct pursuit can also distort the behavior that produced the desirable outcome in the first place.

• Chase a test score too aggressively and you start training the test.

• Chase fatigue and the session becomes a competition in looking tired.

• Chase technical perfection and the athlete stops exploring.

• Chase recovery and a day off becomes another performance task.

• Chase adherence and “fun” becomes a manipulation technique.

The oblique route begins with a direction and a provisional action. Train consistently. Try a workable dose. Solve an interesting movement problem. Play. Collect feedback. Revise the next step and, when necessary, revise the goal itself.

That process may be enjoyable and worthwhile in the present, but it does not have to be. An oblique program can still feel miserable. A direct task can still be satisfying. This is exactly why obliquity and atelic value need separate labels.

Figure 6 shows why oblique and atelic must not be treated as synonyms.

Kay does not argue that directness is always stupid. In the book’s concluding chapter, he reserves it for environments that are stable and for objectives that are transparent and sufficiently one-dimensional. If you know the problem and the relevant cause-and-effect structure, take the direct route. When the athlete, sport, calendar, and adaptation refuse to behave like a vending machine, iterate.

The conceptual stack becomes easier to manage if we keep three questions separate:

Table 1 is the antidote to collecting complementary pairs as if they were Pokémon.

Obliquity remains telic. The future outcome still matters; Kay is questioning the route toward it and showing how the route may reshape the goal. Atelic value makes the stronger claim that the activity can be worthwhile now, whether or not the outcome ever arrives.

Put more bluntly:

Obliquity says the shortest route to an outcome may not be the most effective. Atelicity says the route can still be worth walking even if the outcome never arrives.

This distinction matters because instrumentalism is very good at eating its critics.

We tell athletes to play because play improves skill. We tell them to enjoy training because enjoyment improves adherence. We tell them to rest because rest improves readiness. We tell them to cruise because cruising prepares the next blast. Each statement may be empirically defensible. Collectively, they turn play, enjoyment, rest, and cruising back into employees of future performance.

Sometimes the best way to push the ceiling is to spend a long time raising the floor without staring at the ceiling. That is obliquity.

Sometimes raising the floor is simply a satisfying way to train, even if the ceiling refuses to move. That is atelic value.

We need both ideas. The first makes planning less naive. The second keeps planning from annexing the whole activity.

Deliberate practice and playful practice

Skill acquisition offers the cleanest sporting example.

The classic concept of deliberate practice is telic by design. Practice tasks are selected to improve performance. They target a gap, demand attention, involve repetition or adjustment, and use feedback to move the performer toward a better solution5.

Deliberate play, in the sport-development literature, was positioned differently: informal, player-led, flexible, and pursued for immediate enjoyment rather than for explicit improvement.6 Skill may improve as a second-order effect, but the game is not justified by a promise of later expertise.

That contrast is useful as a prototype. Reality, as usual, refuses to respect our boxes.

In a study of 208 competitive youth athletes, Arne Güllich and colleagues found high enjoyment in both coach-led practice and youth-led play. Motives for both were multidimensional, and effort and enjoyment were largely unrelated. Their conclusion was that the features of practice and play should be treated more as dimensions than as rigid categories.

Good. That is closer to what coaches actually see.

The practical mistake is to treat play as sugar sprinkled on the skill-acquisition vegetables.

The continuum in Figure 7 is deliberately porous. A tightly designed task can be enjoyable, and playful activity can produce serious learning. The point is to vary the dimensions consciously instead of assigning moral superiority to one label.

The coach designs a “fun” game, but secretly every element is optimized for scanning frequency, decision density, high-speed running, or representative information. The athletes may enjoy it, and the design may be excellent. Yet we should be honest: this is still a telic training task wearing a colorful hat.

Protected play is different. Participants have some authorship. Rules can bend. The activity is allowed to develop in ways the coach did not preselect. It may teach something. It may reveal something. It may simply be a good fifteen minutes.

This does not require abandoning deliberate practice. It requires a practice ecology containing both:

• targeted work where improvement is explicit;

• game-like work where improvement and enjoyment coexist;

• protected play that is not hijacked by hidden conditioning targets;

• occasional exploration where the athlete can search for solutions without every attempt being corrected.

The useful coaching variables are not practice or play? but:

• How explicit is the improvement target?

• Who controls the task and its rules?

• How dense is the feedback?

• How much exploration is permitted?

• Is immediate enjoyment incidental, designed, or the point?

Sometimes the session needs to be a lesson. Sometimes it should be a laboratory. Sometimes people should just play the bloody game.

When recovery becomes another job

Recovery is where instrumental thinking performs its best magic trick: it turns not-training into more training.

The athlete finishes the session and begins a second shift:

• complete the recovery protocol;

• hit the mobility target;

• close the activity ring;

• improve the sleep score;

• perform the breathing drill;

• prepare tomorrow’s meals;

• consume “educational” content while doing all of the above.

Then we call this downtime.

To be clear, sport requires recovery. The consensus literature describes recovery as a multifaceted process, and the balance between stress and recovery matters for performance and health.7 Sleep, nutrition, load management, and sensible restoration are not capitalist propaganda. They are part of training.

But that is precisely the point: recovery is still telic. It is justified by what it restores for the future.

Rest and leisure can be different. Burkeman traces how leisure was gradually demoted from something worth choosing for itself to replenishment for more work. Sport reproduces the same hierarchy whenever a day off is allowed only because it increases readiness for the next hard day.

Research on the psychology of rest in athletes gives this distinction some empirical texture. In a qualitative study of university field-hockey athletes and staff, David Eccles and Alexander Kazmier found that wakeful rest included experiences such as not thinking about sport, often supported by activities and environments outside the normal training context.8 That is not just lower physical load. It is detachment from the sporting project.

Figure 8 does not ask us to choose between recovery and rest. Athletes need both.

The distinction does ask us to stop labeling every recovery task as leisure. A deload is a training-load manipulation. A regeneration session is a session. A day filled with mandatory recovery modalities is not necessarily a day off, even if all the activities occur at low intensity.

This is where an unscored period becomes useful:

• no requirement to produce steps;

• no readiness test;

• no mandatory mobility circuit;

• no demand to watch sport, study sport, or think about sport;

• no need to convert the time into self-improvement;

• no requirement to enjoy it correctly.

Watching a movie can simply be watching a movie. A hike can be taken without turning it into zone 2 with better scenery. Music does not need to become an ergogenic aid. The moment may have useful after-effects, but it does not owe us any.

Paradoxically, such time may need structure. Drawing on Judith Shulevitz’s essay “Bring Back the Sabbath”, Burkeman makes the point that people do not always drift effortlessly from striving into genuine rest. Rules, habits, advance preparation, and a shared norm can protect the interruption.

Schedule the boundary. Do not optimize its contents.

Periodization is telic by design

Periodization is sport’s explicit version of time management. It converts a finite calendar into phases, blocks, sessions, and doses, each expected to earn its place by contributing to something later.

Periodization arranges present actions in relation to future demands. Competition provides a date. The plan works backward. Blocks, phases, and sessions become means.

This is not a criticism. It is almost a definition.

The Programming Pyramid also moves from broad human foundations toward sport- and athlete-specific decisions. That map helps decide what training should do. It does not tell us that every minute of the athlete’s life must become training material.

The danger is total periodization:

• the offseason becomes a corrective block;

• the holiday becomes an altitude opportunity;

• leisure becomes active recovery;

• a hobby becomes psychological skills training;

• sleep becomes a score;

• social life becomes either a recovery resource or a threat to readiness.

At that point, the athlete is never off the clock. The plan has eaten the person it was supposed to serve.

Atelic value should therefore enter planning in two ways.

First, as a design criterion inside training. A session can have a clear adaptation target while also supporting capability, agency, connection, and enjoyment. This is not guaranteed, but it can be considered.

Second, as a protected limit around training. Some periods should not be captured by the plan at all. They may include play, hobbies, relationships, wandering, or simple inactivity. Their contents are not prescribed. Their non-prescription is the feature.

This applies across time scales:

• Within a session: bounded choice, exploratory repetitions, or an unscored segment.

• Within a microcycle: a genuine day away from sport tasks, not merely a low-load checklist.

• Within a longer block: space for playful practice and variation that does not require immediate transfer.

• Within a transition period: enough detachment for the athlete to inhabit roles other than performer.

• Across a career: an identity that contains sport without being fully consumed by it.

Do not ask me for the optimal percentage of atelic time. That would be impressively on-brand and completely miss the point.

A practical audit: two questions and one boundary

We do not need another readiness dashboard. We need a better conversation.

For each important session, ask:

1. What should this change later?

2. What could make it worthwhile today?

Then add a third question at the planning level:

3. What should remain unmeasured and unclaimed by the program?

The first question protects training from becoming random entertainment. The second protects it from becoming technically sound and practically dead. The third protects the person from the appetite of the system.

The branches in Figure 9 should not be collapsed into one composite score. They are prompts for judgment and conversation, not the Atelic Readiness Index™.

Consider a standard gym session for a field-sport athlete. The primary work has a clear performance rationale. The coach retains the exercises and loading that matter. But the athlete chooses between two roughly equivalent accessory options, trains one section with a teammate, and finishes with a short movement block that is not scored or converted into a competition. The core dose survives. The athlete also gets authorship, connection, and a chance to move without auditioning for the spreadsheet.

That is not a revolution. It is a better product.

Now consider the opposite case. A technically demanding session must remain tightly controlled because the timing, feedback, or risk requires it. Fine. Do not fake playfulness. Explain the purpose, do the work, and avoid pretending that every session must be entertaining.

The aim is not permanent fun. The aim is a training ecology in which the present is not always sacrificed to the future.

Useful warning signs include:

• “fun” is mentioned only as a way to improve adherence;

• every game hides a conditioning or skill KPI;

• every day off contains a recovery checklist;

• every hobby is praised for helping performance;

• every session is judged mainly by fatigue, load, or future transfer;

• the athlete cannot name anything training gives them now;

• coaches become anxious when an activity has no measurable justification.

If all value must arrive later, later becomes an increasingly crowded place.

The run does not need an alibi

The 10K matters. A personal best can organize effort, sharpen commitment, and create a memorable test. Competition is one of sport’s great inventions.

But the finish line cannot carry the value of every step that precedes it.

If each run matters only because it prepares the next race, crossing the line merely creates a vacancy for another project. The runner must keep manufacturing future accomplishments to defend the time she spends running.

The atelic alternative does not abolish the race. It changes the relationship:

I am training for a 10K, and this run is already part of my life—not a down payment on it.

The same applies in the gym. We can train for performance, robustness, looks, and health. We can use dose, adaptation, transfer, and periodization. We should.

But we can also lift because lifting is a way of inhabiting capability. We can practice because solving movement problems is interesting. We can play because the game is worth playing. We can rest without turning rest into unpaid employment for tomorrow’s self.

The program should help build the future athlete without consuming the present person.

The finish line is useful.

It is a lousy place to live.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

References and further reading

• Aristotle. Politics, Book VIII.

• Balk, Y. A., de Jonge, J., Oerlemans, W. G. M., & Geurts, S. A. E. (2019). Physical recovery, mental detachment and sleep as predictors of injury and mental energy. Journal of Health Psychology, 24(13), 1828–1838.

• Burkeman, O. (2021). Four Thousand Weeks: Time Management for Mortals. Farrar, Straus and Giroux.

• Côté, J., Lidor, R., & Hackfort, D. (2009). ISSP position stand: To sample or to specialize? International Journal of Sport and Exercise Psychology, 7(1), 7–17.

• Eccles, D. W., & Kazmier, A. W. (2019). The psychology of rest in athletes: An empirical study and initial model. Psychology of Sport and Exercise, 44, 90–98.

• Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406.

• Güllich, A., Faß, L., Gies, C., & Wald, V. (2020). On the empirical substantiation of the definition of “deliberate practice” and “deliberate play” in youth athletes. Journal of Expertise, 3(1), 1–19.

• Gray, J. (2002). Straw Dogs: Thoughts on Humans and Other Animals. Farrar, Straus and Giroux.

• John, D. (n.d.). Ramping up—Training rules for the big push. On Target Publications.

• Jovanović, M. (2017). Anti-glycolytic training for CrossFit?. Complementary Training.

• Jovanović, M. (2020). Strength Training Manual: The Agile Periodization Approach. Volume One and Two: Theory (First paperback ed., second release; ISBN 979-8604459898). Author.

• Kay, J. (2010). Obliquity: Why Our Goals Are Best Achieved Indirectly. Profile Books.

• Kellmann, M., et al. (2018). Recovery and performance in sport: Consensus statement. International Journal of Sports Physiology and Performance, 13(2), 240–245.

• Kerr, W. (1962). The Decline of Pleasure. Simon & Schuster.

• Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic definitions and new directions. Contemporary Educational Psychology, 25(1), 54–67.

• Ryan, R. M., Frederick, C. M., Lepes, D., Rubio, N., & Sheldon, K. M. (1997). Intrinsic motivation and exercise adherence. International Journal of Sport Psychology, 28, 335–354.

• Setiya, K. (2014). The midlife crisis. Philosophers’ Imprint, 14(31), 1–18.

• Shulevitz, J. (2003, March 2). Bring back the Sabbath. The New York Times Magazine.

• Sutherland, R. (2019). Alchemy: The Dark Art and Curious Science of Creating Magic in Brands, Business, and Life. William Morrow.

• Teixeira, P. J., Carraça, E. V., Markland, D., Silva, M. N., & Ryan, R. M. (2012). Exercise, physical activity, and self-determination theory: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 9, 78.

• Woolley, K., & Fishbach, A. (2016). For the fun of it: Harnessing immediate rewards to increase persistence in long-term goals. Journal of Consumer Research, 42(6), 952–966.

One of the Agile Periodization Skool members asked me to recommend books that are not directly about physical preparation, training theory, sport science, or programming, but which have nevertheless had a strong influence on my work.

The more I thought about the question, the harder it became to draw a clean border. Some of the resources below have nothing to do with sport. Others deal with learning, philosophy, psychology, statistics, complexity, writing, coding, or scientific reasoning, but have obvious consequences for coaching and performance. A few resources in the skill-acquisition section are more directly connected to practice design, but I included them because they changed the way I think about learning rather than simply giving me another collection of drills.

This is not a list of “the greatest books ever written,” and I am not endorsing every argument made by every author. Some of these authors disagree strongly with one another. That is partly the point.

I have also included a few books and intellectual detours that influenced me when I was younger, but that I would now read with substantial caution. An intellectual autobiography becomes dishonest when it includes only the respectable books we would display behind us during a Zoom call. Influence is not the same as endorsement. Some books help because they are right; some help because they are wrong in an interesting way; and some push us toward better questions only after we discover where their map breaks.

I consider myself a pluralist. By this I do not mean that every explanation is equally good, that truth is merely personal, or that we should throw incompatible ideas into an intellectual soup. I mean that complex phenomena can usually be viewed through multiple partial models, operating at different levels, built for different purposes, and emphasizing different aspects of reality. Physiology, biomechanics, psychology, ecological dynamics, statistics, subjective experience, network science, and practical coaching judgment can all reveal something useful without any one of them owning the whole athlete.

It is useful to distinguish between two forms of pluralism.

Pragmatic pluralism concerns how we frame the problem in the first place. Before asking which training method works, we need to ask what is actually going wrong, for whom, according to which criteria, and in which context. What initially appears to be a conditioning problem might instead be a pacing, tactical, technical, motivational, recovery, communication, or measurement problem.

Integrative pluralism begins once the problem has been provisionally framed. It asks which combination of partial models and perspectives can help us understand and address it. Improving repeated-sprint performance, for example, might require insights from physiology, biomechanics, skill acquisition, psychology, load management, and the athlete’s own experience.

This distinction connects directly to Problemming: Problem-Led Planning. Problemming comes before programming. The first task is not to select exercises, methods, or loading schemes, but to construct, negotiate, and continually refine the performance problem. Planning then becomes an attempt to intervene in that problem rather than an exercise in applying a preferred method.

This is also one of the central ideas behind Agile Periodization:

Models are tools, not truths.

Every model is a Small World representation of a larger, messier, and partly unknowable world. It highlights some features, compresses others, and ignores the rest. The coach’s task is not to discover the one model that destroys all competing models. The task is to develop a useful zoo of models, understand their assumptions and limitations, select the smallest useful set for the current problem, act, observe, and update.

In this sense, Agile Periodization combines pragmatic and integrative pluralism within an iterative process:

Frame the problem, select useful models, intervene, monitor, and reframe through feedback.

Pluralism does not remove judgment. It makes judgment more important.

For each section, I have tried to explain the problem the resource helps solve, the change it can induce, how it connects to performance, and roughly how difficult it is. The Amazon links are non-affiliate search links because editions and regional stores change. Wherever a legal free book, paper, course, or website exists, I have linked it directly.

If I remember something important later, I will add it.

If you are interested in more about this topic, please check the Agile Periodization: Philosophical Foundations Online Course

1. Uncertainty, Decisions, and Robust Thinking

Nassim Nicholas Taleb — The Incerto

Books: Fooled by Randomness, The Black Swan, Antifragile, Skin in the Game, and The Bed of Procrustes.

Taleb is probably the most visible philosophical influence on Agile Periodization. His work helps solve the problem of making highly confident plans in environments where predictions are fragile, rare events matter, and our models inevitably leave something out. The central change is moving from “What is the optimal plan?” toward “What happens when this plan is wrong?” In performance, this leads to robustness, optionality, via negativa, the barbell strategy, avoiding ruin, maintaining slack, and preserving the ability to adapt. I would begin with Fooled by Randomness, then The Black Swan, and then read Antifragile (although I’ve read them in the opposite order). Difficulty: moderate; Taleb can be combative, repetitive, and deliberately provocative, but the ideas are worth wrestling with.

Gerd Gigerenzer — Heuristics and Ecological Rationality

Books: Gut Feelings, Risk Savvy, Simple Heuristics That Make Us Smart, and Rationality for Mortals.

Gigerenzer helps solve the assumption that more information, more variables, and more calculation must produce a better decision. His work shows that a simple heuristic can be rational when it is matched to the structure of the environment. This is directly relevant to coaching because we rarely possess complete data, stable relationships, or unlimited time. The practical change is learning to construct simple, transparent, and adaptable rules rather than hiding uncertainty beneath mathematical complexity. Risk Savvy was particularly influential for me because it clearly separates risk, where probabilities are reasonably known, from uncertainty, where they are not. Start with Gut Feelings or Risk Savvy. Difficulty: easy to moderate.

Daniel Kahneman — Thinking, Fast and Slow

Book: Thinking, Fast and Slow.

This is a valuable and influential introduction to judgment, framing, overconfidence, heuristics, and predictable errors in reasoning. I would treat it as a good read rather than scripture, and I would read it together with Gigerenzer because their disagreement is more useful than choosing a camp. Kahneman makes us more suspicious of the stories we build around limited evidence; Gigerenzer reminds us that heuristics can be adaptive in the environments for which they evolved. In performance, this tension matters when interpreting test scores, athlete feedback, injury stories, trends, and the apparent success of a program. Difficulty: easy, although long.

Peter Bevelin — Seeking Wisdom

Book: Seeking Wisdom: From Darwin to Munger.

This is one of my favorite collections of mental models. It helps solve the problem of approaching every decision through one narrow professional framework. Bevelin draws from psychology, evolution, probability, incentives, and Charlie Munger’s emphasis on avoiding obvious stupidity before searching for brilliance. Its connection to performance is straightforward: training decisions are never merely physiological. They involve communication, incentives, uncertainty, behavior, measurement, and trade-offs. The book is dense but modular, so it can be revisited by topic. Difficulty: moderate.

Brian Christian and Tom Griffiths — Algorithms to Live By

Book: Algorithms to Live By: The Computer Science of Human Decisions.

This is one of the most influential and enjoyable books on my list. It translates ideas from computer science into everyday problems such as searching, sorting, scheduling, caching, stopping, randomness, and the explore–exploit dilemma. It changes the way you see messy human decisions: many have a recognizable structure even when they do not have a perfect solution. Explore–exploit is especially important for Agile Periodization. We need to exploit what currently works while preserving enough exploration to discover something better. Difficulty: easy and highly readable.

Richard Meadows — Optionality

Book: Optionality: How to Survive and Thrive in a Volatile World.

Meadows takes Taleb’s idea of optionality and turns it into a practical book about careers, money, habits, relationships, and the design of a life that does not collapse when one forecast fails. The problem it solves is premature lock-in: committing so much to one path that changing direction becomes ruinously expensive. In performance, optionality means preserving several viable exercises, tactical solutions, developmental routes, and scheduling options rather than pretending that the first plan is destiny. Difficulty: easy and practical.

John Kay — Obliquity and Radical Uncertainty

Books: Obliquity, Radical Uncertainty, and The Long and the Short of It.

Kay’s idea of obliquity is that complex aims are often achieved indirectly. Happiness, trust, excellence, and durable performance can become harder to reach when pursued through one crude metric. Radical Uncertainty, written with Mervyn King, fits Taleb and Gigerenzer by distinguishing calculable risk from situations where the model itself is uncertain. The performance lesson is blunt: maximizing one number can damage the system that produces the number. Difficulty: easy to moderate.

2. Model Thinking, Pluralism, and the Zoo of Models

Scott E. Page — Model Thinking

Course: Model Thinking on Coursera.
University page: Model Thinking at Michigan Online.

This might be one of the most influential courses I have taken. Scott Page introduces models from economics, mathematics, political science, sociology, networks, statistics, and complexity science. The purpose is not to make everyone a specialist in every field, but to demonstrate how different simplified models reveal different structures inside the same problem. One model may reveal incentives, another feedback, another thresholds, another contagion, another path dependence, and another coordination. For performance practitioners, the course helps prevent disciplinary tunnel vision: when physiology is your only model, every problem looks physiological; when biomechanics is your only model, every problem becomes a matter of forces and angles; when ecological dynamics is your only model, everything becomes a constraint or affordance. All can be useful. None is sufficient for every problem. Difficulty: beginner to moderate.

Scott E. Page — The Model Thinker

Book: The Model Thinker: What You Need to Know to Make Data Work for You.

The book develops Page’s many-model paradigm into a large library of mathematical, statistical, computational, and conceptual models. Its deeper message is not that we should apply every model to every problem, but that intelligent reasoning improves when we possess several models and understand how their conclusions overlap, complement, or conflict. In sport, an average, an individual trend, a causal model, a predictive model, a coach’s observation, and an athlete’s report are not interchangeable; they answer different questions. The practical sequence is simple: clarify the problem, choose a few genuinely different models, inspect their assumptions, compare what they reveal and conceal, consider the cost of being wrong, act provisionally, and update. Difficulty: moderate; take the course first and use the book partly as a reference.

Gabriel Weinberg and Lauren McCann — Super Thinking

Book: Super Thinking: The Big Book of Mental Models.

This belongs beside Page’s Model Thinking and Bevelin’s Seeking Wisdom. It is lighter, more visual, and easier to dip into. The value is not memorizing a hundred clever labels so you can sound intelligent at dinner. The value is recognizing recurring structures such as opportunity cost, feedback loops, forcing functions, bottlenecks, Goodhart’s law, and second-order effects. For coaches, it provides a practical bridge between abstract pluralism and everyday decisions. Difficulty: easy.

Farnam Street — The Great Mental Models

Resources: Farnam Street’s mental-model library and The Great Mental Models series.

Farnam Street helped popularize the idea of building a latticework of models from several disciplines rather than collecting isolated facts. The material can occasionally become “toolbox tourism,” where we admire the tools but never repair anything, so the antidote is application: take one decision, use two different models, and record what each reveals. The performance link is the same as Page’s: no single discipline should become the universal hammer. Difficulty: easy.

Sandra D. Mitchell — Unsimple Truths

Book: Unsimple Truths: Science, Complexity, and Policy.
Publisher: University of Chicago Press.

Sandra Mitchell provides a philosophical foundation for integrative pluralism. She challenges the expectation that mature science must eventually reduce every complex phenomenon to a few simple universal laws. Human beings, athletes, teams, and organizations are evolved, adaptive, historical, multilevel, and context-sensitive systems. A training response can involve tissue adaptation, coordination, perception, expectation, emotion, sleep, nutrition, prior experience, group dynamics, and coaching behavior at the same time. These levels interact, but they are not always reducible to one privileged language. The change this book induces is learning that different models may compete in some situations while complementing one another in others. Difficulty: moderate and philosophically dense.

Sandra D. Mitchell — Biological Complexity and Integrative Pluralism

Book: Biological Complexity and Integrative Pluralism.

This is the deeper and more technical companion to Unsimple Truths. Mitchell argues for integration without demanding total unification. That distinction matters to me. I want physiology, psychology, biomechanics, ecological dynamics, athlete experience, and data to inform each other; I do not need one of them to prove that it is the only real level of explanation. In performance, this protects us from turning a useful model into an ideology. Difficulty: high.

A Practical Version of Pluralism

My practical pluralism rests on a few commitments. There is one reality, but many partial representations of it. Models are built for purposes, so a map can be excellent for one task and useless for another. Description, explanation, prediction, and intervention are different jobs. Models can complement one another without being fused into one giant equation. When models disagree, that disagreement can reveal different assumptions, definitions, levels, objectives, or time scales. Finally, pluralism does not excuse indecision: the coach still has to act before certainty arrives.

Agile Periodization is partly an operating system for this pluralism. We begin with prior knowledge, models, experience, and best practices; build a Minimum Viable Plan; act in the Large World; observe what happens; and revise. A physiological model may help estimate dose, an ecological model may help design practice, a psychological model may improve communication, a statistical model may quantify uncertainty, and athlete experience may reveal context absent from the spreadsheet. The review asks which model was useful, which assumption failed, what we did not see, and what the smallest worthwhile adjustment might be.

If you are interested in more about this topic, please check the Agile Periodization: Philosophical Foundations Online Course

3. Quality, Meaning, and Philosophical Orientation

Robert M. Pirsig — Zen and the Art of Motorcycle Maintenance

Books: Zen and the Art of Motorcycle Maintenance and Lila.

Pirsig’s work influenced me more deeply than most conventional coaching or productivity books. It addresses the division between analytical knowledge and lived experience, between classical and romantic understanding, and between what can be formally defined and what is nevertheless recognized as Quality. In performance, we often act as though numbers and experience belong to different worlds. Pirsig encourages us to see them as complementary modes of engagement with the same reality. Zen and the Art of Motorcycle Maintenance is accessible as a story but demanding as philosophy; Lila is denser and develops the Metaphysics of Quality further.

Jordan B. Peterson — Early Academic Work

Book: Maps of Meaning: The Architecture of Belief.
Lectures: search for the full University of Toronto Maps of Meaning and Personality and Its Transformations lecture series on Peterson’s official lecture archive or YouTube channel (from 2017, if memory serves me well).

For me, Peterson’s earlier university work is far more representative of his intellectual contribution than newer short clips, political commentary, or the compressed internet version of his ideas. The older courses integrate mythology, personality, neuroscience, development, narrative, order, chaos, and action. The distinction between the world as a “Place of Things” and a “Forum for Action” was especially influential: an objective description of an athlete does not automatically tell us what that athlete ought to do next. The book is difficult; the long-form university lectures are the best entry point.

Joseph Campbell — Myth and Transformation

Books: The Hero with a Thousand Faces and The Power of Myth.

Campbell offers a language for understanding development as a journey through uncertainty rather than a smooth accumulation of positive adaptations. The Hero’s Journey should not be imposed as a rigid template, but it captures something important about leaving familiar territory, confronting problems, being changed by them, and returning with a new capability. In performance, this is often more realistic than imagining progress as a straight line on a graph. Start with The Power of Myth. Difficulty: easy to moderate.

Iain McGilchrist — Modes of Attention

Books: The Master and His Emissary and The Matter with Things.

McGilchrist explores contrasting modes of attending to the world: isolated parts versus living wholes, representation versus participation, certainty versus openness, and manipulation versus understanding. His work connects with the Map–Territory tension and with the danger of allowing our analytical tools to replace the phenomena they were created to illuminate. The performance connection is not a ready-made training method; it is a warning that measurement and decomposition can improve knowledge while also narrowing attention. Difficulty: high.

Adam Frank, Marcelo Gleiser, and Evan Thompson — The Blind Spot

Book: The Blind Spot: Why Science Cannot Ignore Human Experience.

This book gives language to one of my long-running problems with scientism. Science is indispensable; scientism is the belief that the third-person description provided by science exhausts reality. The Blind Spot argues that science is itself grounded in human experience and cannot coherently erase the experiencing subject from the picture. This connects directly with Peterson’s distinction between the world as a Place of Things and the world as a Forum for Action. A force plate belongs to the Place of Things. The athlete deciding whether to attack, retreat, trust, persist, or protect belongs to the Forum for Action. Both are real; neither should impersonate the other. Difficulty: moderate.

My Beef with Scientism and Pseudoscience

I have a problem with both. Pseudoscience avoids the discipline of evidence, testing, and correction. Scientism takes the success of scientific methods and inflates it into a total philosophy of life. One turns every feeling into proof; the other treats everything that cannot be measured cleanly as unreal. In coaching, pseudoscience gives us magical devices and confident nonsense, while scientism gives us dashboards pretending that meaning, trust, fear, intention, and judgment do not exist.

The alternative is not a polite fifty-fifty compromise. Evidence should constrain our claims. Experience should constrain our abstractions. The test tells us something about the Place of Things; it does not automatically close the Is/Ought gap in the Forum for Action.

Yuval Noah Harari — Sapiens and Intersubjective Reality

Books: Sapiens, Homo Deus, and 21 Lessons for the 21st Century.

I do not agree with every historical generalization Harari makes, but he is exceptionally good at changing the scale of the question. His distinction among objective, subjective, and intersubjective realities is useful. A barbell is objective; pain is subjectively experienced; money, nations, rankings, brands, and professional status are sustained through shared belief. Performance organizations live in all three realities at once. A contract may be intersubjective, fear may be subjective, and the broken bone remains stubbornly objective. Difficulty: easy and provocative.

Myth, Symbol, and the “Mystical” Register

Peterson, Campbell, and McGilchrist also opened a door toward material that might loosely be called mystical: myth, ritual, symbol, the sacred, transformation, and experiences that resist clean reduction to variables. I do not use “mystical” as permission to abandon evidence or replace explanation with vague spiritual language. I use it to acknowledge that human beings do not live only inside mechanical descriptions. They also live through stories, values, identities, symbols, commitments, and felt significance.

This matters in performance because an athlete is not merely a collection of tissues and outputs. Competition can involve fear, sacrifice, belonging, status, shame, courage, ritual, and meaning. These experiences affect action even when they cannot be captured fully by a dashboard. The pluralist response is not to choose mythology over science. It is to recognize that scientific, phenomenological, narrative, and symbolic accounts may illuminate different parts of the same human situation.

Peterson’s early Maps of Meaning material is useful for the relationship among order, chaos, narrative, and action. Campbell is useful for recurring patterns of transformation and return. McGilchrist is useful for understanding how the manner in which we attend can change the world we disclose. These resources are difficult to convert into a checklist, which is probably part of their value.

Robert Bly — Iron John

Book: Iron John: A Book About Men.

Iron John is part myth, part poetry, part psychology, and part cultural diagnosis. It is not a scientific theory of masculinity, and reading it as one would be a category error. Its value is symbolic: initiation, grief, responsibility, mentorship, anger, and the development of mature strength. It fits Campbell and Peterson better than it fits a laboratory manual. Difficulty: easy to read, harder to interpret.

Richard C. Schwartz — Internal Family Systems

Books: No Bad Parts and Internal Family Systems Therapy.

IFS offers a compassionate model of the mind as a system of “parts” with different protective roles rather than one perfectly unified rational controller. I find the model useful for understanding internal conflict without immediately turning it into moral failure: one part wants the difficult session, another wants safety, another wants approval, and another would prefer to disappear into a motorcycle catalogue. It should still be treated as a model, and serious mental-health problems belong with qualified clinicians. Difficulty: easy in No Bad Parts, technical in the clinical texts.

Joe Dolce — Modern Psychedelics (Currently Reading)

I am currently reading Joe Dolce’s Modern Psychedelics. I am not placing it on the starting shelf and I am not presenting it as a practical how-to. My interest is in the history, clinical research, cultural meaning, and the way altered states challenge simplistic models of consciousness and experience. This is an area where legal, medical, developmental, and psychological risks matter, so curiosity should not be confused with endorsement.

Norman Melchert and David R. Morrow — The Great Conversation: A Historical Introduction to Philosophy

Book: The Great Conversation: A Historical Introduction to Philosophy.

This historical introduction presents philosophy as an ongoing conversation in which major thinkers develop, challenge, and revise one another’s ideas. It combines accessible explanations with substantial excerpts from primary philosophical works, making it a useful bridge between a general overview and reading the original authors directly. For a coach, its value is not in providing ready-made answers, but in demonstrating how assumptions, concepts, and arguments evolve through disagreement. It improves the questions we ask, the problems we frame, and the reasoning we use before selecting methods. Difficulty: moderate.

4. Complexity, Chaos, and Systems

Complexity Explorer — Santa Fe Institute

Website: Complexity Explorer.

Complexity Explorer is one of the best collections of free educational material on complexity science. These courses help solve the problem of applying linear, mechanical explanations to systems composed of many interacting parts, feedback loops, delays, changing relationships, and emergent behavior. Athletes, teams, organizations, and training processes contain these features. The useful change is not to start calling everything “complex,” but to become more precise about emergence, adaptation, instability, prediction, and unintended consequences.

Melanie Mitchell — Introduction to Complexity

Course: Complexity Explorer courses.

This is the best general entry point. It introduces dynamics, chaos, fractals, information, networks, evolution, self-organization, computation, and agent-based modeling without requiring advanced mathematics at the beginning. It changes the way you think about how relatively simple local rules can create organized global behavior. For coaches, it encourages humility about direct cause-and-effect stories and greater attention to interactions. Difficulty: beginner to moderate.

David P. Feldman — Introduction to Dynamical Systems and Chaos

Course: Introduction to Dynamical Systems and Chaos.

The course covers phase space, bifurcations, sensitive dependence, attractors, and limits to prediction. The key lesson is that deterministic does not automatically mean predictable and that small changes can sometimes produce qualitative changes in behavior. In performance, this helps us think about transitions, instability, coordination, and why the same intervention may not produce the same trajectory twice. Difficulty: moderate, with some mathematics.

Bill Rand — Introduction to Agent-Based Modeling

Course: Introduction to Agent-Based Modeling.

Agent-based modeling lets us construct a system from individual agents, rules, interactions, and an environment, and then observe what emerges. This is relevant to team behavior, facility flow, athlete scheduling, equipment bottlenecks, and the interaction of individual choices inside a larger training system. It changes you from a passive user of models into someone capable of building and testing simplified worlds. Difficulty: moderate.

Melanie Mitchell — Complexity: A Guided Tour

Book: Complexity: A Guided Tour.

This is probably the best book to accompany the introductory Complexity Explorer course. It offers an accessible overview of complex systems across computation, biology, evolution, and collective behavior. The book does not give you a new periodization formula. It gives you a vocabulary for emergence, networks, adaptation, feedback, and the limitations of reductionism. Difficulty: moderate.

David P. Feldman — Chaos and Dynamical Systems

Books: Chaos and Fractals: An Elementary Introduction and Chaos and Dynamical Systems.

These provide a more formal companion to Feldman’s course and help distinguish genuine nonlinear dynamics from the loose everyday use of the word “chaos.” The performance connection lies in concepts such as attractors, stability, transitions, sensitivity, and the difference between describing the general behavior of a system and predicting its exact path. Difficulty: moderate to technical.

Steven Strogatz — Nonlinearity Made Understandable

Books: Nonlinear Dynamics and Chaos, Sync, The Joy of x, and Infinite Powers.

Strogatz is unusually good at moving between mathematics and intuition. Nonlinear Dynamics and Chaos is the technical text; Sync, The Joy of x, and Infinite Powers are more accessible. His work helps make attractors, coupling, synchronization, phase transitions, and nonlinear change feel less like mystical vocabulary and more like actual mathematical ideas. Difficulty: easy in the popular books, technical in the textbook.

James Gleick — Chaos and The Information

Books: Chaos and The Information.

Gleick is a storyteller of ideas. Chaos provides the intellectual history of nonlinear science; The Information follows the development of information as a scientific and cultural concept. These books are useful when the technical textbooks feel like being hit with a differential equation before coffee. They show how new concepts emerge, spread, and reorganize several disciplines. Difficulty: easy to moderate.

Jordan Ellenberg — Mathematical Habits of Mind

Books: How Not to Be Wrong and Shape.

Ellenberg shows that mathematics is not merely calculation but a way of seeing structure, curvature, uncertainty, and hidden assumptions. How Not to Be Wrong is particularly useful for anyone interpreting percentages, trends, averages, and apparently impressive numbers in sport. The book does not make you a mathematician, but it makes you a more difficult customer for bad quantitative arguments. Difficulty: easy.

Ronald T. Kneusel — The Art of Randomness

Book: The Art of Randomness: Randomized Algorithms in the Real World.

This is a practical Python-oriented book about using randomness in simulation, algorithms, experiments, and computation. It sits nicely between complexity, coding, and statistical thinking. Randomness is not only something to remove from data; it can be used deliberately to explore models, estimate quantities, stress-test assumptions, and build algorithms. Difficulty: moderate; basic Python helps.

Donella Meadows — Thinking in Systems

Book: Thinking in Systems.

Meadows gives one of the clearest introductions to stocks, flows, feedback loops, delays, resilience, and leverage points. It helps solve the common mistake of intervening at the most visible symptom instead of examining the system that repeatedly produces it. An injury problem may also be a scheduling problem, a communication problem, a squad-size problem, an incentive problem, or a facility problem. Difficulty: easy to moderate.

Dave Snowden — Cynefin and Sense-Making

Resource: The Cynefin framework.

Snowden’s work distinguishes clear, complicated, complex, chaotic, and confused situations. It helps solve the problem of applying the same management method everywhere. Some situations reward standard operating procedures; others reward expert analysis; genuinely complex situations require safe-to-fail probes, observation, and adaptation. For performance staff, this helps clarify when a protocol is appropriate and when experimentation is unavoidable. Difficulty: moderate.

Yaneer Bar-Yam — Making Things Work: Solving Complex Problems in a Complex World

Book: Making Things Work: Solving Complex Problems in a Complex World.

Bar-Yam provides an accessible introduction to complex-systems concepts such as emergence, networks, patterns, evolution, distributed organization, and the relationship between parts and wholes. He then applies these ideas to real-world problems in healthcare, education, engineering, conflict, and organizational design. For Agile Periodization and Problemming, the central lesson is that complex performance problems cannot be understood or solved through isolated variables, rigid hierarchies, or predetermined plans alone. The structure of the intervention must match the complexity of the problem, using distributed knowledge, iterative adaptation, and coordination across multiple levels. Difficulty: easy to moderate.

5. Dispositionalism, Causal Powers, and the Unique Athlete

Rani Lill Anjum, Samantha Copeland, and Elena Rocca — Rethinking Causality, Complexity and Evidence for the Unique Patient

Free open-access book: Download or read through OAPEN.
Amazon search: Rethinking Causality, Complexity and Evidence for the Unique Patient.

This is the best practical entry point into causal dispositionalism, and it is legally available as a free PDF. The book challenges the idea that causality is exhausted by regular associations between isolated variables. A disposition is a causal power or tendency that may manifest only when it meets suitable partners and conditions. A seed may be disposed to grow, but whether it grows depends on soil, water, temperature, damage, and many other factors. The same intervention can therefore contribute to different outcomes in different people without either the intervention or the person having “failed.”

This is deeply relevant to performance. A training method does not carry a fixed effect inside it. It has tendencies that manifest through an interaction among athlete, history, current state, execution, dose, environment, timing, expectations, recovery, and competing causal factors. The dispositional view encourages local knowledge, theoretical reasoning, epistemic humility, and methodological pluralism. It also provides a better language for individual propensities than simply attaching a population probability to one athlete. Difficulty: moderate and very readable for a philosophy-of-science book.

Stephen Mumford and Rani Lill Anjum — Causation and Powers

Books: Getting Causes from Powers, Causation: A Very Short Introduction, and What Tends to Be: The Philosophy of Dispositional Modality.

These books develop the philosophical account in more depth. The important practical change is replacing overly rigid “if A, then B” reasoning with “A tends toward B when the relevant causal partners are present and preventers are not dominant.” This does not reject prediction or evidence. It makes prediction fallible and context-sensitive, which is exactly how coaching feels in the Large World. Start with Causation: A Very Short Introduction after the open CauseHealth book. Difficulty: moderate to high.

Why Dispositionalism Matters for Performance

Dispositionalism gives us a way to speak about capacities without turning them into fixed substances. An athlete may possess a disposition to express force, tolerate pace, learn a movement, or recover from a particular dose, but manifestation depends on the total configuration of enabling and inhibiting factors. This fits the Agile idea that the same plan is not the same intervention once the athlete, context, and timing change.

It also changes how we interpret non-response. Instead of saying that a method “does not work” or an athlete “did not respond,” we can ask which causal partners were missing, which preventing factors were present, whether the intended threshold was reached, and whether our theoretical account was wrong. This is a richer starting point for individualization than simply dividing athletes into responders and non-responders after the fact.

6. Network Science and Network Psychometrics

Albert-László Barabási — Network Science

Free online book: Network Science.
Amazon search: Network Science by Albert-László Barabási.

Barabási’s book is one of the best general introductions to networks and is available free online. Network science studies systems through nodes and the relationships among them. The key change is moving from asking only what the parts are to asking how the pattern of connections influences system behavior. Degree, hubs, communities, path length, robustness, contagion, and network growth can matter as much as the average properties of individual nodes. In performance, this can apply to communication networks inside a staff, exercise relationships, symptom interactions, movement options, tactical passing structures, or the way training components influence one another. Difficulty: moderate; the opening chapters are accessible and later chapters become mathematical.

Adela-Maria Isvoranu, Sacha Epskamp, Lourens Waldorp, and Denny Borsboom — Network Psychometrics with R

Book: Network Psychometrics with R: A Guide for Behavioral and Social Scientists.
Publisher: Routledge.

This is the practical and technical book. Traditional psychometrics often assumes that an unobserved latent variable causes correlations among observed items. For example, a hidden construct called “strength” might be treated as the common cause of several tests. Network psychometrics offers another possibility: observed variables may directly or indirectly interact, and the larger construct may emerge from that network rather than sit behind it as a hidden cause. The performance consequence is significant. A cluster of correlated tests does not automatically prove that one real, stable quality causes them all, and a change in one test does not guarantee a proportional change in the others. Difficulty: technical; basic R and statistical modeling help.

Sacha Epskamp and Colleagues — Network Psychometrics

Free chapter/preprint: Network Psychometrics on arXiv.
Free tutorial paper: Estimating Psychological Networks and Their Accuracy.

Epskamp’s free material is a good way to explore the area before buying the larger book. It introduces graphical models, network estimation, and the relationship between network and latent-variable approaches. One of the most important lessons is methodological humility: estimated networks are models with sampling error and assumptions, not photographs of causal reality. In sport, attractive node-and-edge diagrams can easily become another form of dashboard theater unless the data, time scale, model, and uncertainty are appropriate. Difficulty: technical.

Schmittmann and Colleagues — “Deconstructing the Construct”

Paper: Search for “Deconstructing the construct: A network perspective on psychological phenomena”.

This paper influenced the latent-versus-observable discussion in my Strength Training Manual. It contrasts reflective, formative, and network perspectives on constructs. The reflective model treats a latent construct as the cause of observed variables. The formative model treats the construct as an aggregate formed from observations. The network model treats the construct as an emergent system of interacting components. The practical lesson is not that network models are automatically true. All three are Small World representations. The value lies in seeing that our choice of model changes what we believe a test measures, what an intervention targets, and what kind of transfer we expect. Difficulty: moderate to technical.

A Performance Example

Suppose bench press, military press, push-ups, and medicine-ball throws correlate. A reflective interpretation might propose an underlying upper-body strength or power construct causing all four scores. A formative interpretation might define a summary index from those tests. A network interpretation might focus on direct relationships among movement skill, tissue capacity, coordination, anthropometry, and task familiarity.

These models imply different interventions. If the latent construct is treated as the cause, we may believe that “training strength” should improve all manifestations. If we take a more constructionist or network perspective, we intervene on actual tasks and relationships and remain less confident that improvement will generalize. This is why test batteries and factor scores do not automatically tell us what to train.

7. Theory Building: Phenomena, Theories, Models, and Observations

Denny Borsboom and Colleagues — Theory Construction Methodology

Free paper: Theory Construction Methodology: A Practical Framework for Building Theories in Psychology.

This is one of the most useful resources on how theories are actually built rather than merely tested. The framework contains five broad steps: identify the empirical phenomena that require explanation; construct a verbal prototheory; translate explanatory principles into a formal model; test whether the model can reproduce the phenomena; and evaluate whether the theory is parsimonious, plausible, and adequate. The problem this solves is the common habit of jumping from a correlation or statistical model directly to a causal story. In performance science, we frequently possess data and methods but no explicit theory connecting components, mechanisms, and observations. Difficulty: moderate.

Eiko I. Fried — “Theories and Models: What They Are, What They Are For, and What They Are About”

Free preprint: OSF preprint.

Fried provides a compact and important distinction between theories, theoretical models, statistical models, and phenomena models. A theory explains by representing a target system; a model is a deliberately simplified representation; and a statistical model does not become an explanatory theory merely because it fits data. This matters in performance because we often mistake a regression equation, factor model, machine-learning model, or correlation network for an explanation of the system. The paper induces a healthy question: what exactly is represented here, and which assumptions connect the data model to the causal story? Difficulty: moderate.

Haslbeck, Ryan, Robinaugh, Waldorp, and Borsboom — “From Data Models to Formal Theories”

Free paper: Modeling Psychopathology: From Data Models to Formal Theories.

This paper clarifies the path from observations to scientific explanation. Data models summarize patterns in observations. Robust phenomena are stable patterns or capacities that deserve explanation. A verbal theory proposes mechanisms. A formal theory makes those assumptions explicit enough to simulate, derive predictions, and risk failure. The authors emphasize that establishing robust phenomena must come before elaborate theory building; the hardest phenomenon to explain is one that does not reliably exist. The performance connection is immediate: before constructing a grand theory of readiness, fatigue, “explosive strength,” or injury risk, first establish what the phenomenon is and whether it is robust across measurements, people, and contexts. Difficulty: technical but conceptually valuable.

Paul Smaldino — “Models Are Stupid, and We Need More of Them”

Chapter: Search the chapter and related author material.

Smaldino’s argument is that formal models are necessarily incomplete and therefore look stupid when compared with reality. Their strength is precisely that they expose assumptions, mechanisms, omissions, and contradictions. A vague verbal explanation can survive almost any observation; a formal model can be simulated and shown to fail. This is why learning to code is so powerful: it allows you to turn words into an executable model and discover where your intuition breaks. Difficulty: moderate.

A Working Vocabulary for Theory Building

I find the following distinctions useful:

Observations or data are recorded events, measurements, reports, or traces. A jump height, an RPE score, a missed session, a video clip, and an interview statement are observations. They already depend on instruments, definitions, sampling, and context.

Data models summarize patterns in observations. Means, regressions, factor models, time-series models, and estimated networks are data models. They can describe or predict without explaining the underlying system.

Phenomena are relatively stable patterns, regularities, capacities, or effects inferred from data and considered worthy of explanation. A phenomenon is not identical to one noisy data set.

Robust phenomena survive reasonable changes in measurement, sample, analysis, and context. Robust does not mean universal or exceptionless. It means stable enough that theorizing about it is not merely fitting noise.

A target system is the part of the Large World the theory aims to represent. We must decide which components and relationships are included and which are ignored.

A theory proposes explanatory principles about how the target system produces the phenomenon.

A prototheory is an early verbal account that organizes plausible components and mechanisms without yet being fully formal.

A formal model encodes enough of the theory in mathematics, logic, or computer code that we can inspect assumptions, simulate behavior, and derive consequences.

A statistical model expresses relationships in data. It can support a theory, but it is not automatically the theory itself.

A prediction is a consequence expected under specified assumptions. Predictions can test parts of a theory, but prediction alone does not establish causal explanation.

An intervention asks what might happen if we deliberately change something. This requires assumptions beyond association.

For a coach, the practical sequence is:

1. Observe something meaningful.

2. Check whether it is more than a one-off story or measurement artifact.

3. Define the phenomenon carefully.

4. Generate more than one plausible explanation.

5. Make assumptions explicit.

6. Build the simplest model capable of producing the phenomenon.

7. Compare the model with reality.

8. Intervene cautiously.

9. Observe what changes.

10. Update or abandon the model.

The plan is a hypothesis, not proof.

8. Test Theory, Measurement, and the Problem of the “True Score”

Denny Borsboom — Measuring the Mind

Book: Measuring the Mind: Conceptual Issues in Contemporary Psychometrics.

This book is a philosophical examination of what psychological tests measure and what a latent variable is supposed to be. It is especially relevant because sport science routinely borrows psychometric language—validity, reliability, constructs, true scores, latent traits—without examining the assumptions underneath it. The book helps solve the mistake of treating a mathematical decomposition as a literal map of reality. Difficulty: high, but selected chapters are extremely valuable.

Classical Test Theory

The familiar Classical Test Theory equation is:

Observed score = True score + Error

This model is elegant and useful, but the word “true” creates philosophical confusion. The true score is not automatically the athlete’s real capacity or the valid value of a construct. It is defined through hypothetical repeated measurement under specified assumptions. Those assumptions include stability, parallel measurements, and a particular treatment of error that may fit laboratory instruments better than changing human performers.

In performance, an observed bench press, sprint time, jump, wellness score, or technical rating is shaped by instrumentation, biological variation, task familiarity, strategy, motivation, fatigue, pain, environment, and actual adaptation. Calling everything except one hidden stable value “error” can erase meaningful dynamics. Measurement error matters, but variation is not always meaningless noise.

The Circular Performance Model

Article: Extending the Classical Test Theory with Circular Performance Model.
Related free material: bmbstats measurement-error chapter.

I developed the Circular Performance Model as a more phenomenological extension of Classical Test Theory for coaches and sport scientists. The aim is not to claim that it is the final true model. It is to make visible the circular relationship among capacity, manifested performance, observation, interpretation, intervention, and future performance. Tests do not passively reveal a timeless true score; testing and training occur inside a changing system. This encourages us to treat measurements as evidence within a decision process, not as direct commands. Difficulty: moderate.

What a Test Can and Cannot Tell You

A test may be reliable without measuring the construct you think it measures. It may be valid for comparing groups but poor for detecting meaningful individual change. A score may predict an outcome without identifying a useful intervention. Several correlated tests may reflect a latent common cause, a constructed index, a network of interactions, shared task demands, or a mixture of all of these.

Therefore, every test should be accompanied by practical questions:

• What phenomenon is this test intended to represent?

• What is the measurement process?

• Which sources of variation are present?

• How large is the error relative to the smallest important change?

• Is the inference about a group or an individual?

• Is the task descriptive, predictive, or causal?

• What decision will change because of the result?

• What is the cost of a false alarm or missed signal?

• Does the test add information beyond what we already know?

• What should we observe next?

A number without a decision context is often just decoration.

9. Learning How to Learn

Barbara Oakley and Terrence Sejnowski — Learning How to Learn

Course: Learning How to Learn on Coursera.
Book: A Mind for Numbers.

This is the Coursera course which introduces focused and diffuse thinking, chunking, retrieval, procrastination, memory, and illusions of competence. It solves the common problem of confusing familiarity with learning: rereading and highlighting can feel productive without producing durable knowledge. The course is useful for athletes, students, researchers, coaches, and anyone trying to learn coding or statistics. Difficulty: beginner.

Peter C. Brown, Henry L. Roediger III, and Mark A. McDaniel — Make It Stick

Book: Make It Stick: The Science of Successful Learning.

This should not be confused with Made to Stick by Chip and Dan Heath (also on the list). Make It Stick deals with retrieval practice, spacing, interleaving, generation, desirable difficulty, and the gap between immediate performance and long-term learning. This distinction is fundamental in coaching: a smooth practice session may create the appearance of competence without retention or transfer. The book changes how you study, teach, review, and schedule practice. Difficulty: easy and highly practical.

Josh Waitzkin — The Art of Learning

Book: The Art of Learning.

Waitzkin uses his experience in chess and martial arts to discuss resilience, depth, presence, pattern recognition, feedback, and performing under pressure. It is not an academic textbook; it is a reflective account of what high-level learning feels like from the inside. It connects technical refinement with adaptability and the psychological experience of competition. Difficulty: easy and enjoyable.

Anders Ericsson and Robert Pool — Peak

Book: Peak: Secrets from the New Science of Expertise.

Peak gives an accessible account of deliberate practice: clear objectives, focused effort, feedback, correction, and work beyond comfortable automaticity. I would read it critically because not every sport behaves like chess, music, or another relatively stable task. In open sport, perception, coordination, uncertainty, opponent behavior, and representative information matter. The useful lesson is not “ten thousand hours,” but that repeated activity is not the same as practice designed to change capability. Difficulty: easy.

10. Skill Acquisition, Ecological Psychology, and Practice Design

Rob Gray — How We Learn to Move

Book: How We Learn to Move.

This is probably the most accessible introduction to ecological dynamics, variability, perception–action coupling, and constraints-led coaching. It helps solve the problem of treating movement as a fixed technique copied from an ideal model rather than an adaptable solution emerging from the performer–environment relationship. Coaches will gain better questions about feedback, repetition, variability, and transfer without immediately entering the deepest theoretical literature. Difficulty: easy to moderate.

Ian Renshaw, Keith Davids, Daniel Newcombe, and Will Roberts — The Constraints-Led Approach

Book: The Constraints-Led Approach: Principles for Sports Coaching and Practice Design.

This is a practical bridge between ecological-dynamics theory and coaching. It reframes the coach as a designer of learning environments rather than someone who transfers the correct movement from the coach’s head into the athlete’s body. It helps answer how task, environmental, and individual constraints can be manipulated without turning practice into either rigid drilling or random play. Difficulty: moderate.

Chris Button, Ludovic Seifert, Jia Yi Chow, Duarte Araújo, and Keith Davids — Dynamics of Skill Acquisition

Book: Dynamics of Skill Acquisition: An Ecological Dynamics Approach.

This is a more comprehensive academic introduction to ecological dynamics. It connects complex adaptive systems, dynamical-systems theory, ecological psychology, representative learning design, self-organization, variability, and individual learning environments. The major change is shifting the unit of analysis from the isolated athlete to the athlete–environment system. Difficulty: technical.

Jia Yi Chow and Colleagues — Nonlinear Pedagogy in Skill Acquisition

Book: Nonlinear Pedagogy in Skill Acquisition.

This book focuses more explicitly on teaching and practice design. It explains why learning trajectories are nonlinear, why variability is not automatically error, and why practice should preserve meaningful information from the performance environment. It connects naturally with problemming, representative practice, adaptability, and transfer. Difficulty: technical but applied.

James J. Gibson — The Ecological Approach to Visual Perception

Book: The Ecological Approach to Visual Perception.

Gibson is one of the philosophical foundations beneath ecological dynamics. The concept of affordances shifts attention from neutral objects to possibilities for action relative to a performer. In sport, athletes do not merely process objects; they perceive gaps, grips, surfaces, distances, timing opportunities, passing lanes, and opponent vulnerabilities. The book changes how we think about information and perception, but it is not the easiest starting point. Difficulty: high.

Where the Constraints-Led Approach Meets Pluralism

I find the ecological and constraints-led perspective very useful, but I do not treat it as a Grand Unified Theory. It is especially powerful for representative practice, perception–action coupling, adaptability, and the design of problems that invite useful behavior. It is less directly suited to every question about tissue loading, recovery, organization, motivation, or long-term resource allocation.

A pluralist coach can use ecological dynamics to design the practice task, physiology to estimate cost, biomechanics to inspect loading, psychology to support communication, and athlete feedback to understand lived experience. The useful question is not “Which school wins?” but “Which model helps with this part of the problem, and where does it stop helping?”

A sensible reading sequence is:

1. How We Learn to Move

2. The Constraints-Led Approach

3. Dynamics of Skill Acquisition

4. Nonlinear Pedagogy in Skill Acquisition

5. Gibson, if you want the deeper philosophical foundation

11. Psychology, Behavior, and Change

Jonathan Haidt — The Rider and the Elephant

Books: The Happiness Hypothesis and The Righteous Mind.

Haidt’s rider-and-elephant metaphor is one of the most useful ways to think about the relationship between conscious reasoning and automatic, emotional, embodied behavior. The rider can plan and explain, but the elephant supplies much of the energy and direction. This matters whenever coaches assume that giving better information will automatically change behavior. Knowledge matters, but habits, identity, trust, emotion, environment, and belonging often matter more. Difficulty: easy.

Chip Heath and Dan Heath — Switch

Book: Switch: How to Change Things When Change Is Hard.

Switch converts Haidt’s metaphor into a practical framework: direct the rider, motivate the elephant, and shape the path. It helps solve the problem of treating resistance to change as laziness or stupidity. A new monitoring system, gym routine, nutrition behavior, or staff process can fail because the environment makes the desired action difficult. The book pushes us to redesign the path rather than merely lecture the person. Difficulty: easy and practical.

Steve Peters — The Chimp Paradox

Book: The Chimp Paradox.

This is not a perfect scientific model of the mind, but it is a memorable and usable metaphor for emotion, impulse, self-regulation, and internal conflict. Its strength is communicative: athletes and coaches can use the language quickly without a degree in psychology. Its limitation is the temptation to treat the metaphor as literal neuroscience. Difficulty: easy.

Robert Sapolsky — Human Behavioral Biology

Course: Stanford Human Behavioral Biology lectures.
Book: Behave.

Sapolsky is one of the best lecturers I have encountered. His work examines behavior across several time scales: what happened a second before, hours before, during development, across culture, and throughout evolutionary history. This discourages single-cause explanations of motivation, aggression, fatigue, anxiety, and decision-making. For anyone who works with humans, the full lecture series is worth the time. Difficulty: the course is accessible; Behave is long but readable.

Dan Ariely — Irrational Behavior

Course and books: Search for Dan Ariely’s behavioral-economics books.

Ariely’s older course on irrational behavior was an entertaining introduction to how framing, defaults, social norms, and context influence decisions. I would treat the material as a useful entry into behavioral economics, not as the final authority on human behavior. The practical connection is that athletes, staff, and customers do not respond only to information; they respond to how choices are structured. Difficulty: easy.

Books I Read When I Was Younger: Pickup, “Red Pill,” and the Long Way Around to Psychology

When I was younger, books such as The Mystery Method, Neil Strauss’s The Game, and parts of the later “red pill” literature were useful in a limited and slightly embarrassing way: they pushed a shy young man to meet people, notice social dynamics, and stop assuming that attraction and communication were entirely random. They were also seductive because they offered simple rules for a complicated domain.

I would not recommend reading them as manuals for manipulating women or as comprehensive theories of human relationships. Their better contribution was sending me toward psychology, evolutionary psychology, behavioral biology, status, signaling, attachment, and communication. Their shadow side was adversarial thinking, overconfident stories about men and women, and the temptation to convert frustration into ideology.

I remain sharply critical of some strands of contemporary gender politics and of explanations that treat “patriarchy” as the universal cause of everything. I am equally cautious about replacing that with an equally totalizing “gynocracy explains everything” model. Both can become ideological compression algorithms: impressive because they explain everything, useless because they can explain anything. Keep the useful lessons about agency, incentives, responsibility, and social courage; discard contempt, manipulation, and tribal certainty.

Sapolsky’s free behavioral-biology lectures were a much better next step. They widened the frame from tactics to biology, development, environment, culture, and evolutionary history. The journey from pickup books to behavioral biology was not the most elegant academic pathway, but few real learning paths look elegant in retrospect.

12. Productivity and Personal Workflow

Daniel Markovitz — A Factory of One

Book: A Factory of One.

This remains one of my favorite productivity books. It applies Lean principles to individual knowledge work: visualize work, reduce waste, standardize repeatable processes, limit work in progress, and improve flow. It solves the problem of overloaded and poorly designed workflow rather than blaming a lack of motivation. For coaches, researchers, writers, and independent creators, it helps turn scattered responsibilities into a visible operating system. Difficulty: easy and immediately practical.

Daniel J. Levitin — The Organized Mind

Book: The Organized Mind.

The main lesson I took from this book was to externalize memory. Trying to keep projects, reminders, unresolved decisions, and obligations active in your head consumes attention. The practical change is building trusted systems for capture, organization, review, and retrieval, leaving more mental room for coaching, writing, analysis, and problem-solving. Difficulty: easy, although longer than necessary.

Jim Benson and Tonianne DeMaria Barry — Personal Kanban

Book: Personal Kanban.

The two key ideas are visualizing work and limiting work in progress. A Kanban board makes invisible commitments visible and exposes bottlenecks, overload, and unfinished work. This applies to athlete-management projects, testing, reports, rehabilitation tasks, content production, and staff collaboration. Difficulty: easy.

David Allen — Getting Things Done

Book: Getting Things Done.

GTD provides a workflow for capturing inputs, deciding what they mean, defining the next action, placing them in a trusted system, and reviewing that system. It solves the anxiety created by vague, unprocessed commitments. The most important idea is not a particular app or folder structure; it is separating capture, clarification, organization, and execution. Difficulty: easy to moderate.

Stephen R. Covey — The 7 Habits of Highly Effective People

Book: The 7 Habits of Highly Effective People.

Covey connects daily work with roles, values, responsibilities, and longer-term direction. Productivity systems can make us efficient at doing things that do not deserve to be done. Covey’s contribution is putting prioritization, responsibility, and values above mere speed. Difficulty: easy.

Paul Gibbons — Reboot Your Life

Book: Search for Reboot Your Life by Paul Gibbons.

This is a short, practical book on organizing areas of responsibility, goals, projects, and tasks. It helped provide a simple hierarchy for connecting daily actions to larger commitments. The performance connection lies in reducing role confusion and making projects actionable rather than keeping them as vague ambitions. Difficulty: easy.

J. D. Meier — Getting Results the Agile Way

Resource: Search for the book and the free Agile Results material.

This applies agile ideas to personal productivity through short planning horizons, weekly outcomes, reflection, and adjustment. It is repetitive in places, but the key idea fits Agile Periodization: do not design a perfect personal system before acting; create a usable version, review it, and update it. Difficulty: easy.

Cal Newport — Focus and Sustainable Knowledge Work

Books: Deep Work, So Good They Can’t Ignore You, Digital Minimalism, A World Without Email, and Slow Productivity.

Newport’s work addresses the mismatch between valuable work and environments dominated by messaging, shallow tasks, fragmented attention, and visible busyness. Deep Work is the best entry point. Coaches and researchers need uninterrupted attention for programming, analysis, coding, writing, and learning difficult material. The practical change is protecting focus deliberately instead of waiting for free time to appear. Difficulty: easy.

Oliver Burkeman — Time, Finitude, and Limits

Books: Four Thousand Weeks, Meditations for Mortals, and The Antidote.

Burkeman is a useful antidote to productivity systems that quietly promise total control. His work reminds us that time is finite and choosing one commitment necessarily means neglecting another. The goal is not to complete everything but to choose what deserves limited attention. In performance, endless testing, monitoring, exercises, content, and interventions can crowd out the work that matters. Difficulty: easy and reflective.

Money, Optionality, and the Motorcycle Problem

I should probably read personal-finance books more often because motorcycles remain an extremely effective intervention for converting savings into noise, suspension parts, and stories.

Morgan Housel — The Psychology of Money

Book: The Psychology of Money.

Housel’s central contribution is that money behavior is not solved by intelligence alone. Time horizon, envy, risk tolerance, status, luck, and personal history shape decisions. The performance connection is surprisingly direct: a technically optimal plan that ignores behavior is not optimal in practice. Difficulty: easy and excellent.

J. L. Collins — The Simple Path to Wealth

Book: The Simple Path to Wealth.

Collins argues for simple, low-cost, long-term investing and avoiding unnecessary complexity. I like the philosophy even when my garage occasionally files a formal objection. The useful change is moving from prediction and constant tinkering toward a robust system with fewer opportunities for self-sabotage. Difficulty: easy.

Aaron Clarey — Bachelor Pad Economics

Book: Bachelor Pad Economics.

I enjoyed parts of this book as a blunt personal-finance manual aimed at young men. The budgeting, incentives, career, and downside-protection material can be useful; the broader social and gender commentary should be read critically rather than adopted as a package. My summary would be: useful arithmetic, uneven sociology. Difficulty: easy.

Money Books as Via Negativa

The main lessons are boring because the useful parts of money management are often boring: spend less than you earn, avoid ruin, understand fees, preserve optionality, insure the catastrophic downside, and allow compounding to work. The difficult part is following these principles while websites keep showing me a very reasonably priced motorcycle that “will hold its value.”

13. Agile Project Management, Lean, DevOps, and Visual Management

Agile Periodization did not emerge only from training theory. It was heavily influenced by project management, software development, Lean production, Scrum, Kanban, visual management, and later DevOps. These fields faced a familiar problem: how do we coordinate people and produce valuable work when requirements change, information arrives late, plans become obsolete, and failure is discovered only after large commitments have already been made?

The transferable lesson is not that a training week is software or that athletes are products. The lesson is procedural: work in shorter cycles, make priorities visible, limit unfinished work, build feedback into the process, review outcomes and methods separately, automate repeatable tasks, and preserve the ability to change direction.

My older Complementary Training material on Scrumban boards, managing teams with Trello, and the Agile Periodization Manifesto shows how directly these ideas entered my coaching framework.

Ken Schwaber and Jeff Sutherland — The Scrum Guide

Free guide: The official Scrum Guide.

The Scrum Guide is the shortest and best place to understand the formal framework before reading interpretations of it. Scrum organizes complex work around a clear goal, a prioritized backlog, short sprints, frequent inspection, and adaptation. The important performance lesson is not to copy software ceremonies literally. It is to separate long-term direction from short-term commitment and to create regular moments where the plan, results, and process can be reviewed. Difficulty: beginner and short.

Kenneth S. Rubin — Essential Scrum

Book: Essential Scrum: A Practical Guide to the Most Popular Agile Process.

This is one of the clearest visual explanations of Scrum and one of the books I used when translating agile ideas into planning for performance. It solves the problem of knowing the vocabulary without understanding how goals, backlogs, sprint planning, daily coordination, review, and retrospective fit together. The coaching equivalent is a transparent system where everyone knows what matters now, what is waiting, what is blocked, and what has changed. Difficulty: easy to moderate.

Andrew Stellman and Jennifer Greene — Learning Agile

Book: Learning Agile.

This is useful because it compares several agile approaches rather than presenting one framework as a religion. It reinforces a pluralist lesson: Scrum, Kanban, Lean, and related methods solve overlapping but different problems. In coaching, a fixed framework should not become more important than the reason it was adopted. Difficulty: easy to moderate.

Jeff Sutherland — Scrum

Book: Scrum: The Art of Doing Twice the Work in Half the Time.

This is the popular, story-driven introduction. It is less precise than the official guide or Rubin’s book, but it communicates why large batches, hidden problems, delayed feedback, and rigid plans create waste. The useful change is learning to release smaller versions, test assumptions earlier, and improve the system rather than heroicly rescuing it at the end. Difficulty: easy.

David J. Anderson — Kanban

Book: Kanban: Successful Evolutionary Change for Your Technology Business.
Free overview: Essential Kanban Condensed.

Kanban begins with the current process rather than demanding a complete reorganization. Work is visualized, work in progress is limited, flow is managed, policies are made explicit, and changes are introduced evolutionarily. This is highly relevant to performance departments, where too many simultaneous projects, unclear ownership, and invisible blockers create more trouble than lack of expertise. A board is not decoration; it should reveal queues, overload, blocked work, and neglected athletes. Difficulty: moderate.

Jim Benson and Tonianne DeMaria Barry — Personal Kanban

Book: Personal Kanban.

The two central rules are simple: visualize your work and limit work in progress. This applies equally well to an individual coach, a research project, or a multidisciplinary performance team. The deeper change is learning that starting more work is not the same as completing more valuable work. Difficulty: easy.

Gwendolyn Galsworth — Visual Workplace/Visual Thinking

Book: Visual Workplace/Visual Thinking.
Author resources: Visual Workplace books.

Visual management is more than placing colorful cards on a wall. The aim is to embed essential information in the environment so that normal, abnormal, ready, blocked, safe, and unsafe states become easier to perceive. In a performance setting, a good visual system reduces the amount of information people must remember and the number of meetings required to discover basic facts. Difficulty: moderate.

James P. Womack and Daniel T. Jones — Lean Thinking

Book: Lean Thinking.

Lean asks us to identify value, see the value stream, improve flow, allow work to be pulled by demand, and pursue continuous improvement. The relevant performance lesson is not mindless efficiency. It is removing activity that consumes time and attention without improving athlete care, learning, or decisions. Lean also helps expose the cost of handoffs, waiting, rework, duplicated data entry, and excessive inventories of unfinished reports. Difficulty: moderate.

Niklas Modig and Pär Åhlström — This Is Lean

Book: This Is Lean: Resolving the Efficiency Paradox.

This book makes an important distinction between resource efficiency and flow efficiency. Keeping every staff member busy can make the total system slower if work accumulates in queues and athletes wait for decisions. Performance environments often optimize the utilization of specialists while damaging the flow of care. Difficulty: easy and concise.

Eric Ries — The Lean Startup

Book: The Lean Startup.

The Build–Measure–Learn loop and Minimum Viable Product strongly influenced my Minimum Viable Plan and Minimum Viable Performance concepts. The point is not to deliver something poor. It is to create the smallest responsible version capable of testing the most important assumption, learn before making a large commitment, and avoid polishing something that nobody needs. Difficulty: easy.

Gene Kim, Kevin Behr, and George Spafford — The Phoenix Project

Book: The Phoenix Project.

This novelized introduction to DevOps follows an organization overwhelmed by urgent work, dependencies, bottlenecks, rework, and fragile systems. The story is imperfect, but the flow problems are recognizable to anyone who has worked in a multidisciplinary performance department. Its main value is showing that local busyness can coexist with poor system performance. Difficulty: easy.

Gene Kim, Jez Humble, Patrick Debois, John Willis, and Nicole Forsgren — The DevOps Handbook

Book: The DevOps Handbook.

DevOps extends Lean and agile thinking into operation: shorten feedback loops, reduce batch size, automate repeatable work, build quality into the process, monitor the real system, and create shared responsibility instead of throwing work across silos. For sport, the analogy is not technical deployment but integrated delivery. Testing, reporting, communication, and implementation should be part of one feedback system rather than separate departments producing files for one another. Difficulty: moderate.

Nicole Forsgren, Jez Humble, and Gene Kim — Accelerate

Book: Accelerate: The Science of Lean Software and DevOps.

Accelerate is especially interesting because it attempts to evaluate organizational practices empirically rather than relying only on stories. It emphasizes delivery speed, stability, feedback, and organizational capabilities. The performance lesson is that speed and quality are not always opposites; well-designed systems can improve both by reducing handoffs, large batches, hidden defects, and recovery time. Difficulty: moderate.

Peter Compo — The Emergent Approach to Strategy

Book: The Emergent Approach to Strategy: Adaptive Design and Execution.
Author resources: The Emergent Approach.

Compo distinguishes strategy from aspirations, plans, and tactics, and focuses attention on the central rule that addresses the bottleneck blocking the desired outcome. This fits Agile Periodization because a strategy should guide adaptation rather than pretend to predict every step. In coaching, “improve performance” is an aspiration; “do more strength work” is a tactic; the strategic question is which bottleneck currently prevents progress and what guiding rule helps us attack it. Difficulty: moderate.

Dominica DeGrandis — Making Work Visible

Book: Making Work Visible.
Author resources: Making Work Visible.

DeGrandis identifies common thieves of time: too much work in progress, unknown dependencies, unplanned work, conflicting priorities, and neglected work. This book is especially useful for performance departments because invisible work creates imaginary capacity. Everyone appears available until the season, rehabilitation plans, reports, travel, meetings, and urgent requests are placed on one board. Visual management does not solve overload by itself; it makes it harder to lie about it. Difficulty: easy and practical.

Dan Charnas — Everything in Its Place / Work Clean

Books: Everything in Its Place and Work Clean.

These are essentially the same book under different titles; Everything in Its Place was previously published as Work Clean. Charnas is not the chef whose autobiography we are reading. He studied professional kitchens and translated mise-en-place—preparing, arranging, finishing, cleaning as you go, and maintaining readiness—into a system for knowledge work and life.

The connection to coaching is stronger than it first appears. Before a session starts, equipment, information, roles, progressions, contingencies, and communication channels should be in place. During the session, the environment should support flow rather than force constant searching and improvisation. “Everything in its place” is physical, informational, and mental visual management. Difficulty: easy and practical.

What I Took from Agile and Lean

The concepts that transferred most strongly into Agile Periodization are short planning cycles, prioritized backlogs, visible work, limits on work in progress, embedded testing, Minimum Viable Plans, daily alignment, review, retrospective, standard work where useful, safe-to-fail experiments, automation, and continuous improvement.

Agile does not mean having no plan. Lean does not mean doing everything with fewer people. Scrum does not mean holding more meetings. Kanban does not mean moving colorful cards. DevOps does not mean buying another platform.

The point is to create a system that makes reality visible early enough to do something about it.

14. Notes, Writing, and Building Knowledge

Sönke Ahrens — How to Take Smart Notes

Book: How to Take Smart Notes.

This book introduces the Zettelkasten approach through the example of Niklas Luhmann. It solves the problem of accumulating highlights and documents that never become new ideas or finished writing. Instead of merely filing notes by topic, you write ideas in your own words, connect them, and allow larger arguments to emerge from the network. This is useful for books, articles, courses, research projects, coding notes, and long-term intellectual development. Difficulty: easy conceptually; the real challenge is consistent use.

Bob Doto — A System for Writing

Book: A System for Writing.
Website: Bob Doto’s writing site.

Bob Doto offers a more concrete and writer-oriented explanation of Zettelkasten. Doto focuses on turning notes into sequences, arguments, drafts, and finished work rather than building a beautiful note archive. This is especially useful for people who read and collect a great deal but struggle to produce (ahem, me). Difficulty: easy and practical.

Tiago Forte — Building a Second Brain and The PARA Method

Books: Building a Second Brain and The PARA Method.
Free overview: Building a Second Brain and PARA.

Forte focuses on capturing useful material, organizing it around action, distilling it progressively, and expressing it in new work. PARA - Projects, Areas, Resources, Archives - is useful because it organizes information by current relevance rather than by an imaginary permanent taxonomy. The warning is the same as with Zettelkasten: a second brain can become a second procrastination system. Notes should eventually become decisions, code, articles, courses, or better coaching. Difficulty: easy.

Chip Heath and Dan Heath — Made to Stick

Book: Made to Stick.

Made to Stick explains why concrete, unexpected, credible, emotional, and story-shaped ideas are remembered while technically accurate abstractions disappear. This is useful for coaching education, presentations, manuals, feedback, websites, and explaining complex concepts to athletes. Difficulty: easy.

Writing as Learning

Writing is not merely the final stage where completed ideas are reported. Writing is part of thinking. A good note system, a reproducible document, and version control make it possible to trace how an idea developed, where assumptions changed, and which evidence altered the argument. This is why learning Quarto, Git, and basic coding improves more than technical productivity. These tools change the quality of thought by making models, evidence, and revisions visible.

The note system must serve output. When maintaining the system becomes a substitute for writing, the system has failed.

15. Coding as a Professional Superpower

Learning to code is one of the strongest recommendations I can make.

Coding allows you to move beyond consuming tools built by other people. You can clean your own data, simulate ideas, inspect assumptions, automate repetitive work, build dashboards, create small applications, produce reports, and make parts of your thinking executable.

Modern AI has lowered the entry barrier, but basic understanding remains important. Without it, you cannot reliably judge whether generated code is correct, understand why it failed, protect your data, or modify the result beyond superficial prompts.

Learning R was one of the most important professional-development decisions I made. It improved my statistics, visualizations, writing, simulation, research, and ability to build packages and applications. If I were starting again today, I would probably begin with Python because of its broad use in automation, AI, general programming, machine learning, and web development. I would still recommend R when the immediate goal is statistics, visualization, research, and reproducible reporting.

Do not spend six months watching tutorials without building anything. Choose a small real problem: clean an athlete-testing file, create one useful graph, automate a report, simulate a training idea, build a simple calculator, or reproduce an analysis from a paper. Learn what you need while making the project.

Starting with Python

Free resources: Python for Everybody and Automate the Boring Stuff with Python.

Python for Everybody provides a gradual introduction to programming and computational thinking. Automate the Boring Stuff quickly demonstrates how code can manipulate files, spreadsheets, text, and repetitive processes. Together they solve the problem of learning syntax without seeing why it matters. Difficulty: beginner.

Starting with R

Free book: R for Data Science.
My training: Complementary Training courses.

R for Data Science gives a practical route through import, cleaning, transformation, visualization, and reproducible workflows. It teaches programming through actual analytical problems. R is particularly strong when analysis, visualization, statistics, and reporting need to live together. Difficulty: beginner to moderate.

Hadley Wickham — Advanced R and Package Development

Free books: Advanced R and R Packages.
Book publishing: bookdown.

These resources gave me the skills and confidence to package ideas rather than leave them as one-off scripts. A package forces you to define functions, inputs, outputs, documentation, tests, and reusable structure. This changes coding from solving today’s problem into building a tool that future you and other people can use. Difficulty: intermediate to advanced.

Hadley Wickham and the R Ecosystem

Free books: R for Data Science, Advanced R, R Packages, and ggplot2: Elegant Graphics for Data Analysis.

Hadley Wickham did not merely write packages; he helped create a coherent grammar for data work. R for Data Science is the best practical start, ggplot2 teaches a grammar of graphics, Advanced R explains the language more deeply, and R Packages shows how one-off scripts become reusable tools.

The R community is unusually open, diverse, and often politically progressive. I do not need to agree with every cultural signal or conference sticker to recognize excellent software, documentation, and generosity. Code should be judged by whether it works, is understandable, and survives contact with new data - not by how its author votes. Difficulty: beginner in R for Data Science, advanced in the others.

Shiny

Resource: Shiny for R.

Shiny turns R analyses into interactive web applications without requiring you to become a full-time front-end developer. It is useful when coaches or athletes need to upload data, change assumptions, explore scenarios, or receive individualized reports. The important change is moving from a private analysis toward a usable decision tool. Difficulty: moderate.

16. Quarto and Reproducible Documents

Website: Quarto.

Quarto is an open-source publishing system that combines prose, references, code, results, tables, and figures inside the same source document. It can produce articles, reports, presentations, websites, dashboards, and books using R, Python, Julia, and other computational tools.

The problem Quarto solves is the separation between analysis and reporting. In a traditional workflow, data are analyzed in one program, graphs are exported manually, numbers are copied into Word or PowerPoint, and every update requires repeated editing. This is slow and creates opportunities for error.

In a reproducible document, the code that creates a result lives beside the explanation of that result. When data or analysis changes, the document is rendered again and the tables, figures, and numerical results update from the source.

For performance work, this produces fewer copy-and-paste errors, transparent calculations, repeatable athlete reports, consistent formatting, easier collaboration, and an auditable link between raw data and conclusions. A single source can generate HTML, PDF, Word, slides, or a website.

A useful first project is a one-page report containing a title, one imported data file, one cleaning step, one table, one figure, one automatically calculated result, and references. Do not begin by trying to write a 400-page book. Difficulty: beginner to moderate once Markdown is familiar.

17. Git, GitHub, and Version Control

Free book: Pro Git.
Interactive lessons: GitHub Skills.
GitHub: github.com.

Git records meaningful versions of a project. GitHub provides a place to store, share, discuss, and collaborate on those projects.

The problem is familiar: files named analysis-final, analysis-final-2, analysis-final-real, and analysis-final-real-corrected. We lose track of what changed, why it changed, and which version generated a particular result.

Git lets you create a history of changes, return to an earlier working version, inspect what was added or removed, experiment on a separate branch, collaborate without emailing files, and connect code, documentation, issues, and publications. This matters even when working alone. Your most unreliable collaborator is often your future self, who opens the project six months later and cannot remember why a decision was made.

You do not need to master advanced commands. Begin by creating a repository, adding files, making one meaningful commit, viewing the history, pushing it to GitHub, and learning how to restore a previous version. Use a graphical interface at first if the command line creates unnecessary friction.

Used together, code, Quarto, Git, and GitHub create a powerful workflow:

Data + analysis + explanation + references + version history = a reproducible project.

Difficulty: confusing during the first few days and enormously valuable afterward.

18. Zotero and Reference Management

Website: Zotero.
Quick-start guide: Zotero documentation.
Quarto citations: Quarto citation guide.
Optional text-based workflow: Better BibTeX for Zotero.

A reference manager is not optional once scientific reading and writing become serious. Zotero collects bibliographic metadata, stores PDFs and notes, organizes sources, and generates citations and reference lists. The problem it solves is not merely formatting. It creates a traceable library of where claims came from and prevents references from being rebuilt manually in every document.

Zotero works well with Word, Google Docs, LibreOffice, and Quarto. With Quarto, references can live in a bibliography file and be cited using stable keys. Better BibTeX is useful for people working heavily with Markdown, LaTeX, Git, and reproducible documents because it can maintain more predictable citation keys and exports.

My suggested minimal workflow is:

1. Save every serious source to Zotero rather than leaving it in browser tabs or random folders.

2. Correct the title, authors, year, DOI, and publication metadata when importing.

3. Attach the legal PDF or a link to the source.

4. Add a short note explaining why the source matters and which project it supports.

5. Use tags sparingly; collections and search are usually enough.

6. Insert citations from Zotero into Word or export a bibliography for Quarto.

7. Keep the manuscript, code, bibliography, and figures in the same Git repository when possible.

AI can help summarize and search literature, but it should not be trusted to invent bibliographic records from memory. A curated reference library remains the source of truth for the sources you actually used.

Difficulty: beginner. The return on the first hour of setup is enormous.

19. Statistics, Models, and Scientific Reasoning

Richard McElreath — Statistical Rethinking

Book: Statistical Rethinking.
Lectures and resources: Richard McElreath’s site.

This is one of the most influential technical books I have read. It teaches Bayesian statistics, but more importantly it teaches generative reasoning: what process might have produced these observations, which assumptions does the model make, and what would we expect to see if the model were true? It replaces the question “Which test should I run?” with “What is my model of the process, what did I believe before seeing the data, and how should the observations update that belief?” Athlete data are noisy, hierarchical, incomplete, and context-dependent, so this language is natural for performance work. Difficulty: technical, but McElreath is an outstanding teacher.

Andrew Gelman, Jennifer Hill, and Aki Vehtari — Regression and Other Stories

Book: Regression and Other Stories.
Free author PDF: Regression and Other Stories.

This book treats regression as a practical language for comparisons, predictions, adjustment, uncertainty, and causal questions rather than a formula that returns coefficients. It is less theatrical than Statistical Rethinking but exceptionally useful. The change is learning to build, check, criticize, and communicate models instead of simply reporting significance. Difficulty: technical.

David Spiegelhalter — The Art of Statistics and The Art of Uncertainty

Books: The Art of Statistics and The Art of Uncertainty.

Spiegelhalter is one of the best statistical communicators. The Art of Statistics explains how data can be used to learn about the world without pretending that analysis is automatic. The Art of Uncertainty broadens the frame to chance, ignorance, luck, and communication. These books are ideal for coaches who need statistical literacy without immediately entering a Bayesian monastery. Difficulty: easy to moderate.

Geoff Cumming — Understanding the New Statistics

Book: Understanding the New Statistics.

Cumming emphasizes effect sizes, confidence intervals, estimation, meta-analysis, and visual reasoning rather than binary worship of the p-value. This was highly influential for magnitude-based thinking. I would still combine it with Bayesian, causal, and decision-theoretic perspectives, but it is an important corrective to “significant or nothing happened.” Difficulty: moderate.

Peter Bruce, Andrew Bruce, and Peter Gedeck — Practical Statistics for Data Scientists

Book: Practical Statistics for Data Scientists.
Code: Official code repository.

This is a pragmatic bridge among statistics, Python, R, and data-science workflows. It covers exploratory analysis, sampling, experiments, regression, classification, and machine learning without becoming a pure theory book. It is useful when you know enough terminology to be dangerous but need to see how the pieces fit into actual work. Difficulty: moderate.

Aubrey Clayton — Bernoulli’s Fallacy

Book: Bernoulli’s Fallacy: Statistical Illogic and the Crisis of Modern Science.

Clayton provides a sharp historical and philosophical critique of common frequentist reasoning and of the confusion between probability in the data and probability in hypotheses. The book is polemical, which makes it enjoyable and occasionally unfair. Read it beside Spiegelhalter, Gelman, McElreath, and Cumming rather than as the book that finally defeated every statistician who annoyed you. Difficulty: moderate.

Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani, and Jonathan Taylor — An Introduction to Statistical Learning

Free book and courses: An Introduction to Statistical Learning.

This was one of my favorite entrances into statistical learning. It introduces regression, classification, resampling, regularization, trees, support-vector machines, clustering, and related methods with a practical orientation. It solves the problem of learning machine-learning algorithms as isolated tricks by placing them inside a coherent predictive workflow. Difficulty: moderate; versions are available with R and Python labs.

Max Kuhn and Colleagues — Predictive Modeling Done Properly

Books: Applied Predictive Modeling, the free Feature Engineering and Selection, and the free Tidy Modeling with R.

Kuhn’s work is excellent because it treats predictive modeling as a workflow rather than a beauty contest among algorithms. Preprocessing, resampling, tuning, leakage, class imbalance, performance estimation, feature engineering, and deployment all matter. In performance, a model with an impressive in-sample fit can still fail immediately with new athletes or future sessions. Tidy Modeling with R, written with Julia Silge, is the modern practical route into the tidymodels ecosystem. Difficulty: moderate to technical.

Galit Shmueli — Explanation versus Prediction

Paper: “To Explain or to Predict?”.

Shmueli’s distinction is essential: explanatory and predictive modeling are related but not identical tasks. A model can fit a plausible causal story while predicting poorly, or predict accurately without identifying an intervention. This helps prevent the common performance-science mistake of treating R-squared, feature importance, or prediction accuracy as proof of mechanism. Difficulty: moderate and short.

Causal Inference

Books: The Book of Why and the free Causal Inference: What If.

Causal inference asks what would happen under intervention, not merely what variables are associated. This is crucial in performance because a characteristic associated with elite status is not automatically a useful training target, and a variable that predicts injury is not necessarily a cause we can change. Causal diagrams are valuable because they force assumptions onto paper before statistical software hides them behind output. The Book of Why is accessible; Causal Inference: What If is technical.

20. bmbstats and Magnitude-Based Reasoning

Free book: bmbstats: Bootstrap Magnitude-Based Statistics for Sports Scientists.
Package: bmbstats documentation.
Source code: bmbstats on GitHub.

I wrote the bmbstats book to organize three classes of statistical tasks that are frequently confused: description, prediction, and causal inference. Description summarizes what happened in the observed sample. Prediction estimates unknown or future observations. Causal inference asks what might happen under an intervention. Statistical inference can appear in all three, but the claims are different.

The book also covers frequentist inference, Bayesian estimation, bootstrap methods, effect magnitudes, the smallest effect size of interest, measurement error, reliability, individual change, predictive model validation, and common sport-science designs. The main lesson is not a specific test. It is learning to identify what type of claim you are trying to make and whether the design, data, and model can support it.

A second lesson is that every statistical model is a Small World. Logical optimality inside the model does not guarantee good decisions in the Large World. A model should be judged by purpose, assumptions, uncertainty, practical consequences, and the cost of being wrong.

The book and package are free and open source. They are also examples of why coding, reproducible documents, Git, and public documentation matter: an idea can become something others can inspect, reproduce, criticize, and improve. Difficulty: moderate to technical.

21. Machine Learning and Interpretability

Brett Lantz — Machine Learning with R

Book: Machine Learning with R.
Code: Fourth-edition repository.

Lantz offers a hands-on introduction to common machine-learning methods through practical R examples. It is less philosophical than Molnar and less workflow-oriented than Kuhn, but it is approachable and useful for building confidence. The best way to use it is to reproduce examples and then replace the toy data with one of your own projects. Difficulty: beginner to moderate.

Christoph Molnar — Modeling Mindsets

Book and information: Modeling Mindsets.

This short book explains different cultures and purposes behind statistical modeling and machine learning. It solves the problem of arguing over methods before clarifying what the model is meant to accomplish. Are we predicting, estimating, explaining, discovering, or supporting a decision? This fits the pluralist perspective because different modeling mindsets can be useful for different jobs. Difficulty: moderate and short.

Christoph Molnar — Interpretable Machine Learning

Free online book: Interpretable Machine Learning.
Print search: Amazon.

Complex machine-learning models can predict well while giving little understanding of why a prediction was made. Molnar explains interpretable models, feature importance, partial-dependence methods, accumulated local effects, LIME, Shapley values, and the limitations of each method. In performance, a model that labels an athlete “high risk” without a credible explanation is difficult to trust or act upon. Interpretability does not magically convert prediction into causality, but it helps inspect model behavior and identify strange dependencies. Difficulty: technical, although the conceptual chapters are accessible.

Timo Freiesleben and Christoph Molnar — Supervised Machine Learning for Science

Free online book: Supervised Machine Learning for Science.

This book asks how machine learning can contribute to scientific understanding rather than merely prediction. It links flexible prediction, study design, uncertainty, explanation, and scientific inference. It is particularly relevant to sport scientists who want to use powerful models without pretending that every feature-importance plot reveals a causal mechanism. Difficulty: technical.

A sensible sequence is:

1. Modeling Mindsets

2. selected chapters of Interpretable Machine Learning

3. Supervised Machine Learning for Science

22. Communication, Teaching, and Presenting Ideas

Nancy Duarte — Resonate and slide:ology

Books: Resonate and slide:ology.

Duarte’s work helped me see presentations as structured arguments and journeys rather than decorated documents. The audience needs contrast, movement, orientation, and a reason to care. Slides should support transformation, not function as a transcript of everything the speaker knows. This is directly relevant to coach education and scientific communication because correct information that cannot be understood or remembered is practically weak. Difficulty: easy.

Patrick Winston — “How to Speak”

Lecture: Patrick Winston — How to Speak.

This lecture is one of the most practically useful pieces of material on explaining difficult ideas. Winston discusses promises, structure, repetition, stories, examples, and how to give an audience landmarks. It changes presentations from “show everything I know” to “help the audience build a useful mental model.” Difficulty: beginner.

Patrick Lencioni — The Five Dysfunctions of a Team

Book: The Five Dysfunctions of a Team.
Author resources: The Table Group.

I really loved this book. Lencioni uses a business fable to organize team failure around trust, conflict, commitment, accountability, and collective results. It is not a scientific theory of teams, but it is a very useful conversation model. Performance departments often possess enormous expertise and still fail because nobody trusts enough to disagree honestly, decisions remain ambiguous, and accountability arrives only after the season. Difficulty: easy and memorable.

Erin Meyer — The Culture Map

Book: The Culture Map: Breaking Through the Invisible Boundaries of Global Business

This book is outstanding for coaches working across different countries and sporting cultures. Meyer shows how expectations around authority, feedback, trust, disagreement, and decision-making differ across cultures. A coach who asks players for input may be viewed in Sweden as democratic, secure, and competent, while some Serbian players may interpret the same behavior as evidence that the coach has no clear answer or plan. The behavior is identical, but its meaning changes with the cultural context.

This is highly relevant to Problemming and the constraints-led approach. Athletes accustomed to direct instruction, technical drills, repetition, and frequent correction may initially experience exploratory practice as unstructured or insufficiently serious. Culture and previous coaching experience are therefore constraints within the learning environment. The task is not simply to apply the constraints-led approach, but to build enough structure, trust, and shared understanding for athletes to engage with it. Culture is not background noise; it shapes what athletes perceive as credible coaching.

Chris Voss and Tahl Raz — Never Split the Difference

Book: Never Split the Difference.
Author resources: Black Swan Group.

This book makes negotiation feel less like clever argument and more like calibrated listening. Tactical empathy, labeling, mirrors, calibrated questions, and attention to hidden constraints are useful when speaking with athletes, head coaches, managers, collaborators, and clients. The methods can become gimmicky if applied mechanically, so the deeper lesson is curiosity under pressure: do not rush to solve the problem before you understand the other person’s world. Difficulty: easy and immediately useful.

Donald Miller — Building a StoryBrand

Book: Building a StoryBrand.

This belongs beside Joseph Campbell, although it is a commercial simplification rather than comparative mythology. Campbell describes recurring structures of departure, ordeal, transformation, and return; Miller turns a similar story logic into a practical communication framework. Its main value is forcing the business or creator to stop making itself the hero. The athlete, coach, customer, or reader is the hero; the product or organization is the guide. Use the framework for clarity, not as a seven-box religious ceremony. Difficulty: easy.

Rory Sutherland — Alchemy

Book: Alchemy: The Dark Art and Curious Science of Creating Magic in Brands, Business, and Life.

This is a must-read. Sutherland shows why behavior and perceived value cannot always be understood through narrow rational or economic models. Small changes in meaning, framing, context, and perception can produce effects that appear disproportionate from a mechanical perspective. This connects StoryBrand, behavioral economics, evolutionary psychology, and the Forum for Action: people respond not only to what a thing is, but to what it means. It is useful for communication, product design, coaching, motivation, and understanding why technically equivalent options are rarely psychologically equivalent. Difficulty: easy, funny, and repeatedly useful.

Crucial Conversations

Book: Crucial Conversations: Tools for Talking When Stakes Are High.

This is a communication book rather than a philosophy book, but it belongs here because performance work repeatedly produces conversations with high stakes, disagreement, emotion, hierarchy, and incomplete information. The book helps people notice when dialogue becomes unsafe, separate observable facts from the story built around them, create shared purpose, and discuss difficult issues without choosing between avoidance and aggression.

The connection to performance is immediate: selection, playing time, role clarity, injury, return to play, poor adherence, staff conflict, and honest feedback cannot be solved by a dashboard. Technical competence without conversational competence often produces technically correct systems that nobody trusts. Difficulty: easy and practical.

23. A Minimum Professional Foundation for a Coach or Sport Scientist

The list above is intentionally broad. A beginner still needs a technical foundation. If someone told me today that they wanted to become a serious coach or sport scientist, I would not tell them to read only philosophy, complexity, or productivity. I would combine foundational domain knowledge with statistics, coding, reproducibility, and contrasting theories of learning.

Roger M. Enoka — Neuromechanics of Human Movement

Book: Neuromechanics of Human Movement, sixth edition.

Enoka integrates mechanics, muscle physiology, neurophysiology, and motor control to explain how the nervous system controls muscle actions and produces movement. This is one of the best foundations for understanding force, muscle activation, fatigue, sensory feedback, and movement without separating biology from mechanics. It is demanding, but it gives a coach a durable vocabulary that survives changing trends. Difficulty: technical; work through it slowly.

George A. Brooks, Thomas D. Fahey, and Kenneth M. Baldwin — Exercise Physiology: Human Bioenergetics and Its Applications

Book: Human Kinetics edition page.

Use the latest available edition. This is a serious physiology text with a strong emphasis on bioenergetics, metabolism, energy storage, oxygen transport, lactate, and the responses and adaptations to exercise. It helps prevent conditioning discussions from becoming a collection of zones, intervals, and slogans detached from physiology. Difficulty: technical and better used as a long-term reference than a quick read.

Richard A. Schmidt and Timothy D. Lee — Motor Control and Learning: A Behavioral Emphasis

Book: Motor Control and Learning, sixth edition.

Schmidt and Lee represent a more traditional information-processing and behavioral approach to motor control and learning. I still consider this a foundational text even though I now place it beside ecological dynamics rather than above it. It provides concepts related to motor programs, feedback, practice schedules, transfer, memory, and movement behavior. Reading it alongside How We Learn to Move or Dynamics of Skill Acquisition creates a productive opposition rather than an artificial consensus. Difficulty: technical.

Read Opposing Skill-Acquisition Traditions Together

A coach should be able to understand both the information-processing tradition and ecological dynamics. Schmidt and Lee help explain one tradition; Gray, Gibson, Davids, Renshaw, Chow, and colleagues explain another. The aim is not to perform a social-media conversion from one camp to another. Compare their units of analysis, concepts of information, assumptions about representation, roles of variability, and implications for practice design.

The friction between models is educational. It forces you to explain what each model sees, what it ignores, and which practical problem it helps solve.

Statistics and Measurement

Start with the free bmbstats book to distinguish description, prediction, causal inference, measurement error, and practical magnitude. Then move into Regression and Other Stories and Statistical Rethinking when ready.

Do not begin by collecting statistical tests. Learn to define the phenomenon, understand the measurement process, visualize the data, specify the claim, inspect uncertainty, and connect the analysis to a decision.

Programming and Data Literacy

Learn enough R or Python to import, clean, visualize, simulate, and report data. I would probably begin with Python today for general programming and AI, or R when statistics and reporting are the immediate focus. The language matters less than completing real projects and learning how data are structured.

A sport scientist who cannot inspect and manipulate data is permanently dependent on vendors, templates, and other people’s assumptions.

Git, Quarto, and Reproducibility

Use Git from the first project. Use Quarto to keep prose, code, tables, figures, and references together. A result that cannot be regenerated after the data changes is fragile. Reproducibility is not only an academic requirement; it is quality control for practical work.

Zotero and Scientific Reading

Build a Zotero library from the beginning. Do not store the scientific literature as an accidental pile of downloads. Record what each source claims, what design produced the claim, what limitations matter, and why the paper entered your project.

Psychology Without One Grand Theory

Sport psychology is difficult because useful practical metaphors, clinical models, personality theories, social psychology, behavioral economics, and experimental psychology operate at different levels and with different standards of evidence. I would begin with Haidt, Sapolsky, The Chimp Paradox, and Peterson’s older university lectures, but treat each as a model rather than literal final truth.

The most important practical capacities are listening, creating trust, giving feedback, managing difficult conversations, understanding incentives, and knowing when a qualified psychologist is needed.

Philosophy Is Part of Professional Training

Pirsig, Page, Mitchell, Taleb, Gigerenzer, Campbell, McGilchrist, and the theory-building literature are not decorations added after the “real science.” They help us ask what a construct is, what a model represents, what counts as evidence, what kind of claim we are making, and how we should act when evidence is incomplete.

A coach who never examines these questions can become highly skilled at using tools while remaining unclear about the problem those tools are supposed to solve.

A Minimal Modern Research Toolkit

For the sport-science route, I would treat the following as basic professional infrastructure rather than optional nerd decoration:

• Zotero for references and PDFs.

• R or Python for data manipulation, simulation, and analysis.

• Git and GitHub for version control.

• Quarto for reproducible reports, articles, presentations, and books.

• A plain-text note system that supports writing rather than replacing it.

• Enough command-line literacy to understand where files, environments, and dependencies live.

• Basic causal diagrams before running complicated causal models.

• Clear separation among description, prediction, explanation, and intervention.

The tools will change. The habits should survive the tools.

A Practical First-Year Sequence

A realistic first-year sequence might be:

1. Read Enoka and Brooks selectively alongside daily coaching.

2. Read Schmidt and Lee, then contrast it with Gray and the constraints-led literature.

3. Learn basic R or Python through one real data project.

4. Put the project under Git version control.

5. Write the report in Quarto.

6. manage the sources in Zotero.

7. use bmbstats and an introductory regression text to analyze the project.

8. take Learning How to Learn, Model Thinking, and an introductory Complexity Explorer course.

9. read Algorithms to Live By, Risk Savvy, and A Factory of One.

10. keep a short retrospective of what changed in your coaching and what remained only an interesting idea.

The goal is not to finish a curriculum. The goal is to build the capacity to learn, test, communicate, and revise for the rest of your career.

24. Where I Would Start Today

Trying to read everything before doing anything is another form of procrastination. I would combine reading with one small practical project.

Step 1: Learn how to learn

Complete Learning How to Learn and read Make It Stick. These improve how you approach every later book, course, and technical skill.

Step 2: Learn to think with multiple models

Take Scott Page’s Model Thinking course. Do not memorize every model. For each useful model, record what problem it addresses, what it assumes, what it emphasizes, what it ignores, and what could happen if it is applied incorrectly.

Step 3: Improve decisions under uncertainty

Read Algorithms to Live By, followed by Risk Savvy or Fooled by Randomness. These provide a useful entry into heuristics, explore–exploit, and fragile prediction.

Step 4: Build a visible personal workflow

Read A Factory of One, create a simple Kanban board, and limit active work. Do not build a perfect productivity cathedral. Make current commitments visible and reduce switching.

Step 5: Learn basic coding

Choose Python or R and complete one real project. I would probably start with Python today for broad applicability. I would choose R if my immediate focus were statistics, visualization, research, and reporting.

Step 6: Use Git immediately

Place the first project under version control. Do not wait until you consider yourself a programmer.

Step 7: Reproduce the project in Quarto

Write a short explanation, include the code, generate one table and one graph, and render the report. At this point, you have already built the foundation of a professional reproducible workflow.

Step 8: Build a Zotero library

Save the sources used in the project to Zotero, correct the metadata, attach notes, and connect the library to the Quarto bibliography. Reference management should be part of the research workflow rather than a panic performed before submission.

Step 9: Learn statistical thinking through real questions

Use the free bmbstats book as an applied introduction to description, prediction, causal inference, uncertainty, and measurement error. Move to Statistical Rethinking when ready for a deeper technical commitment.

Step 10: Learn theory construction

Read Fried’s short paper on theories and models, then Borsboom and colleagues’ Theory Construction Methodology. Take one concept from your work and try to separate observations, robust phenomena, target system, verbal theory, formal model, and predictions.

Step 11: Study complexity and networks

Complete Melanie Mitchell’s introductory Complexity Explorer course, read the opening chapters of Barabási’s free Network Science, and then explore Feldman or Epskamp depending on whether you are more interested in dynamics or psychometrics.

Step 12: Study skill acquisition

Begin with How We Learn to Move and The Constraints-Led Approach. Use them to redesign one real practice problem rather than collecting another hundred games.

Step 13: Read the deeper philosophy gradually

Read Pirsig, Taleb, Mitchell, and the dispositionalism material slowly. These books are not checklists. Their value appears when they change the questions you ask in practical situations.

Optional Short Route: Twelve Resources Before the Rabbit Hole

For someone who wants a smaller route before this list turns into a second career, I would start with:

1. Algorithms to Live By

2. Risk Savvy

3. the free Model Thinking course

4. Super Thinking

5. Zen and the Art of Motorcycle Maintenance

6. Alchemy

7. A Factory of One

8. Make It Stick

9. How We Learn to Move

10. The Five Dysfunctions of a Team

11. The Art of Statistics

12. one real coding project written in Quarto and stored with Git

That sequence gives you decisions, uncertainty, models, quality, communication, work design, learning, teams, statistics, and technical agency. It is enough to become dangerous in a useful way.

25. A Compact Starting Shelf

For someone who wants a smaller starting list, I would choose:

1. Algorithms to Live By

2. Risk Savvy

3. Fooled by Randomness

4. the free Model Thinking course

5. Super Thinking

6. Seeking Wisdom

7. Zen and the Art of Motorcycle Maintenance

8. The Blind Spot

9. Alchemy

10. A Factory of One

11. Deep Work

12. Four Thousand Weeks

13. Make It Stick

14. How We Learn to Move

15. The Five Dysfunctions of a Team

16. Never Split the Difference

17. Statistical Rethinking

18. the free Rethinking Causality, Complexity and Evidence for the Unique Patient

19. the free Network Science

20. the free bmbstats book

This sequence moves through decision-making, uncertainty, pluralism, quality, productivity, learning, skill acquisition, causality, networks, and statistical reasoning.

Final Thoughts

Many of the resources that most influenced my coaching contain almost nothing about sets, repetitions, sprinting, conditioning, or periodization.

They deal with uncertainty, complexity, causality, quality, judgment, models, learning, meaning, psychology, networks, writing, coding, measurement, workflow, and the limits of prediction.

These subjects are not distractions from performance. They determine how we define performance, what we decide to measure, how we interpret observations, how we design practice, how we communicate, and how we act when the evidence is incomplete.

Taleb warns that our predictions are fragile. Gigerenzer gives us adaptive heuristics. Christian and Griffiths show the hidden structure of decisions. Page gives us multiple models. Mitchell explains why complex phenomena may require plural explanations. Pirsig reminds us that analysis does not exhaust Quality. Dispositionalism gives us causal powers, propensities, and context-sensitive manifestation. Network science shifts attention from isolated parts to patterns of interaction. Borsboom, Fried, Haslbeck, and Smaldino help separate observations, phenomena, theories, and models. McElreath teaches us to construct and criticize Small Worlds. Ecological psychology changes the unit of analysis from an isolated athlete to an athlete–environment system. Coding allows us to make models executable. Scrum, Kanban, Lean, and DevOps remind us to make work visible, shorten feedback loops, limit unfinished work, and improve the delivery system rather than merely demanding more effort. Zotero keeps the evidence traceable, while Quarto and Git make the work reproducible, transparent, and revisable.

I do not believe coaches need one more ideology. That includes evidence-based scientism, mystical pseudoscience, red-pill total explanations, feminist total explanations, machine-learning solutionism, and whichever coaching cult currently owns the loudest podcast microphone.

They need a richer but usable toolbox.

They need enough physiology to avoid nonsense, enough statistics to understand uncertainty, enough psychology to work with humans, enough complexity science to respect emergence, enough skill-acquisition theory to design meaningful practice, enough technology to test ideas, and enough philosophy to remember that none of these perspectives owns the whole athlete.

The goal is not to become an expert in everything.

The goal is to become capable of changing lenses without becoming blind to the limitations of each one.

Use multiple models.

Do not worship any of them.

Tools, not truths.

Every coach who has planned running-based high-intensity interval training (HIIT) for a group knows the problem. The session is easy to describe in theory: pick a speed, pick a duration, set the recovery, and repeat. The problem starts when the theory has to become cones, distances, groups, timing, substitutions, and 24 athletes who do not have the same profile.

One athlete needs 86 meters. Another needs 92. Another is coming back from injury. The field is shorter than expected. The head coach wants the block finished in 12 minutes. You have a spreadsheet with test data, a calculator, three interval ideas, and one assistant asking where Group 3 should start. This is where conditioning stops being a physiological concept and becomes a coaching problem.

HIIT Builder 3 was created for that middle space. It is a macro-enabled Excel template that helps coaches turn user-entered athlete profiles into running-based HIIT sessions: individual targets, practical groups, printable outputs, saved programs, completed-session records, and enough structure to keep the session from becoming a mess. It is not an automated coach. It does the repetitive calculations and organizes the output. The coach still makes the decision.

This Version 3 upgrade was built with Đorđe Sekulić, who did outstanding hard work taking the previous HIT Builder idea and turning it into a much more complete tool. The original version already had a useful practical foundation. Djordje helped make it cleaner, deeper, and more useful for real coaching workflows.

Let Excel do the boring math. Let the coach coach.

The real problem is not the formula

Coaches often think the problem is calculating the distance. That is part of it, but it is not the whole problem. If an athlete runs at 5 m/s for 20 seconds, the straight-line distance is 100 meters. Fine. Nobody needs a product for that.

The real problem is that the distance changes when you use different anchors, different athletes, different surfaces, standing starts, shuttles, change-of-direction losses, different work durations, different methods, and different group tolerances. Then you have to deliver this to 15, 20, or 30 athletes without turning the warm-up into a staff meeting.

The hidden cost is friction. You calculate, copy, adjust, group, rewrite, print, explain, change it again, and then after the session you probably do not save the exact logic. Next week you rebuild the same session from memory. Coaches do this all the time, not because they are stupid, but because the workflow is fragmented.

HIIT Builder 3 tries to connect the pieces: athlete data, velocity profiles, method selection, intensity, work and rest, start loss, COD loss, grouping, printable outputs, Running Database, and Program Manager. The benefit is not fake precision. The benefit is a cleaner path from intention to field delivery.

A quick physiology map before we put cones down

HIIT is not one thing. A long interval around the aerobic ceiling, a 30:30 session, a shuttle-based intermittent session, and a short sprint interval can all be called HIIT, but they do not create the same problem for the athlete.

A useful coaching map is the four-domain model. The moderate domain is mostly stable when the athlete is fresh. This is your easier aerobic work and volume accumulation. The heavy domain is still controllable, but cost starts drifting. This is where tempo, controlled threshold-style work, and a lot of “not hard, but not free” training lives.

The severe domain is where no true steady state exists. Work above the critical velocity or critical power region is finite. The athlete can hold it for a while, but the system is moving toward the limit. A lot of VO2max-oriented intervals live here. The extreme domain is above the region where oxygen uptake can neatly explain the whole session. Short hard reps, sprint intervals, and repeat-sprint work live here. Heart rate can be late and misleading because the rep is often over before HR catches up.

This map is useful, but it does not tell you what to do automatically. “Severe domain” is not a training session. It is a region of likely response. The method still matters, the athlete still matters, and the week still matters. This is the Agile Periodization problem in one sentence: the map gives you the is; the coach still has to decide the ought.

For HIIT Builder 3, this matters because the workbook lets you work with different speed anchors. Maximum Aerobic Speed (MAS) is a practical anchor for the aerobic ceiling. Maximum Sprinting Speed (MSS) gives you the sprint ceiling. Anaerobic Speed Reserve (ASR), the difference between MSS and MAS, helps when you move above MAS and the athlete’s sprint reserve changes the meaning of the prescription. Intermittent Field Test velocity (vIFT; referring to Martin Buchheit’s 30-15 test) is useful because intermittent field tests include acceleration, deceleration, recovery, and change of direction, which often makes them more relevant to team-sport conditioning than a clean straight-line time trial.

Two athletes can have the same MAS and very different MSS. If you prescribe short fast reps only from MAS, those athletes might receive the same target while experiencing different neuromuscular and metabolic demands. ASR does not solve the universe, but it gives you a better language for some sessions. For a broader explanation of this map, read my Endurance Map Builder article on Substack. For the full background on HIIT prescription, methods, and practical planning, see the HIIT Manual on Amazon.

Corrections are not decoration

The next practical problem is that straight-line math does not automatically survive the field. If the athlete has to start from standing, accelerate, brake, change direction, or run on grass or mud, the target distance needs a correction. Otherwise the spreadsheet says one thing and the athlete experiences another.

In the original HIIT Manual, I used practical start-loss and change-of-direction corrections. The classic quick-and-dirty COD rule is around 0.7 seconds per 180-degree turn for a typical shuttle context. For lighter or average athletes on decent surfaces, this can be a useful default. For heavier athletes, poor surfaces, higher speeds, tighter shuttles, or worse braking mechanics, the cost can be higher.

The basic idea is simple enough:

available running time = work time - start loss - COD losses
corrected distance = target velocity x available running time

Do not turn this into religion. The 0.7-second rule is a heuristic. It was useful because it was simple enough to coach with, not because it captured every athlete, speed, shuttle length, and surface. A rule of thumb is allowed to be imperfect. It just needs to be visible enough that the coach knows what assumption is being made.

Martin Buchheit recently published a useful update on this exact issue: From Rule of Thumb to Mechanistic Formula and the accompanying PDF report. The short version is that the old 0.7-second rule and the 2–3% per COD approach are practical, but they can break down when shuttle length changes, especially in shorter shuttles where turns become a larger part of the task. The newer model makes the COD cost more speed-dependent and individual-dependent.

This is exactly why I prefer tools that make assumptions visible. HIIT Builder 3 does not pretend the correction is perfect. It gives you fields for start loss, COD loss, shuttles, and prescription type so you can adjust the setup and see the consequence. If you disagree with the correction, change it. That is coaching.

The map is not the territory

All of this sits inside a bigger point. MAS, MSS, ASR, IFT, COD corrections, grouping tolerances, and intensity domains are maps. They are not the territory. They help us see something useful, but they do not remove context, judgment, or trade-offs.

That is not a weakness. That is the honest use of a model. The weak move is pretending that the model is reality, hiding the assumptions, and then acting surprised when the field session disagrees. The stronger move is to use clear heuristics, make the assumptions visible, deliver the session, and update from what happened.

If you want the deeper Agile Periodization rabbit hole behind this way of thinking, I cover it in the Agile Periodization: Philosophical Foundations course. I also write about these ideas on the Agile Periodization Substack. The short version is simple: tools, not truths; maps, not territory; decisions under uncertainty, not fake certainty.

What HIIT Builder 3 actually does

The workflow is simple: enter athlete data, build the session, group and deliver, review, and save. That sounds almost too obvious, but most coaches do not have those steps connected in one place.

In the athlete sheet, you enter the speed anchors you have: MSS, MAS, IFT, and related profile information. The workbook can then build velocity profiles and use those anchors in the Builder sheet. This is important because the quality of the output depends on the quality of the inputs. A bad test does not become a good prescription because Excel wore a tie.

In the Builder, you select the method and variation, choose the prescription logic, set intensity and duration, add recovery details, apply start and COD corrections, and decide whether you want ranges or specific outputs. The workbook then calculates velocities and distances and prepares the session logic.

The grouping step is where the tool becomes useful for teams. Perfect individualization is easy when there is one athlete. It becomes a comedy routine when there are 24 athletes and one field. HIIT Builder 3 lets you choose the acceptable difference between individual target distance and group distance. Tight tolerance means better individualization but more groups. Loose tolerance means easier delivery but rougher individualization.

That trade-off is not a bug. That trade-off is the job. The tool makes it visible so the coach can decide.

After the session, the workbook helps you save what happened. This matters because the planned session and the completed session are often different. Some athletes complete fewer reps. Some return-to-play athletes need reduced volume. Some groups overshoot or undershoot. The surface is heavier, the weather is ugly, or the previous sport session left more fatigue than expected. The session teaches you something only if you keep the trail.

Version 3 includes the Running Database and Program Manager so the session does not disappear. You can save the logic, tag it, reload it, adjust it, and build a better version next time. That is much more useful than a one-off calculator.

What to watch on the field

The workbook gives you a starting point. The field tells you whether the starting point was sensible. If athletes miss the target early, check the setup before blaming fitness. The distance might be too aggressive, the correction might be wrong, the group tolerance might be too wide, or the athletes might simply be carrying fatigue from the week.

If mechanics fall apart early, the session has changed. You are no longer training the thing you planned. Sometimes that is acceptable. Often it is just junk with a stopwatch. Short fast sessions need technical quality. Heart rate will not save you there because it lags behind the work.

If the session looks too hard, do not immediately moralize. Maybe the heat did the work of an extra interval. Maybe your MAS estimate is wrong. Maybe the athlete is under-recovered. Maybe the surface is slower than the one used during testing. This is why saving completed work and coach notes matters. Not because the notes are perfect, but because they help you update the next decision.

If the session is too easy, do not automatically add more. First ask what the session was supposed to do. Sometimes extensive work should feel controlled. Sometimes the athlete paced well. Sometimes you underdosed the problem. Same observation, different conclusion, depending on the purpose.

The important point is this: the workbook should not make the coach less important. It should make the coach’s attention available for the things that actually need judgment: movement quality, readiness, communication, adaptation, and the next decision.

Who this is for

HIIT Builder 3 is for coaches who plan running-based conditioning and want less friction between the idea and the delivered session. It is especially useful for strength and conditioning coaches, sport scientists, team-sport coaches, private coaches with groups, and combat-sport coaches who use running intervals as one piece of the conditioning menu.

It is most useful when you work with groups. One athlete can be managed with a calculator and a notebook. A group needs workflow. You need individual targets, but you also need practical lanes, cones, groups, printouts, and a way to save what worked.

It is not for coaches who want a cloud app, automatic GPS import, wearable integration, or an algorithm that tells them what problem to solve. This is a macro-enabled Microsoft Excel workbook. You need desktop Excel and macros enabled. Very glamorous. Like carrying cones.

What you get

HIIT Builder 3 includes the macro-enabled Excel workbook, athlete database, velocity profiles, HIIT drill logic, training guide tables, main Builder sheet, grouping workflow, Program Manager, Running Database, distance calculators, MAS and MSS tools, Curved Run Builder and history, print/PDF export functions, and the instructional walkthrough video.

The Curved Run Builder is experimental, and that is how it should be treated. Use it, test it, see whether the setup makes sense on the field, and send feedback. Team-sport running is not only straight-line running, but a curved run that looks clever in Excel still needs to survive athletes, cones, and actual execution.

The one-line workflow is this: enter athlete data, choose the method, build the session, group the athletes, deliver the work, record what happened, and save the logic for next time. Nothing mystical. Just less chaos.

Get HIIT Builder 3

If you plan running-based HIIT and want a cleaner way to move from athlete profiles to field-ready sessions, you can get HIIT Builder 3 here:

Get HIIT Builder 3 on Payhip

The product includes the Excel workbook and instructional walkthrough. You can watch the instructional video here:

Watch the HIIT Builder 3 instructional video

You can also see the original HIT Builder here:

Original HIT Builder

HIIT Builder 3 was upgraded in collaboration with Đorđe Sekulić, who did the outstanding hard work of turning a useful tool into a much more complete Version 3 workflow. If you use it, send feedback, bugs, friction points, and examples from your context. I am especially interested in the grouping workflow, printable outputs, Program Manager, and Curved Run module.

A good tool does not remove uncertainty. It organizes it well enough so the coach can act, observe, and update. That is what HIIT Builder 3 is trying to do.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

The athletes are arriving. One rack is occupied, the cable station is being repaired, and the athlete who missed last week has just told you that heavy bilateral work feels like a poor life choice today. The plan says:

Lower-body strength, upper-body pull, power, and trunk.

This is where classification either helps or reveals that it was designed by someone who never had to start a session on time.

Part I built the Place of Things map. It described exercises through movement patterns, anatomy, muscle action, equipment, position, direction, and context. That map matters. A good library needs a rich back end.

But the coach in Figure 1 does not need every possible description at once. The coach needs to narrow choices, preserve the job of the session, work around constraints, communicate clearly, and later review what happened.

That is the Forum for Action perspective.

The goal is not to create a smaller classification because coaches are unable to understand complexity. The goal is to compress complexity at the point of use. The detailed map remains available underneath. The visible interface keeps only the distinctions that help someone do something.

A useful practical classification should support four actions:

1. Filter a large library.

2. Plan sessions and longer structures at the right resolution.

3. Substitute intelligently when reality changes.

4. Review broad patterns without pretending the categories explain the whole effect.

Everything else is optional.

Start with the human, then narrow

Before someone is a powerlifter, fighter, footballer, runner, office worker, or person trying to survive parenthood with a functioning back, they are a human. They share a broad body plan and a recurring set of movement possibilities: pushing, pulling, squatting, hinging, lunging, jumping, landing, throwing, rotating, carrying, crawling, walking, running, and controlling the body in space.

This is the base shown in Figure 2. It connects the classification system to the Programming Pyramid: human-specific at the base, sport-specific above it, and athlete-specific at the top.

The layers are nested. They are not separate religions.

A practical shared exercise library should mainly describe the human-specific layer through stable tags. A squat remains a lower-body grinding pattern whether it is selected for a fighter, footballer, powerlifter, or older adult. The sport layer adds what the activity tends to demand, overload, punish, or leave underdeveloped. The athlete layer adds the actual person’s history, tolerance, readiness, preferences, morphology, beliefs, and current problems.

This protects us from two common mistakes.

The first is starting too specific. Coaches try to make every gym exercise look like the sport and end up with poor imitations of practice. The second is pretending that a broad classification solves the individual problem. “Lower push” does not tell us whether the person should back squat, split squat, belt squat, use a machine, hold an isometric, or do nothing heavy today.

The forum-for-action system starts broad because broad categories are stable. It narrows through tags and context because decisions are local.

The gross categories

A practical system needs a small set of broad families that are easy to remember and useful enough to organize a session. Figure 3 gives the front-end map I currently find most useful.

These categories are not mutually exclusive truths. They are broad action families. An exercise can sit near the border, carry several tags, or change practical category when the method changes. That flexibility is a feature.

Grinding

Grinding is the broad family of exercises in which the athlete produces and controls force against resistance without primarily releasing the implement or body into flight. The repetition may be slow or fast. “Grinding” does not mean ugly, maximal, or performed with a face that scares the interns. It means that force production against resistance is the organizing feature.

The familiar movement patterns live here:

• Upper push: bench press, push-up, overhead press, landmine press.

• Upper pull: pull-up, row, pulldown, chest-supported row.

• Squat or lower push: front squat, back squat, split squat, leg press.

• Hinge or lower pull: Romanian deadlift, deadlift, hip thrust, hyperextension.

• Lunge: forward, reverse, lateral, walking, supported, loaded.

• Carry and trunk: farmer carry, overhead carry, Pallof press, plank, rollout.

This language is broad enough for generalists and still useful for specialists. A powerlifter can add a deeper squat-bench-deadlift hierarchy. A bodybuilder can add muscle and line-of-pull tags. The shared library does not have to choose between them. Grinding is the common trunk; specialist systems can grow as branches.

The category also supports substitution. A lower grinding slot does not force a back squat. It preserves a job while leaving room for the athlete, room, phase, and method to determine the exact task.

Ballistic

Ballistic actions involve rapid acceleration, projection, release, flight, or an intention to continue producing force through the movement rather than braking the action early. The category includes several families that specialists may later separate:

• Jumping: squat jump, countermovement jump, box jump, broad jump, hop, bound.

• Throwing: chest pass, overhead throw, scoop toss, rotational throw, slam.

• Swinging and rotating: kettlebell swing, landmine rotation, rotational cable or band actions.

• Olympic-lifting derivatives: hang clean, power clean, high pull, snatch derivative, jerk derivative.

• Fast pushes and pulls: bench throw, explosive push-up, explosive pull-up, fast band or machine work.

• Sprinting and acceleration: starts, short sprints, resisted accelerations, bounds and related actions.

The common feature is not that every exercise shares identical mechanics. It is that the task is organized around rapid expression rather than sustained force against resistance.

This category is especially valuable because ballistic exposure is easy to lose in programs dominated by strength exercises, machines, and local tissue work. A quick review of the week can reveal whether the athlete has done any jumping, throwing, accelerating, or releasing at all.

The category also prevents the opposite mistake: calling every fast-looking exercise “power.” A sloppy high-speed movement with no useful intent is not rescued by the label ballistic. The classification identifies a family. Coaching still determines quality, dose, and relevance.

Control, patterning, coordination, complex, and direct work

This is the drawer that initially feels less elegant. That does not make it useless. Many exercises are not primarily selected because they are heavy grinds or ballistic projections.

Control exercises make position, balance, support, rhythm, or management of degrees of freedom central to the task. Single-leg RDLs, step-downs, unsupported split-stance tasks, controlled trunk work, and balance-oriented holds may live here. These exercises can still be loaded. The label identifies the dominant coaching problem.

Patterning exercises are selected to teach, rehearse, restore, or simplify an action. A goblet squat can be a grinding strength exercise for one person and a patterning exercise for another. A wall drill can be used to organize acceleration posture. A dowel hinge can help someone find the movement before load is added.

Coordination exercises require several body segments, timing relationships, or contralateral actions to work together. Crawling, certain carries, contralateral lifts, transitions, and some medicine-ball sequences may fit. The category should not become an excuse for random complexity. Coordination is a task property, not a compliment.

Complex exercises combine several phases, positions, or movement families. The Turkish get-up is the classic example. Some flows, ground-to-standing tasks, and combined lift-carry-transition exercises also belong here. Complex work can be useful, but complexity creates cost. The athlete must solve the task before substantial force or volume can be expressed. It seems that Instagram is full of these performative exercises, as William Wayland use to call them.

Mobility and direct tissue work deserve visible space because real programs include them. Hip mobility drills, ankle work, thoracic rotation, neck exercises, calf work, Copenhagen planks, hamstring slides, rotator-cuff work, grip exercises, and local isometrics often support robustness, rehabilitation, or availability. They are not leftovers simply because they do not fit neatly into push, pull, squat, and hinge.

I prefer keeping these as related but distinct tags under an “Other” or “Control / Complex / Direct” branch. The branch can evolve with use. A classification without an honest other drawer eventually starts lying.

Gait and cyclical work

Gait and cyclical actions can sit beside the main resistance-training families:

• carries

• sled pushes and pulls

• crawling and rolling

• walking and running

• cycling and rowing

• skiing and swimming

• climbing

Some of these overlap with grinding, ballistic, control, and conditioning. A heavy sled push can be a lower-body grind, gait task, and conditioning exercise. A farmer carry can be grip, trunk, gait, total-body strength, and a beautifully simple way to keep a large group moving.

Multiple membership is not a database failure. It is an honest description of a multi-purpose task.

Secondary tags: add detail only when it changes a choice

The gross category gets us into the right neighborhood. Secondary tags help us find the house.

The most useful shared tags usually come from five groups.

Location can remain very broad: upper body, lower body, torso, total body. These labels support quick filtering and session summaries without pretending the body works as separate machines.

Joints and regions can include neck, thoracic spine, lumbar spine, shoulder, elbow, wrist and forearm, hip, knee, ankle and foot. A coach does not need to mark every joint for every exercise. The useful question is whether the tag helps someone find or manage relevant exposure.

Muscular regions can include chest, upper back, deltoids, triceps, biceps, forearms, quadriceps, hamstrings, glutes, groin, calves, abdominals, low back, traps, and lats. These are particularly valuable for bodybuilders, rehabilitation, and direct support work.

Modifiers describe meaningful task differences: bilateral, unilateral, alternating, contralateral, ipsilateral, supported, unsupported, vertical, horizontal, forward, backward, lateral, diagonal, stable, unstable, isometric, eccentric, paused, explosive, rhythmical, or slow.

Equipment connects the library to reality: bodyweight, barbell, specialized bar, dumbbell, kettlebell, plate, landmine, medicine ball, sandbag, band, cable, rings, suspension trainer, machine, sled, and so forth.

A tag should earn its place by changing search, setup, substitution, communication, or review. If a label never changes a decision, it is probably metadata cosplay.

Use a hierarchy, but do not force one home

A hierarchy makes the language easier to navigate. Figure 4 shows a practical tree with broad branches and deeper children.

If Grinding is the parent, push, pull, squat, hinge, lunge, carry, and core can sit beneath it. Selecting the parent should reveal all exercises in the family. Selecting Grinding + Upper Pull narrows the search. Adding Supported + Dumbbell narrows it again.

If Ballistic is the parent, jump, throw, swing, rotate, Olympic lift, sprint, and fast push or pull can sit beneath it. The coach can search broadly or move to the exact action family.

The tree reduces repeated thinking, but it should not become a prison. An exercise can inherit parent tags and carry additional tags from other branches. A box jump is ballistic and jump, but also lower body, bilateral, bodyweight, and perhaps vertical. A med-ball rotational throw is total body, ballistic, throw, rotational, contralateral or ipsilateral depending on execution, and medicine ball.

The important design choice is multiple membership. The exercise is not forced to live in one folder. The hierarchy organizes the vocabulary. Tags describe the object from several useful directions.

This also allows different users to open different levels. A general user may search only by movement type and equipment. A clinician may open joint, muscle, tissue, support, and action tags. A specialist may add sport-specific branches. The same library supports several resolutions through progressive disclosure.

One exercise, several tags

The exercise cards in Figure 5 show how the system becomes useful.

A front squat can be tagged lower body, grinding, lower push, squat, bilateral, barbell, quadriceps, hip, knee, and ankle. A box jump can be lower body, ballistic, jump, explosive, bilateral, and bodyweight. A single-leg RDL can be lower body, control, hinge or lower pull, unilateral, unsupported, hamstring, glute, and dumbbell. A chest-supported row can be upper body, grinding, pull, horizontal, supported, and dumbbell or machine.

Not every tag needs to be visible on the card. The library can store a richer set while showing the fields relevant to the current view.

Suppose the program contains a lower-body ballistic slot. The filter returns jumps, hops, bounds, Olympic-lifting derivatives, and perhaps ballistic sled or swing options depending on the library. Add bodyweight because equipment is limited. The list narrows.

Suppose the athlete needs an upper-body grinding pull with reduced postural cost. Add supported. The library may return chest-supported dumbbell rows, machine rows, cable rows with support, or prone rows.

Suppose the coach needs a lower-body control task that is unilateral and uses a dumbbell. The library returns single-leg RDLs, step-down variations, supported or unsupported split-stance options, and other tasks tagged that way.

The classification is not selecting the exercise. It is reducing decision friction. The coach still decides which option fits the person and objective.

Context is added by the user

Some tags should not be universal library properties.

Priority is contextual. Main, supplemental, assistance, primary, secondary, tertiary, major, and minor describe where an exercise sits in a particular plan. The front squat may be primary for one block and assistance in another.

Specificity is contextual. Competitive, specific developmental, specific preparatory, general preparatory, and rehabilitation labels depend on the sport and athlete. The bench press is a competition exercise for one person and general upper-body work for another.

Tolerance and preference are contextual. An exercise may be available, familiar, provocative, confidence-building, disliked, or temporarily excluded for one athlete.

Facility status is contextual. A cable row exists in the library even when the cable station is unavailable today.

The shared exercise record should contain relatively stable descriptors. The program, athlete profile, phase, and facility add local meaning.

This separation prevents one coach’s worldview from contaminating the whole library. The system can support specialists without pretending their priorities are universal.

Exercise plus method remains the prescription unit

A tag system describes the exercise. It does not complete the prescription.

A split squat tagged lower body, grinding, lunge, unilateral, and dumbbell can still be prescribed as heavy strength work, slow eccentric tissue loading, an isometric hold, a low-cost reintroduction task, or a rhythmical circuit exercise. The method changes the intervention.

A practical planning tool therefore needs two connected objects:

• Exercise: what task is being performed?

• Method or scheme: how is it loaded, dosed, progressed, and controlled?

This matters for analysis as well. Counting the exercise category without the scheme can be misleading. Three sets of heavy front squats, three sets of low-load tempo squats, and three sets of jump squats should not be treated as identical because the word squat appears in all three.

The plan remains a hypothesis. Tags help describe the hypothesis. The athlete’s response still has to be observed.

Filter, plan, and substitute

The workflow in Figure 6 is where the classification earns its keep.

Filter

Filtering reduces the library to a manageable menu. The coach can begin from the slot rather than from memory.

• lower body + ballistic + bodyweight

• upper body + grinding + pull + supported

• torso + isometric + band

• lower body + direct work + groin + bodyweight

• total body + gait + carry + dumbbell

This is useful for experienced coaches because it reduces search time. It is useful for developing coaches because it makes the logic of exercise selection visible.

Plan

Categories can be used as program slots before exact exercises are chosen. A session might be planned as:

1. control or mobility

2. ballistic

3. lower-body grind

4. upper-body pull

5. direct tissue or accessory work

The slot preserves the philosophy. The exercise is selected later according to the athlete, facility, phase, and day.

This is the meeting point between top-down and bottom-up planning. Top-down says what kind of work the session needs. Bottom-up asks what can actually be done here, with these people and resources.

The gym matters. A program is not executed in empty space. Equipment quantity, anchors, walking paths, group size, station capacity, and coaching visibility all shape the available choices. The same logic appears in The Gym Is Not Just a Room: facility and program design create each other.

Substitute

Substitution is not random variety. It is controlled adaptation.

If the cable row is unavailable, the coach can search for exercises that retain upper body + grinding + pull and then select ring row, chest-supported dumbbell row, band row, machine row, or another option that fits the group.

If the athlete needs reduced axial loading, the coach can preserve lower body + grinding + lower push while changing equipment, support, and setup. Belt squat, split squat, leg press, hack squat, or another task may fit.

The tags protect the job while allowing the form to change.

This is important in team settings. A substitution must work not only biomechanically, but logistically. A theoretically perfect alternative that creates a fifteen-minute queue is not an alternative. It is a scheduling problem wearing a biomechanics hat.

Review without worshipping the counts

Classification can help review what was planned and what was completed. Figure 7 shows the right level of ambition.

A coach can ask:

• Did the week contain any ballistic exposure?

• Was upper-body work dominated by pushing?

• Did lower-body training rely almost entirely on bilateral tasks?

• Were direct calf, groin, hamstring, shoulder, neck, or grip needs addressed?

• Which equipment created repeated queues?

• Which slots were substituted most often?

• Which exercise categories were prescribed but regularly skipped?

These questions can expose blind spots. They do not produce verdicts.

Ten sets of pulling do not tell us whether the athlete performed them well. A week containing three ballistic exposures may still be poorly timed. A balanced push-pull ratio can coexist with excessive fatigue, poor technique, or irrelevant exercise selection.

The categories are a map for reflection. The review also needs athlete feedback, observed quality, completion, readiness, symptoms, session flow, and the sport schedule.

The plan is a hypothesis. The completed week provides evidence. The next plan is an update.

A minimum viable classification

A usable starting system can be surprisingly small.

Gross movement type

• Grinding

• Ballistic

• Control / Patterning / Coordination / Complex / Direct Work

• Gait / Cyclical

Location

• Upper Body

• Lower Body

• Torso

• Total Body

Movement family

• Push

• Pull

• Squat / Lower Push

• Hinge / Lower Pull

• Lunge

• Core / Carry

• Jump

• Throw

• Swing / Rotate

• Olympic Lift

• Sprint

Modifiers

• Bilateral / Unilateral

• Supported / Unsupported

• Isometric / Eccentric / Paused / Explosive

• Vertical / Horizontal / Lateral / Rotational

Equipment

• Bodyweight

• Barbell

• Dumbbell

• Kettlebell

• Band / Cable

• Machine

• Medicine Ball

• Sled

• Other

Then add joint, muscle, tissue, and deeper positional tags only when actual use demands them.

This is not minimal because detail is bad. It is minimal because a system must survive contact with users. A coach should be able to add an exercise without completing a census. The classification can grow from real searches, substitutions, and review questions.

Rules for keeping the Forum for Action useful

The first rule is to start broad and narrow only as needed. A gross category gets you near the answer. Secondary tags reduce the menu. Context makes the final choice.

The second rule is to allow several tags per exercise. Exclusive boxes produce clean diagrams and poor libraries.

The third rule is to separate stable properties from local meaning. Movement type and equipment can live in the library. Priority and specificity usually belong to the program and user.

The fourth rule is to keep an honest Other branch. Control, patterning, coordination, complex work, mobility, rehabilitation, and direct tissue tasks will not always separate neatly. Let the branch evolve.

The fifth rule is to hide complexity until it is useful. The database can be rich while the interface remains simple. Different users can open different layers.

The sixth rule is to let practice edit the system. Add tags when coaches repeatedly need them. Hide tags that nobody uses. Classification is a working hypothesis about useful distinctions, not a constitution.

The seventh rule is to remember the exercise plus method. Selecting the task does not complete the intervention.

The eighth rule is to never confuse tag counts with training effects. The classification describes what was planned and performed. It does not explain the entire response.

The classification should disappear into the coaching

A good tool eventually becomes quiet.

The coach should not stand on the floor reciting an ontology. The classification should help the coach find an option, preserve a slot, change the exercise when necessary, organize the group, and notice broad patterns later.

The detailed Place of Things map remains underneath. It gives us vocabulary and prevents one specialist worldview from taking over the entire library. The Forum for Action compresses that map into a practical interface.

The sequence is simple:

Start with the human. Choose the broad action family. Narrow with tags. Add sport and athlete context. Prescribe the method. Observe what happens. Update.

The classification is not the program. It is not the training effect. It is not the athlete.

It is a way of turning a cabinet of possible exercises into the next useful action.

That is enough. The session is starting.

Addendum: tags find, ratings choose

Tags answer questions such as:

• Is this lower body or upper body?

• Is it grinding, ballistic, control, gait, or direct work?

• Does it fill a lower-push, lower-pull, upper-push, or upper-pull slot?

• What equipment does it require?

• Which tissues, regions, and modifiers are relevant?

Those tags create the candidate pool.

Ratings answer a different question:

Among the exercises that could do the job, which one fits this athlete, group, facility, and day?

A Romanian deadlift, hip thrust, and hamstring curl can all appear after filtering for lower body + posterior chain + lower pull. The tags tell us why they belong in the conversation. They do not tell us which one should be selected.

A small set of operational ratings can make the trade-offs visible. The six proposed dimensions are summarized in Figure 9.

The distinction is simple:

Tags describe membership. Ratings describe practical characteristics. Context determines whether those characteristics are desirable.

For a team of twelve with ten minutes and two racks, the Romanian deadlift may beat the hip thrust because it is faster to organize and easier to scale. For one athlete with plenty of time and a specific glute-loading objective, the hip thrust may be the better fit. The rating does not crown a winner. It exposes the compromise.

Keep the ratings coarse

A three-point scale is usually enough:

• 1 — Low

• 2 — Moderate

• 3 — High

More precision looks scientific but often records confidence the coach does not possess. Is setup burden really a 6.8 rather than a 6.2? Probably not. The useful distinction is whether the exercise is easy, manageable, or expensive to deploy.

Even then, the ratings are not universal properties.

Skill demand changes with athlete experience. Setup burden changes with the facility. Group scalability changes with equipment quantity, staff numbers, and session organization. Stability demand changes with the variation and prescription. A split squat performed with light dumbbells is not operationally identical to a heavily loaded rear-foot-elevated version.

The library can therefore hold a default rating, while the program, facility, or coach applies a local override when needed.

Do not collapse everything into one score

The temptation is to calculate an overall exercise score:

loadability + scalability − setup burden − skill demand = best exercise

That produces a neat ranking and a questionable decision.

The dimensions are not independent. Setup burden, equipment footprint, and group scalability overlap. High stability demand can be a cost in a maximum-strength slot and the entire purpose of a control slot. High skill demand can be inappropriate today but valuable across a development phase.

The interface should display the profile and allow the coach to prioritize the dimensions that matter for the current slot. It should not quietly turn six debatable ratings into one authoritative number.

A useful workflow is:

1. Use tags to filter exercises that can preserve the job of the slot.

2. Use ratings to compare the remaining options under current constraints.

3. Select and prescribe the exercise plus method.

4. Record substitutions and friction during delivery.

5. Update the defaults only when repeated experience justifies it.

This is Agile Prioritization applied to the exercise library. The system does not try to solve every future decision in advance. It provides a useful default, exposes the trade-offs, and lets feedback from actual sessions edit the model.

The benefit and the danger

Used well, ratings improve search, substitution, template design, and conversation between coaches. They also allow the system to learn where friction repeatedly appears: exercises that are prescribed often but completed rarely, stations that create queues, or options that repeatedly require simplification.

Used badly, ratings create another layer of administrative theatre. Coaches can spend more time scoring exercises than coaching them. The numbers can become stale, context can disappear, and the software can begin recommending what is easy to measure rather than what matters.

The rule remains the same as for tags:

Add a rating only when it changes a real decision.

The detailed map belongs underneath. The visible profile should remain small, interpretable, and editable through experience.

Tags find the options. Ratings expose the trade-offs. The coach still chooses.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

Do we really need another classification of resistance-training exercises?

Probably not, if the goal is to create one more beautiful tree that gets admired for five minutes and then abandoned in a folder. But the question changes once you try to build an exercise library, write a book, design software, create reusable training templates, or simply explain to another coach why two exercises are similar in one way and very different in another.

At that point, the exercises need names, labels, relationships, and some kind of address. The library in Figure 1 is no longer just a list. It becomes a small world that needs organizing.

This sounds simple until you try it. A back squat is obviously a squat. It is also a lower-body exercise, a bilateral task, a barbell exercise, a compound movement, a knee-and-hip-dominant pattern, an eccentric-concentric action, a compressive loading task, and perhaps a main lift, assistance exercise, general preparatory exercise, competition exercise, teaching exercise, or bad idea for today. All of those descriptions can be correct. None of them is the exercise’s single true identity.

The more seriously you take classification, the more the categories multiply. This is the attraction and the danger of the Place of Things perspective. It attempts to describe what is there as completely as possible. It gives us a map of the territory. The problem is that maps can become so detailed that they stop helping us travel.

This first article builds the detailed map. In Part II, I deliberately cut it down and ask a more practical question: which distinctions help us act?

The fantasy of the one true classification

We often talk about categories as if reality arrived pre-sorted. The exercise is either a push or a pull. It is either functional or non-functional. It is either compound or isolation. The coach’s task is supposedly to identify the correct box and place the exercise inside it.

That view works reasonably well when the job is narrow. A powerlifting coach can organize a large amount of training around squat, bench press, and deadlift families. A bodybuilder can organize the same gym around chest, back, delts, arms, quadriceps, hamstrings, glutes, and calves. Both systems can be coherent because they are built around different practical worlds.

The trouble starts when we mistake a locally useful map for a universal one.

Classification systems make certain properties visible and quietly hide others (Bowker and Star 1999). A muscle-group system highlights local muscular work but says little about balance, support, equipment flow, or whether the task is ballistic. A movement-pattern system reveals broad human actions but may hide the local tissue target that matters in rehabilitation. A specificity system makes sense for a specialist but becomes unstable when the same exercise is used by a fighter, an older adult, and a powerlifter.

Human categories also tend to have fuzzy borders. We often reason through useful prototypes rather than perfect definitions (Rosch 1978). A bench press is an easy prototype for upper-body pushing. A landmine press is less clean. A Turkish get-up is not clean at all. The edge cases are not defects in reality. They reveal that the classification was built for a particular purpose.

This is why the correct starting question is not:

Where does this exercise truly belong?

It is:

Which properties of this exercise matter for the task I am trying to perform?

That question still allows detailed description. It simply removes the fantasy that description is neutral or final.

One exercise, many valid lenses

Take the Romanian deadlift. Figure 2 shows why a single label is never enough when the goal is description.

From a fundamental movement perspective, it is a hinge. Some coaches may call it a lower-body pull. From an anatomical perspective, it involves the hip joint, knee and ankle as supporting joints, trunk and spine, hamstrings, glutes, back musculature, grip, and several tissues that refuse to read our labels. From a segment perspective, it is bilateral and compound. From a movement-action perspective, it is usually a grinding exercise performed through eccentric and concentric actions. From an equipment perspective, it is a free-weight barbell task. From a directional perspective, it is mostly sagittal, with a vertically oriented external load and a posterior displacement of the hips.

Then context enters and ruins the neatness.

For a general strength trainee, the Romanian deadlift may be a main hinge. For a powerlifter, it may be supplemental deadlift work. For a sprinter, it may be general posterior-chain loading. For a bodybuilder, it may be a hamstring and glute exercise. For an athlete returning from irritation, a light dumbbell variant may be a control task used to restore tolerance and confidence. The visible movement is similar, but its role changes.

This is not an argument against classification. It is an argument for multi-dimensional classification. The same exercise can belong to several categories because those categories answer different questions.

The Place of Things approach takes this seriously. It tries to preserve the richness rather than compressing everything too early. That makes it valuable for library design, teaching, search infrastructure, and careful thinking. It also creates a monster if every distinction is presented to the user at once.

Different people see different exercise worlds

A classification is constructed in a relationship with use. Figure 3 makes this practical.

The strength specialist is often organizing exercises around the competitive task. A powerlifter has excellent reasons to distinguish the competition lift from a close variation, a supplemental lift, and assistance work. The movement family is anchored to the sport. A weightlifter may organize the library around snatch, clean, jerk, pulls, catches, starting positions, receiving positions, and technical derivatives. The taxonomy grows out of the competitive grammar.

The bodybuilder sees a different landscape. The same bench press is less important as a formal movement pattern and more important as a way of loading chest, triceps, and anterior deltoids. Grip, incline, machine path, range, stability, and local stimulus become meaningful distinctions. This is not a less sophisticated view. It is sophisticated for a different problem.

The strength generalist uses resistance training to support another activity or a broader life objective. The footballer, fighter, recreational athlete, busy parent, and older client may all need pushing, pulling, squatting, hinging, carrying, jumping, and local tissue work, but none of them competes in “the gym.” The exercise has to fit around sport practice, work, recovery, facility constraints, and the person’s actual tolerance.

The clinician or rehabilitation practitioner may use categories that are nearly invisible to the other three. Support level, symptom behavior, joint range, local tissue target, isometric versus dynamic action, exposure progression, and confidence can matter more than whether an exercise looks athletic.

This is closely related to the broader question of archetypes in the gym and the different domains or aspirations that organize training. Who is using the gym? What are they trying to accomplish? What costs are acceptable? The answers influence which categories deserve attention.

The Place of Things map tries to contain all of these views. That is its ambition. It is also the reason the map becomes enormous.

Building the Place of Things map

The simplest way to understand the full system is not as one hierarchy, but as a collection of descriptive lenses. Figure 4 shows the structure. Each lens answers a different question.

Fundamental movement pattern

This is the language most general strength coaches already use. The major families usually include squat or knee-dominant work, hinge or hip-dominant work, upper-body push, upper-body pull, core or trunk work, gait, and a collection of other recurring human tasks such as carrying, crawling, climbing, sled work, rolling, grip, neck, breathing, groin, and hip flexion.

The advantage is immediate recognizability. A coach can look at a week and ask whether it contains broad exposure to pushing, pulling, squatting, hinging, carrying, and trunk work. The categories are useful for program balance and substitution.

The weakness is that movement patterns are not natural laws. Is a split squat a squat, lunge, unilateral lower push, control exercise, or all four? Is a sled push gait, lower-body push, conditioning, or a total-body grind? Is a Turkish get-up core, mobility, control, coordination, or complex work? The system becomes useful when multiple labels are allowed and brittle when every exercise needs one home.

Anatomical description

Anatomy gives us several possible layers.

We can classify by joints: shoulder, elbow, wrist and forearm, hip, knee, ankle and foot, trunk and spine. The spine can be divided further into cervical, thoracic, lumbar, and rib-cage regions. This is useful when a coach or clinician needs to search by region, manage exposure, or build a local support menu.

We can classify by muscles or muscular regions: neck, chest, back, deltoids, triceps, biceps, forearms, quadriceps, hamstrings, glutes, adductors, calves, abdominals, low back, hip flexors. A bodybuilding system can go much deeper because the local target is central to the task.

We can classify by location: upper body, lower body, torso, total body. This is coarse, but coarse is often useful. It supports simple planning, filtering, and workload summaries without pretending we can isolate the body cleanly.

We can also classify by segments and limbs. Is the exercise compound or isolated? Bilateral or unilateral? Double-arm or single-arm? Double-leg or single-leg? Ipsilateral or contralateral? Alternating or simultaneous? Is the working limb straight or bent? These distinctions matter for setup, balance, coordination, local loading, and facility organization.

A rehabilitation-oriented system may add functional muscle classes such as global mobilizer, global stabilizer, and local stabilizer. It may also identify the intended target tissue: muscle, tendon, ligament, joint, nerve, or another structure. These labels can be useful, but they should not seduce us into imagining that the rest of the organism politely switches off.

Movement type and function

Movement pattern describes the broad shape. Movement type describes what kind of action the exercise is trying to organize.

Grinding exercises usually involve producing force against meaningful resistance without releasing the implement or body into flight. Squats, presses, rows, hinges, carries, and many trunk exercises live here. “Grinding” does not mean the rep must be painfully slow. It identifies a family in which force and control against resistance are central.

Ballistic exercises involve acceleration through the action, release, flight, projection, or rapid expression. Olympic lifting derivatives, jumps, throws, punches, kicks, fast pulls, sprint actions, and swings may sit here. The family is broad enough that it immediately demands subcategories.

Patterning exercises are used primarily to teach, restore, rehearse, or simplify an action. A goblet squat may be a strength exercise for one person and a patterning tool for another. The label describes the job, not the object alone.

Control exercises make management of position, balance, support, tempo, or degrees of freedom a major part of the task. Single-leg RDLs, step-downs, unsupported split-stance work, and many trunk exercises can live here.

Coordination and complex exercises involve several interacting segments, transitions, or tasks. These labels are easy to abuse. “Complex” should not become a compliment, and “coordination” should not mean random circus difficulty. The category is useful when task organization is genuinely part of what the coach wants to preserve.

The full map can also include gait and cyclical modes: carry, sled, crawl, walk, run, cycle, row, ski, swim, climb. These actions often sit awkwardly inside a resistance-training taxonomy, but excluding them makes the library less useful for actual preparation.

Muscle action and contraction strategy

The next lens describes how force is produced and managed.

An exercise can emphasize eccentric action through slow lowering, overload, catching, or overspeed. It can emphasize isometric action through holding, pushing against an immovable object, switching, catching, quasi-isometric work, or oscillatory strategies. It can be eccentric-concentric, concentric-only, reactive, rhythmical, or maximal and explosive.

The library can also contain equipment-defined actions such as isokinetic, isoinertial, or near-constant-force resistance. This is where classification becomes technical very quickly. The distinction may be essential in a laboratory, a rehabilitation setting, or a specialized facility, but unnecessary for a coach trying to find three sensible pulling options before the group arrives.

There is another complication: some of these labels may describe the exercise, while others describe the method used with the exercise. Is a slow squat a separate exercise or a normal squat prescribed with a slow tempo? Is an isometric split squat a separate library entry or a split squat with an isometric scheme? The answer depends on how much the setup, communication, search, and analysis change. We will return to this shortly.

Positional, postural, and directional descriptors

A detailed system can describe whether the task is frontal, sagittal, or transverse; anterior, posterior, lateral, or medial; vertical, horizontal, forward, backward, rotational, or diagonal. It can describe flexion, extension, abduction, adduction, pronation, supination, traction, and compression.

It can also describe the body’s position: prone, supine, standing, kneeling, half-kneeling, seated, supported, unsupported, stable, unstable, inclined, declined, flat, bent, straight, flexed, extended, open-chain, or closed-chain.

These descriptors are useful because they allow meaningful filtering. A coach may need an upper-body horizontal pull that is chest-supported. A clinician may need a supported unilateral lower-body exercise. A team coach may need a standing band exercise that can be performed by twelve athletes without a queue.

The danger is obvious. Once every possible descriptor is displayed, selecting a row becomes a tax return.

Resistance, load, and equipment

Equipment is one of the most practical lenses because it connects programming to the actual facility.

The library may include bodyweight, barbells, specialized bars, plates, dumbbells, kettlebells, sandbags, waterbags, weight vests, landmines, chains, medicine balls, hammers, clubs, strongman implements, bands, suspension systems, rings, unstable surfaces, sliders, immovable objects, machines, cable stacks, pneumatic devices, flywheels, motorized resistance, manual resistance, and other increasingly expensive ways to make the athlete uncomfortable.

This classification is not merely logistical. Equipment changes affordances (Gibson 1979). A dumbbell split squat and a barbell back squat are not interchangeable objects with different handles. They differ in setup, stability, loading ceiling, coordination, supervision, spatial demand, and cost to the session.

Facility design and exercise classification therefore influence each other. The idea is developed further in The Gym Is Not Just a Room. A useful exercise library needs to know what equipment an exercise requires because the program eventually has to pass through a real room with finite racks, plates, benches, anchors, and walking paths.

Specificity, priority, and context

This is the most unstable branch.

Exercises are often classified as competition, main, supplemental, assistance, primary, secondary, tertiary, major, minor, general preparatory, specific preparatory, specific developmental, competitive, or rehabilitation work. These labels can be extremely useful inside a defined program.

They are poor candidates for universal properties in a shared library.

A back squat is a competition exercise for a powerlifter in a squat event. It may be a main general-strength exercise for a rugby player, supplemental work for a weightlifter, an optional variation for a bodybuilder, a patterning task for a novice, or temporarily inappropriate for an athlete whose back is irritated. The exercise has not changed. The relationship between the exercise, person, sport, phase, and objective has changed.

The Place of Things system can store possible contextual labels, but it should resist hard-coding them as if they belong to the exercise forever. Context is often added by the user, not discovered inside the object.

Exercise plus method

An exercise name is not a complete prescription. Figure 5 shows the same split squat used in several different ways.

The exercise gives us a task structure: stance, support, movement path, equipment, and broad coordination problem. The method adds load, repetitions, proximity to failure, tempo, range of motion, pauses, density, progression, and intention. Together they form the practical intervention.

A heavy split squat for five repetitions is not the same training task as a thirty-second isometric hold. A slow eccentric version is not the same as a low-load rhythmical version. A supported split squat used after missed training is not the same as an unsupported version used to challenge control.

This matters for classification because we need to decide what belongs in the exercise library and what belongs in the prescription scheme.

A useful rule is to create a separate exercise entry when the distinction changes setup, search, communication, safety, or analysis enough to matter. A front-foot-elevated split squat may deserve its own entry. A supported split squat may deserve its own entry. A tempo of 30X1 probably belongs to the scheme. Three repetitions in reserve certainly belong to the scheme. An isometric push against pins may deserve its own entry because it is a different physical task, not merely a slower rep.

There is no perfect universal boundary. The decision is part of software and library design. The important point is to avoid two opposite errors: creating a separate exercise for every tiny prescription detail, or collapsing genuinely different tasks under one vague name.

The distinction also protects us from assuming that the exercise determines the effect. The same nominal task can create different internal doses and different adaptations across people and days. A related discussion appears in Training Effects: The Dose Is Not the Effect. Classification can describe the planned task. It cannot guarantee the organism’s response.

When the tree becomes a jungle

The full map looks impressive because it is impressive. It captures many meaningful properties. It can support sophisticated search, research, documentation, and analysis. It can also create the interface in Figure 6.

The first problem is entry burden. If every new exercise requires the user to select movement pattern, joints, muscles, limb involvement, tissue target, action type, contraction strategy, position, direction, equipment, range, tempo, priority, and specificity, most users will stop adding exercises. The software may be theoretically complete and practically empty.

The second problem is classification disagreement. Coaches will not agree on every label, and that is not always because one coach lacks knowledge. The categories may be answering different questions. One coach tags a single-leg RDL as hinge. Another tags it as control. A third emphasizes hamstring. A fourth sees a contralateral trunk task. Requiring one answer turns a useful difference of perspective into a data-cleaning argument.

The third problem is false precision. A long list of tags can create a scientific appearance without improving a decision. The fact that an exercise is sagittal, vertically loaded, bilateral, closed-chain, eccentric-concentric, compound, free-weight, knee-and-hip dominant, and quadriceps-glute focused still does not tell us whether it belongs in today’s session.

The fourth problem is context leakage. Labels such as main, assistance, sport-specific, rehabilitation, or athlete-specific are not stable properties. If they are baked into the shared exercise record, one user’s logic becomes another user’s clutter.

The fifth problem is analysis theatre. A database can calculate the number of sets by joint, muscle, direction, equipment, and movement category. Some summaries are useful. Others create the warm comfort of numbers without telling the coach whether the work was performed well, tolerated, timed appropriately, or relevant to the objective.

This is the Is/Ought gap in a tracksuit. More description does not automatically produce better action.

Why the full system still matters

The answer is not to delete the detailed map. The Place of Things perspective remains useful in several roles.

First, it acts as a design dictionary. When building an exercise library, it reminds us which properties might matter and prevents the system from being trapped inside one specialist worldview.

Second, it supports progressive disclosure. The database can remain rich underneath while showing only a few relevant fields to the user. A team coach may open movement type and equipment. A clinician may open joint, tissue, support, and action. A specialist may add a competition-specific layer.

Third, it supports better search later. A tag that seems unnecessary today may become useful when the library grows, the facility changes, or the software needs to suggest substitutions.

Fourth, it makes the limitations visible. Once you see how many valid descriptions exist, it becomes harder to claim that one classification reveals the essence of the exercise.

The Place of Things is therefore best treated as a back-end ontology, a reference map, or a sandbox for thinking. It should not necessarily be the coach’s front-end interface.

That distinction is important. The map can be detailed. The dashboard can be simple. The underlying library can know that a chest-supported dumbbell row is upper body, pull, grinding, horizontal, bilateral, supported, dumbbell, shoulder-and-elbow involving, and back focused. The coach may only need to see upper pull + supported + dumbbell.

Completeness belongs in the basement. Action happens on the ground floor.

The bridge to Part II

The full classification teaches us something important: an exercise is not one thing. It is a node in several overlapping systems. It has a movement shape, anatomical relationships, equipment demands, force actions, positions, directions, methods, and contextual roles.

That richness is worth preserving. But the coach cannot live inside the archive.

Part II crosses the bridge shown in Figure 7. Instead of asking how completely we can describe the exercise, it asks what the classification must help us do. The system will become smaller, less pure, and more useful. It will start with gross categories, allow a hierarchy of tags, keep multiple memberships, and separate stable descriptors from user-defined context.

The Place of Things gives us the vocabulary.

The Forum for Action decides which words need to be spoken now.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

References

Bowker, Geoffrey C., and Susan Leigh Star. 1999. Sorting Things Out: Classification and Its Consequences. Cambridge, MA: MIT Press.

Gibson, James J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.

Rosch, Eleanor. 1978. “Principles of Categorization.” In Cognition and Categorization, edited by Eleanor Rosch and Barbara B. Lloyd, 27–48. Hillsdale, NJ: Lawrence Erlbaum Associates.

Every coach has had this experience. You spend an evening building a clean program. The progression makes sense, the percentages are calculated, every exercise has its place, and the final spreadsheet looks like a small engineering project. Then the athlete walks into the gym with sore adductors, poor sleep, a painful thumb from sparring, and the news that the squat rack is occupied for the next twenty minutes.

Your beautiful plan is not necessarily wrong. It is simply a map meeting the territory.

That distinction matters because most strength-programming software is built as if the map should control the territory. It asks us to define the exact exercise, exact set, exact repetition target, exact load, and exact session before the athlete arrives. In return, it gives us the warm comfort of precision. The problem is that this precision often becomes administrative work for the coach and a prison for the athlete.

The Agile Strength Planner was built from a different assumption: a prescription should help the athlete find the right training zone quickly, but it should not pretend that the coach can predict every interaction in advance. The plan provides a ballpark, an intention, and quality controls. The athlete and coach still need to make decisions.

That is not a weakness in the system. That is coaching. The central tension is summarized in Figure 1.

The promise of software - and the hidden bill

I am not against apps. Apps can be excellent for collecting training logs, analyzing trends, distributing programs, and keeping information synchronized. The problem begins when the interface quietly starts dictating the coaching process. The tool no longer serves the coach; the coach begins serving the tool.

This usually happens in small, annoying ways. You need to enter a separate maximum for every assistance exercise. You can only see one session at a time, so the three- or four-week progression disappears behind menus. The athlete needs to carry a phone around the gym, unlock it, find the next exercise, enter a set, and repeat the ritual. The coach spends more time formatting the program than thinking about the program.

None of these actions is catastrophic. That is why the cost is easy to ignore. But the friction compounds across athletes, sessions, weeks, and seasons. A tool that saves two minutes in theory can steal hours once you include setup, data entry, correction, athlete education, and troubleshooting.

There is also a philosophical cost. When software asks for an exact number, we feel pressure to provide one. A load of 107.5 kg looks more professional than a range of 100-110 kg. Yet the extra decimal does not mean we understand the athlete better. It may only mean that our spreadsheet is more confident than we should be.

The Agile Strength Planner tries to reduce that friction. It uses Excel where Excel is useful: storing information, connecting tables, performing calculations, updating benchmarks, parsing prescriptions, and building reusable templates. Then it deliberately returns the program to a simple output: a printable card that can sit on a clipboard, hang near a station, or travel with the athlete.

Digital planning. Analog delivery. Human coaching.

The middle ground between rigid and vague

Strength prescription often gets pushed toward two unhelpful extremes.

At one extreme, the coach prescribes everything with false precision: four sets of six repetitions at 72.5 percent, followed by three sets of eight at 64 percent, with no room for the athlete who slept badly, performed hard technical training earlier, or simply responds differently to the generic rep-max table.

At the other extreme, the coach writes something like “three hard sets” or “work up to a heavy five.” This preserves freedom, but it can create an unnecessary search problem. The athlete spends several sets figuring out where to begin, overshoots the target, or stays too conservative because the instruction provides no useful anchor.

The practical middle ground, illustrated in Figure 2, is a ballpark estimate plus quality control. You might prescribe three to five sets of four to six repetitions at 65-72 percent, then add “keep at least two repetitions in reserve” or “all repetitions should remain fast and technically clean.” The percentage helps the athlete enter the right neighborhood. The quality control helps the athlete choose the right house.

This is the distinction between precision and significance. A precise statement can still be useless. A less precise statement can be more meaningful because it helps the athlete act under real conditions.

The planner does not remove uncertainty. It organizes it.

The prescription puzzle

The first practical problem is simple: how do you provide useful load recommendations without testing a one-repetition maximum for every exercise in the program?

You cannot realistically test the back squat, front squat, split squat, step-up, Romanian deadlift, dumbbell bench press, landmine press, single-arm press, pulldown, chin-up, row, and every other variation you might use. Even if you could, the process would consume training time, increase fatigue, and produce numbers with a shorter shelf life than we like to admit.

The usual alternatives are equally unsatisfying. You can leave the athlete to guess the load. You can enter values manually for every exercise. Or you can prescribe percentages only for a tiny set of main lifts and abandon useful guidance everywhere else.

The Agile Strength Planner uses a different solution: benchmark movements as reference points. The basic network logic is shown in Figure 3.

Instead of pretending that every exercise needs an independent maximum, the planner allows exercises to be related to a smaller number of benchmark lifts. A dumbbell bench press can be related to the barbell bench press. A Romanian deadlift can be related to a squat, deadlift, or hex-bar benchmark, depending on how you organize your system. The exact relationship is editable because it is a coaching prior, not a universal law.

This is important. The relationship is not “true” in the scientific or metaphysical sense. It is useful enough to provide a conservative starting point. You begin with the best available estimate, observe what the athlete does, and update the benchmark or relationship when reality disagrees.

That is Bayesian updating without the statistics lecture: start with what usually works, test it against what actually happens, and revise the plan.

New athletes do not need to wait for perfect testing

This benchmark logic also helps when a new athlete enters the system. Rather than postponing percentage-based work until a formal testing battery is complete, you can begin with conservative prescription maximums based on the information you have. The initial estimate may use body mass, training experience, known performance, and coach judgment.

The goal is not to guess the athlete’s true one-repetition maximum with laboratory accuracy. The goal is to provide a safe and useful first draft so the athlete can train now. Over the next sessions, plus sets, RIR feedback, velocity data, technical quality, and normal coaching observation give you better information. The benchmark can then be updated.

This is the Minimum Viable Program logic applied to strength prescription: good enough to start, conservative enough to protect the downside, and flexible enough to improve.

The perfect initial number does not exist. A responsible starting number and a tight feedback loop do.

The prescription atom

The core building block inside the planner is what I call the prescription atom. Its three-part structure is shown in Figure 4. It combines three things:

1. The exercise and its relationship to a benchmark.

2. The athlete’s latest available benchmark or prescription maximum.

3. The training scheme written in normal coaching language.

When these elements are combined, the planner calculates the relevant load while preserving the original instruction. This matters because coaches do not think only in isolated numbers. We think in combinations: sets, repetitions, ranges, intent, RIR, tempo, technical cues, back-off work, isometrics, rehabilitation options, and what the athlete should do when the first plan does not fit.

A prescription might read:

3-5 sets of 6-8 repetitions at 60-70%. Keep at least two repetitions in reserve. Stop the set if position deteriorates.

The planner identifies the percentage range, applies it to the selected athlete and exercise, rounds the result according to your equipment, and returns a clean text prescription. The coaching language remains visible. The calculation is automated, but the intention is not lost.

Why a text prescription is more powerful than ten input boxes

Most software separates the prescription into fields: sets, reps, percentage, tempo, RIR, rest, and notes. That structure is useful for databases, but it often fragments the coaching idea. It becomes difficult to write a progression that includes a top set, two different back-off options, an isometric hold, and a qualitative instruction without fighting the interface.

The Agile Strength Planner treats the prescription as a string that can contain both calculable percentages and normal language. That gives you room to build prescriptions such as:

• Work to a top set of five while keeping two repetitions in reserve.

• Repeat the first set for three additional sets.

• Choose one hypertrophy, isometric, or rehabilitation option.

• Complete three to five explosive repetitions while speed remains high.

• Perform a 15-30 second hold using a percentage of the relevant prescription maximum.

The point is not to make every prescription complicated. The point is to avoid forcing every coaching idea into the same rigid form.

Form should follow the problem.

Modularity: give options without producing chaos

The word individualization often creates an image of a coach writing a completely unique program for every athlete. In theory, that sounds impressive. In practice, it can become an administrative nightmare, especially with teams, group settings, or athletes who train remotely.

The alternative is not one-size-fits-all programming. The alternative is modularity.

A modular program has stable training intentions and flexible components. You may prescribe an explosive block, a lower-body strength slot, a horizontal press, a vertical pull, and an assistance block. Inside each slot, the athlete can have one primary option and several acceptable alternatives. The options are not random; they belong to the same functional role in the program.

This is where the planner’s “buffet” logic becomes useful. Figure 5 shows how one stable training intention can support several pre-approved paths. Suppose the goal is lower-body strength for three to five sets of three to five repetitions. One athlete may back squat. Another may use a hex bar because of shoulder irritation. A combat athlete with a painful thumb may choose a machine or belt squat. A remote athlete may use the best option available in a hotel gym.

The coach can provide estimated loads for every option. The athlete gets freedom, but not confusion.

A plan should provide guardrails, not handcuffs

Optionality is sometimes criticized because the athlete might choose the easiest movement. That can happen. But rigid prescriptions do not eliminate poor decisions; they merely hide them until the athlete skips the session, performs the exercise badly, or changes it without telling you.

The answer is not less autonomy. The answer is better constraints and better communication. Define the purpose of the slot. Limit the acceptable options. Explain the decision rule. Review what the athlete selected and why.

For example:

Choose the option that allows hard, technically clean lower-body work without increasing current pain. Stay above two repetitions in reserve. Record the option and top load used.

That is not a free-for-all. It is a decision system.

The best coach is not the one who makes every decision forever. The best coach gradually improves the quality of the athlete’s decisions.

The Scheme Builder: heuristics made editable

Percentage-based programming depends on rep-max relationships, but these relationships vary between athletes, exercises, techniques, tempos, and training histories. A generic table can still be useful, as long as we remember what it is: a prior.

The built-in Scheme Builder, also called the Prescription Generator, allows the coach to create reusable progressions while keeping the assumptions visible and editable. You can adjust the expected number of repetitions at 80 percent, use different multiplication factors for grinding or ballistic work, account for the number of sets, add percentage or repetition buffers, and organize linear or undulating progressions.

This is particularly useful for athletes who are not strength-sport specialists. A football player or MMA fighter may need to preserve speed, technical quality, and energy for sport practice. The load appropriate for one hard set is not automatically appropriate for five sets. The load that allows a grinding repetition is not automatically appropriate when the intention is explosive execution.

As summarized in Figure 6, the Scheme Builder helps convert those coaching decisions into a reusable prescription. It does not prove that the percentages are correct. It saves you from rebuilding the same logic every time.

You can watch the dedicated Scheme Builder walkthrough here:

Why I still want paper in the gym

The phrase “back to paper” can sound nostalgic, as if I want to return to a time before useful technology. That is not the point. I want to use technology at the part of the process where it creates leverage, and remove it where it creates friction.

Excel is excellent for building the program. It can store athlete records, identify the latest benchmark, relate exercises to reference movements, parse percentages, round loads, recycle templates, and generate multiple athlete-specific outputs. A printed card is excellent for delivering the program on the gym floor. It is visible at a glance, does not need a battery, does not send notifications, and can display three or four weeks of progression on one page.

This matters even more in team training. A coach can print athlete-specific loads and place the sheets at relevant stations. Athletes can move through the session without opening an app between sets. The coach can see the entire workout and spend attention on movement, effort, communication, and flow.

The card also becomes a simple training log. The athlete can circle an option, write the actual load, record repetitions, or leave a note. That information can later be used to update the benchmark or next cycle.

The process is not “paper instead of technology.” It is the feedback loop depicted in Figure 7:

Build → calculate → print → coach → observe → update.

Three practical use cases

1. The new athlete

A new athlete arrives without reliable testing data. The coach creates conservative starting benchmarks, generates the first training card, and lets the athlete begin. The early sessions are treated as information-gathering rather than as a final verdict. Loads are adjusted from RIR, performance, and observation. After a few iterations, the prescription maximum becomes more representative.

The athlete trains immediately. The coach learns without wasting a week on ceremonial testing.

2. The remote combat-sport athlete

A fighter is training in another city. The coach does not know exactly how the athlete feels after the morning skill session, what equipment will be free, or whether a small injury will change the exercise choice. A rigid plan creates either compliance theater or constant messaging.

A modular card solves a large part of the problem. The training intention remains stable, but the athlete can choose from pre-approved options and still receive useful load ranges. The quality controls define how hard the work should feel. The athlete records what was done and the coach updates the next iteration.

The program provides direction without pretending the coach is standing beside the athlete.

3. The team workout

A team has one training block and several athletes sharing stations. The coach needs athlete-specific loads, but does not want twenty athletes staring at phones. The planner creates a team sheet or several station sheets with calculated loads for the selected progression step.

The sheets are printed and placed around the gym. The athletes know where to start. The coach retains the ability to adjust. The session moves.

This is not glamorous. It is useful. Useful usually wins.

What the planner does - and what it does not do

The Agile Strength Planner performs the repetitive work that software should perform. It stores athletes, benchmarks, exercises, relationships, schemes, and templates. It pulls the latest benchmark, calculates prescription maximums, parses percentage ranges, rounds loads, and produces individual or team outputs.

It does not know whether the athlete slept well, whether the last sparring session was excessive, whether a painful movement is dangerous, whether the prescribed exercise makes sense for the current sporting problem, or whether the athlete is lying about RIR. It cannot bridge the Is/Ought Gap for you.

That is why the planner should be understood as a coaching tool, not an automated coach.

The goal is not to remove judgment. The goal is to stop wasting judgment on arithmetic and formatting.

The basic principle

Let the workbook handle the repetitive calculations. Let the coach define the intention, constraints, and acceptable options. Let the athlete provide information from the session. Then update.

What you receive

The current Agile Strength Planner package includes the Excel workbook, a library of approximately 200 exercises and movement options, athlete and benchmark tables, the modular Prescription function, the Scheme Builder, individual and team-program templates, printable card layouts, example prescriptions, and instructional videos.

The workbook is intentionally extendable. You can add exercises, create new benchmarks, modify relationships, build your own schemes, and redesign the printouts around your environment. The included material gives you a working system, not a locked black box.

The full package is summarized visually in Figure 8 and contains:

• The Agile Strength Planner Excel workbook.

• Athlete and historical benchmark management.

• Approximately 200 exercises and movement options.

• Editable exercise-to-benchmark relationships.

• Modular prescription parsing with ranges and coaching cues.

• The Scheme Builder / Prescription Generator.

• Individual, remote, team, circuit, and buffet templates.

• Printable three- and four-step strength cards.

• A detailed approximately one-hour walkthrough.

• A separate Scheme Builder instructional video.

• Eligible bug fixes and minor updates.

The offer, without funnel gymnastics

The price of the Agile Strength Planner is €99.

That price is not based on the number of cells in the workbook. A spreadsheet is cheap. The value is in the relationships, logic, templates, prescription engine, Scheme Builder, examples, and the hours of development that allow you to begin with a working system rather than an empty grid.

The more important calculation is what the tool removes. If it saves you from manually calculating and formatting several programs, rebuilding the same schemes, entering dozens of assistance-exercise maximums, or correcting confusion during team sessions, it can repay the purchase in a small number of coaching hours.

More importantly, it can reduce the cognitive residue that comes from boring administrative work. The coach’s scarce resource is not information. It is attention.

Use that attention to watch athletes, ask better questions, communicate, and make decisions.

Get the Agile Strength Planner

Price: €99

The purchase includes the workbook, templates, Scheme Builder, approximately 200 exercise options, the detailed instructional walkthrough, the separate Scheme Builder video, and eligible minor updates.

View the Agile Strength Planner on Payhip

Free for Agile Periodization Skool members

The Agile Strength Planner is available free of additional charge to Agile Periodization Skool community members. The School is where I share courses, working frameworks, tools, PDFs, discussions, and ongoing updates. It also gives us a place to stress-test the planner with coaches who are using it in different sports and environments.

This is not charity and it is not a trick. The community makes the product better. Coaches report bugs, challenge assumptions, suggest templates, share use cases, and reveal problems I would not see alone. In return, members receive access to the tool and the broader educational ecosystem.

Join the Agile Periodization School

Watch before you buy

I do not want you to purchase the planner based on a polished two-minute sales video and then discover that the workflow does not fit your environment. The detailed walkthrough is intentionally long. It shows the logic, the tables, the templates, the modular function, the exercise options, and the limitations.

Full Agile Strength Planner walkthrough:

Scheme Builder / Prescription Generator:

Watch the videos. Look at the workflow. Decide whether it solves a problem you actually have.

That is a better sales process than manufacturing urgency.

Who this is for

The planner is built for strength and conditioning coaches, personal trainers, team coaches, remote coaches, combat-sport practitioners, and serious professionals who already use Excel and want a faster way to move from a programming idea to a usable gym-floor prescription.

It is especially useful when you:

• Work with multiple athletes or groups.

• Use percentage-based loading but do not want to test every exercise.

• Prefer ranges and quality controls over false precision.

• Need exercise options because athletes and environments change.

• Want three or four weeks visible on one printed card.

• Need team-specific loads without phones in the gym.

• Want editable templates rather than a closed application.

It may not be for you if you want a fully cloud-based athlete-management system, automated messaging, a mobile training diary, or a polished app that hides all underlying logic. This is an Excel-based coaching tool. Its strength is transparency, flexibility, and speed of modification.

It also requires a recent desktop version of Microsoft Excel for Mac or Windows with VBA macros enabled. It is not designed for Google Sheets, Apple Numbers, Excel Online, or mobile spreadsheet applications.

The larger idea

The Agile Strength Planner began as a practical answer to a programming problem, but it reflects a larger view of coaching technology.

I do not think the next step is always more data, more dashboards, more fields, and more precision. Sometimes the next step is better compression: a few useful benchmarks, a sensible range, clear quality controls, modular options, a printed card, and a coach who is paying attention.

The model should help us act. The plan should survive contact with reality. The technology should remove friction without removing judgment.

A good tool does not make the coach less important. It frees the coach to do the work that only a coach can do.

Automate the calculations. Preserve the flexibility. Keep the coach in control.

Get the Agile Strength Planner for €99

Access it through the Agile Periodization Skool

There comes a time in every coach’s life when he must climb the mountain, look into the clouds, receive revelation, and return to the village with a message that nobody asked for. This is that moment. I have descended from Mount Common Sense, carrying stone tablets, slightly dehydrated, probably under-recovered, and ready to annoy both traditional coaches and the new ecological priests.

The message is simple: stop starting with drills. Stop starting with methods. Stop starting with exercises. Stop starting with your favorite model, your favorite paper, your favorite guru, or the magical set-and-rep scheme that “built monsters” in 1987. Start with the problem the athlete must solve, then design backward.

This is Problemming: problem-led planning and prescription. Not another ideology. Not another template. Not another church. More like a working irritation. A reminder that the athlete does not perform inside your spreadsheet, your drill library, or your theoretical identity. The athlete performs inside a problem.

So here are the commandments.

I. Thou shalt start with the problem, not the drill

Most coaches do not need more drills. They already have enough drills saved in Instagram folders, YouTube playlists, WhatsApp groups, and that one Google Drive called “GOOD STUFF” that has not been opened since 2019. The bottleneck is not drill poverty. The bottleneck is problem blindness.

A drill is only useful if we know what problem it is meant to solve. Otherwise, we are just moving athletes through choreography and hoping sport performance accidentally shows up at the end, like a confused Uber driver. The first question is not “What are we doing today?” The first question is “What must the athlete solve?”

II. Thou shalt not worship the correct movement

The traditional technical model loves the “correct movement.” Demonstrate it, repeat it, correct it, repeat it again, polish it until it looks beautiful, and then act surprised when it collapses under pressure. This is the coaching equivalent of training a soldier by teaching him how to salute in a mirror while the enemy politely waits outside.

Technique matters, but technique lives inside a task. A basketball shot is not just about arm mechanics. It is a shot selection under a closeout. A football pass is not just inside-foot contact. It is scanning, timing, pressure, and the next action. An MMA jab is not just a shoulder, hip, and fist. It is distance, threat, timing, counters, exits, and the small inconvenience of another human trying to hit you back.

The commandment is not anti-technique. The commandment is an anti-decontextualized technique worship. The game does not ask for clean shapes. It asks for useful solutions.

III. Thou shalt remember that the dose is invited to the party

Problemming is not anti-dose. Dose matters. Load matters. Reps matter. Rest matters. Progression matters. Energy systems, tissue tolerance, fatigue, and recovery are not imaginary creatures invented by old-school strength coaches to ruin ecological dynamics.

The issue is not dose. The issue is dose without a problem. The problem is the anchor; dose is the regulator. A task can have a great representative design and still be dosed stupidly. A conditioning session can produce a beautiful heart-rate trace and still have almost nothing to do with the sport problem. Both errors are common. Both errors are boring.

IV. Thou shalt design tasks, not just deliver instructions

The puppet-master coach tries to control every movement. He pulls the strings, corrects every error, talks through every repetition, and leaves the session feeling productive because his mouth did cardio. The athletes may look organized, but the question is whether they are learning to solve problems or learning to wait for the next instruction.

The problem-led coach is more like a gardener. You do not pull every leaf into position. You design the conditions: soil, water, sunlight, spacing, boundaries, pests, timing. Then you observe what grows and adjust. This is not “do nothing” coaching. Neglect a garden and you do not get ecological dynamics. You get weeds.

Good coaching is not measured by how much you talk. Sometimes you instruct. Sometimes you demonstrate. Sometimes you change the task. Sometimes you constrain the space. Sometimes you ask a question. Sometimes you shut up and let the athlete discover the information you were about to over-explain.

V. Thou shalt not confuse sport-specific with sport-looking

A cable punch is not automatically boxing-specific. A lunge with a ball is not automatically football-specific. A band pull that looks vaguely like swimming is not automatically useful. This is cosplay specificity: dressing as an exercise in a sports costume and pretending transfer has been summoned. It is the coaching equivalent of wearing a lab coat to make your opinion look more scientific.

Representative does not mean “looks similar.” It means the task preserves enough of the information that regulates the action. In tennis, that might mean reading the opponent, not just hitting balls from a machine. In martial arts, that might mean distance, timing, threat, and consequence, not just abusing a heavy bag that cannot file a complaint. In team sports, that might mean scanning, pressure, space, and decision-making, not cone ballet.

But there is another side to this commandment. Some useful training does not look like the sport because it is not trying to train the sport form directly. It is trying to build the underlying substance. A squat does not look like a double-leg takedown. A Romanian deadlift does not look like a football sprint. A row does not look like wrestling. Neck isometrics do not look like MMA. Good. They are not supposed to win a costume contest. They are there to build tissue capacity, strength reserve, robustness, and action capabilities that may support the sport later.

The question is not “Does it look like the sport?” The question is “What is this intervention for?” If the goal is skill transfer, preserve the information that guides action. If the goal is physical preparation, build or protect the substance that allows the athlete to practice and compete. Same athlete. Different job.

VI. Thou shalt explore before thou pretendest to know

Sometimes we do not know what the problem is. This is inconvenient for coaches because uncertainty is bad for selling systems and looking wise on podcasts. But with a new athlete or group, the honest answer is often: “I do not know yet. Let us probe.”

Exploration means using broad tasks to reveal bottlenecks. Who panics under pressure? Who looks good in drills but disappears when the opponent becomes real? Who is physically overwhelmed? Who is strong enough but perceptually late? Who loses rhythm when the task changes? Who cannot coordinate basic human movement without looking like a folding chair in a storm?

One practical way to do this is the problem sandwich. You start with a problem, not a lecture. Let the athletes wrestle with the task first. This explores both the athletes and the problem itself. They feel where they get stuck, and you get to see what actually breaks: perception, timing, coordination, confidence, physical capacity, emotional control, or simple misunderstanding.

Then, when the struggle is visible, you insert the middle layer of the sandwich: a principle, cue, analogy, demonstration, or short piece of direct instruction. Now the athlete has a reason to listen. The instruction lands on a lived problem instead of floating in the air like coaching perfume. After that, you return to the problem. Same problem, slightly changed problem, or a more costly version of the problem. The athlete tests the principle under pressure, and you observe whether the behavior actually improves.

This is exploration turning into exploitation. We begin broadly, probe reality, identify the bottleneck, provide the smallest useful intervention, and then put the athlete back into action. We stop pretending we knew from the first warm-up. This is not a weakness. This is coaching under uncertainty.

VII. Thou shalt use instruction like medicine, not confetti

Instruction can be useful. Demonstration can be useful. Explicit coaching can be useful. Anyone saying “never instruct” should be forced to teach beginners how to avoid dangerous joint locks using only interpretive dance.

But instruction has a dose. Too much of it can poison the task. Athletes become self-conscious, internally focused, dependent on the coach, and disconnected from the information in the environment. They start thinking about elbow angle, knee tracking, shoulder position, hip rotation, breathing, and whether the coach is silently disappointed in their existence.

This is why the problem sandwich works so well. Instruction is often more useful after the athlete has encountered the problem (see the problem sandwich above). First, they feel the gap. Then the coach offers a cue or principle that helps organize the solution. Then the athlete returns to the task and tests it again. The instruction is no longer random information dumped on a cold brain. It becomes an answer to a question the athlete now actually has.

The question is not “explicit or implicit?” The question is “What information does this athlete need right now, and when will it be most useful?” Sometimes the answer is a cue. Sometimes it is an analogy. Sometimes it is a constraint. Sometimes it is a question. Sometimes it is silence. The coach is not paid by word count, which is tragic for some of us.

VIII. Thou shalt respect Substance ~ Form

Substance is the underlying resource: tissue, strength, mobility, aerobic power, robustness, capacity. Form is how that resource is organized and expressed: timing, perception, coordination, rhythm, decision-making, and behavior under constraints. The traditional coach often reduces sport to substance. The new ecological zealot sometimes reduces everything to form. Both can become ideological. Same disease, different costume.

An apple is not only nutrients, and coffee is not only caffeine. You can describe coffee as 150-200 mg of caffeine plus fluid and aroma compounds, but coffee is also grinding beans, boiling water, taking a break, reading the newspaper, sitting with friends, or having ten quiet minutes before the day starts punching you in the face. Same substance. Different form. Different meaning.

Training is like that. A bigger squat means one thing for a powerlifter and another thing for an MMA fighter. For the powerlifter, the squat is the sport problem. For the MMA fighter, it is one possible substance input that may or may not help solve fighting problems. Same gym. Different meaning.

And this is where many coaches lose the plot. They hear “problem-led” and assume every useful training effect must come from a representative task. Not so fast. We still need substance work. We still need physical preparation. We still need to build qualities directly when the sport does not reliably build them, when the sport builds them too slowly, or when the sport keeps threatening them like a drunk uncle at a wedding.

I like Anastasios Karamitros’ line here: the purpose of strength and conditioning is to make athletes better at their sport by increasing both the quality and quantity of their sports practice. That is the point. Physical preparation is not there to cosplay as the sport. It is there to top up what the sport misses and protect what the sport threatens.

That substance work can also be organized from broad to specific. At the broad end sits human-specific work: basic movement capability, tissue tolerance, coordination, strength, running, jumping, throwing, carrying, and not collapsing when asked to move like a functioning mammal. Then we move to sport-specific work: the demands the sport repeatedly creates, the qualities it misses, and the threats it tends to overload. Then we narrow to athlete-specific work: this body, this injury history, these weak links, these capacities. Finally, we narrow to event-specific work: travel, taper, schedule, climate, surface, weight cut, or whatever fresh nonsense the calendar has prepared for us.

The skill side can follow a similar broad-to-specific logic, but the object of attention changes. We can begin with broad sport problems that reveal how athletes perceive, decide, and act. Then narrow toward role or position problems, then athlete-specific learning problems, and finally event- or opponent-specific tactical problems. The point is not to memorize the pyramid like it is sacred geometry. The point is to avoid rushing into athlete- or opponent-specific cleverness before the generic problem is even visible.

So yes, the best way to improve MMA preparation is still to do MMA. The best way to improve swimming is still to swim. The best way to improve football is still to play football. But the sport does not always give us every quality we need, and it can be too costly to use the sport itself as the only tool. That is why physical preparation matters. It tops up what the sport misses and protects what the sport threatens. Problem-led skill work and substance work are not enemies. They are teammates who occasionally annoy each other, which is the healthiest kind of relationship.

IX. Thou shalt defend before thou attackest

Nassim Taleb’s barbell strategy is a useful way to think under uncertainty: put most exposure on the safe, robust side and a smaller portion on the risky, high-upside side. In coaching language: protect the downside first, then explore upside with bounded bets.

The left side is not “boring old basics because grandpa said so.” It is the foundation that keeps the athlete alive in the process: human-specific capabilities, tissue tolerance, basic sport problems, recovery, safety, and the removal of stupid risks. The right side is where we experiment: athlete-specific constraints, opponent-specific scenarios, event-specific tuning, tactical bets, and interventions that might unlock upside.

Fragile coaches jump straight to novelty because novelty feels intelligent. Robust coaches cover the base, then experiment. Defend before you attack. Then attack.

X. Thou shalt iterate and update, or thou shalt become an ideology

This is the final commandment, and it is the one that saves problemming from becoming another stupid flag. The plan is a hypothesis. The task is a probe. The prescription is a bet. The athlete is not obligated to validate your model. Reality is not your intern.

This is also where Problemming connects directly with Agile Periodization. We do not write the perfect plan and then defend it like sacred scripture. We work in short loops. We design a task or prescription, observe the response, review what happened, run a small retrospective, and update the next iteration. The session is not only training. It is also information.

In practice, this means we need small “ceremonies” that keep the learning loop alive. Before the session: what problem are we probing today? During the session: what behavior are we trying to see? After the session: what actually happened? What transferred under pressure? What collapsed when the cost increased? What did the athlete notice? What should we keep, remove, simplify, dose differently, or test next time?

Different practice intentions matter here. Sometimes we practice to maintain: keep the skill alive with low cost. Sometimes we practice to develop: stretch the athlete into a harder problem where learning is expected to be ugly. Sometimes we practice to transfer: make the task more representative and costly, so the skill has to survive closer to performance conditions. Sometimes we need test-match practice: not because we worship performance in training, but because we need to see what holds when pressure, scoring, fatigue, opposition, or consequence enters the room.

This is the tension between performance and learning. A session that looks clean may only prove that the task was too easy. A session that looks messy may be productive learning, or it may be chaos with branding. The coach has to judge the difference. Are we maintaining confidence? Developing a new solution? Testing transfer? Preparing for competition? Different goals require different levels of difficulty, feedback, opposition, and cost.

If the athlete does not improve, update. If the task does not create the behavior, update. If the dose is too costly, update. If the athlete is solving the wrong problem beautifully, update. If your favorite method does not work here, grieve privately and update publicly.

Traditional coaches can become trapped in drills and corrections. New-school coaches can become trapped in terminology and vibes. The problemming coach should be trapped by neither. The job is not to be loyal to a school. The job is to be loyal to the athlete, the problem in front of you, and the feedback reality keeps sending while you are busy pretending your plan was obvious all along.

The short version, before someone makes a certification

Problemming means we start from the problem athletes must solve, then design the task, dose, instruction, environment, and physical preparation around that problem. It is not anti-technique, anti-S&C, anti-instruction, anti-data, or anti-drills. It is anti-randomness disguised as methodology.

Use drills when they help. Use dose when it regulates. Use instruction when it clarifies. Use constraints when they reveal. Use physical preparation when the body needs substance. Use representative tasks when information matters. Use exploration when you do not know. Use exploitation when the bottleneck is visible. Use the problem sandwich when instruction needs to land on lived experience. Use the barbell strategy when uncertainty is high. Use Agile Periodization loops to probe, review, retrospect, and update. Use humility always, preferably before reality slaps you with it.

The commandments are not sacred. They are tools, not truths. But if your session begins with a drill and only later tries to invent a reason for it, you may need to climb the mountain again.

Bring water.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

Why this discussion matters

Every coach has had this moment. The plan is clean, the microcycle makes sense, the loads are calculated, and the week has a beautiful internal logic. Then reality walks into the gym with muddy shoes. The athlete slept badly, the head coach changed the session, the field is unavailable, Tuesday's sparring turned into a small civil war, and now your optimal plan looks like a museum piece from a more obedient universe.

This is not a failure of coaching. This is coaching. The failure is pretending this is the exception rather than the normal working condition.

This article is a fictional dialogue between two coaches, but the tension is real. Jake Reynolds is the reductionistic strength coach. He likes clean categories: strength, power, conditioning, mobility, readiness, fatigue. He is not stupid. In fact, he is useful because he forces clarity. But he tends to move too quickly from what I can measure to what I should prescribe.

Ryan Miller is more pluralistic. He still likes physiology, testing, biomechanics, and training theory, but he is suspicious of any model that pretends to remove judgment. He thinks in terms of constraints, bottlenecks, leverage points, trade-offs, and iterative experiments. In other words, he is closer to the Agile Periodization stance: models are tools, not truths.

They are drinking coffee after a mixed gym session. One athlete is a powerlifter preparing for a meet. Another is an MMA fighter in camp. Both squatted today. Same rack. Same bar. Same plates. Different world.

That is the entry point. The overview image in Figure 1 gives the big picture: the model starts with the old substance/form distinction, then becomes more useful through the race-car analogy and the nested intervention logic.

Same squat, different world

Coach Jake: I still think you are making this more complicated than it needs to be.

Coach Ryan: That is always possible. Most coaching theories start as useful tools and end as decorative furniture.

Coach Jake: Here is my problem. Strength is strength. Power is power. Conditioning is conditioning. If the athlete needs a stronger squat, we work on improving the squat. If he needs better aerobic fitness, we can improve it. If he fades in round three, we condition him harder. Why do we need Substance ~ Form on top of that?

Coach Ryan: Because you just jumped from a description to a prescription. That jump is where a lot of coaching stupidity hides.

Coach Jake: Strong opening.

Coach Ryan: I am not saying your categories are useless. Strength, power, endurance, speed, robustness - these are useful abstractions. The problem starts when we treat them like little substances hiding inside the athlete. As if “strength” is a thing in a box, and the coach’s job is simply to fill the box.

Coach Jake: And you think substance and form helps?

Coach Ryan: It can. But only if we use it as a lens, not as a religion.

The back squat is a good example because it is familiar, emotionally loaded, and easy to misunderstand. A powerlifter, a weightlifter, a bodybuilder, a football player, and an MMA fighter can all perform something that looks like a back squat. Same rack. Same bar. Same plates. Same general movement pattern. But the coaching meaning of that squat is not the same.

For the powerlifter, the squat is the sport. It is judged directly. Depth, commands, bar path, attempt selection, body mass, peaking, arousal, and technical consistency all matter because the squat itself is the performance problem.

For the MMA fighter, the squat is usually support work. It might build lower-body strength, tissue capacity, confidence, or general robustness. But nobody gets their hand raised because his back squat looked beautiful in the gym. The fight problem lives somewhere else.

Coach Jake: So the powerlifter and the MMA fighter are both squatting, but they are not really doing the same thing.

Coach Ryan: Exactly. The external exercise is the same. The coaching meaning is different.

This is where the Substance ~ Form distinction earns its keep. It gives us a way to ask: what is the underlying capacity, and how is that capacity organized into action?

A simple definition before we get lost

Coach Jake: Okay, define it cleanly before we start sounding like two drunk philosophers at a conference bar.

Coach Ryan: Fair. In this context, I would define it like this:

Substance is the underlying material, resource, capacity, tissue, physiology, dose, engine, or potential.

Form is the organization, expression, skill, timing, rhythm, tactic, task, intention, or constraint that gives substance its meaning in action.

That is the simple version. In running, VO₂max, critical speed, W′, glycogen, and heat tolerance can be treated as substance. Pacing, rhythm, tactical positioning, surges, and when to commit can be treated as form. In MMA, aerobic base, repeat power, strength, grip, neck robustness, and weight-cut recovery can be treated as substance. Pace, distance management, shot selection, transitions, pressure, and composure under threat can be treated as form.

The old philosophical anchor is Aristotle’s matter and form. Bronze is not a statue until it is organized as a statue, which is the first panel in Figure 1. The bronze is the material; the statue is the formed expression. That is a useful starting point, but I would not push it too far. Athletes are not bronze. They adapt. They anticipate. They get tired, emotional, bored, injured, and occasionally do something brilliant that ruins our tidy model.

That is why the race-car analogy is often better for coaches. Figure 3 makes the same point visually: the car has an engine, chassis, fuel, brakes, tyres, and aero. But the lap time also depends on the driver, racing line, timing, pacing, traffic, tyre management, and when to spend the battery. Even this analogy is imperfect, because cars are more decomposable than people. But it gets the point across.

Coach Jake: And in the squat?

Coach Ryan: For the powerlifter, squat performance has substance and form. Substance might be muscle mass, neural drive, leverages, connective tissue capacity, and fatigue resistance. Form might be stance, bracing, descent speed, depth, bar path, sticking-point strategy, commands, and attempt selection.

For the MMA fighter, the squat itself might mostly be substance at the fight level. It supports the athlete’s ability to tolerate and express force. But whether that substance matters depends on the form of action in the cage. A stronger squat might help a fighter. Or it might not. It depends on the bottleneck.

As Table 1 suggests, the problem gives the exercise its meaning. That does not mean the exercise has no physiological reality. Of course it does. A heavy squat is still a heavy squat. But the coaching significance of that squat depends on the athlete, the sport, the phase, the constraints, and the bottleneck.

The specialist sees a different gym

Coach Jake: This is where specialists and generalists split, right?

Coach Ryan: Yes. A specialist sees the gym differently because the gym task is much closer to the competitive task.

A powerlifter lives inside the squat, bench press, and deadlift. A weightlifter lives inside the snatch and clean & jerk. A bodybuilder lives inside local mechanical tension, muscle size, fatigue management, and body-part development. The gym is not merely a preparation environment. It is the actual performance world, or very close to it.

For a generalist, such as an MMA fighter or field sport athlete, the gym is usually one step removed. Sometimes two or three steps removed. The gym supports the sport problem, but it does not replace it. That changes the level of precision that makes sense.

Coach Ryan: With a powerlifter, I can care deeply about a 2.5 kg jump, the exact opener, squat depth, whether the hips shoot back, whether the bar drifts forward, and whether the athlete is peaking too early. These details are not decoration. They are close to the target. See Figure 2.

With an MMA fighter, if I obsess over whether he squats 170 or 175 kg, I might be polishing the wrong stone. Maybe his lower-body force is not limiting anything. Maybe he gasses because he overcommits on every takedown. Maybe he loses clinch exchanges because of head position, not strength. Maybe he looks “unconditioned” because he cannot relax under threat.

Coach Jake: So the leverage point changes.

Coach Ryan: Exactly. That is the whole point.

If a powerlifter misses a squat, the leverage point might be inside the squat itself. Maybe he loses position out of the hole. Maybe his bracing fails. Maybe his attempt selection was too aggressive. Maybe the training block built strength but not enough exposure to heavy singles.

If an MMA fighter gets taken down or fades in the third round, the leverage point may not be inside the squat at all. It might be distance management, grip endurance, pacing, poor carbohydrate intake, inefficient wrestling, or simply too much accumulated fatigue from the week.

A reductionistic coach sees the failed action and immediately reaches for a quality: “he needs more strength,” “he needs more conditioning,” “he needs more power,” “he needs more mobility.” Maybe. But maybe not. A better first question is not “which quality do I want to improve?” It is:

What is the actual bottleneck?

Not the favorite quality. Not the fashionable metric. Not the exercise you like prescribing. The bottleneck.

The nested rabbit hole

Coach Jake: Earlier you said every form has substance and every substance has form. That sounds nice, but also like the kind of phrase that can get annoying very quickly.

Coach Ryan: It can. But it is useful.

The Substance ~ Form split is scale-dependent. What appears as substance at one level can be split again into substance and form at a lower level. What appears as form at one level has its own substrate underneath.

Take endurance running. At the race level, we might say:

• Substance: VO₂max, threshold, critical speed, W′, glycogen, heat tolerance.

• Form: pacing, rhythm, surges, positioning, when to commit.

Now zoom into VO₂max:

• Substance: oxygen delivery, cardiac output, hemoglobin, blood volume, mitochondrial capacity.

• Form: running economy, breathing pattern, posture, stiffness, technical expression.

Now zoom into pacing:

• Substance: physiological budget, speed reserve, fatigue tolerance, perception of effort.

• Form: even pace, negative split, surge and float, sit and kick, responding to opponents, committing at the right time.

So yes, pacing is form at the race level. But pacing itself has substance and form when we zoom in.

Figure 4 shows the endurance example as a static slide. Figure 5 gives the same logic.

Coach Jake: So we can keep zooming.

Coach Ryan: Yes, but we should not keep zooming just because we can. That is how coaches disappear into the swamp. The rule is simple:

Decompose only until you find a useful lever.

The rabbit hole is infinite. Coaching is not.

The practical point of Table 2 is not to create a perfect taxonomy. The point is to avoid premature prescription. “Bad pacing” might mean the athlete has poor judgment, poor race experience, poor speed reserve, poor heat tolerance, poor perception of effort, or the wrong tactical model. Same symptom. Different lever.

Work rate: measured in endurance, inferred in MMA

Coach Jake: In running or cycling, at least we can measure work rate. Power, pace, velocity, heart rate, maybe lactate. It is not perfect, but you have a handle.

Coach Ryan: Exactly. Endurance sports are not simple, but they are friendlier to measurement. In cycling, power gives you a clean external work-rate signal. In running, velocity or pace gives you something useful, even if terrain, wind, heat, and biomechanics complicate it.

MMA is different. What is work rate in a fight? Number of strikes? Takedown attempts? Scramble density? Clinch pressure? Ground control? The cost of being pinned under a heavy human? The effort of defending a bad position? The panic tax of being stuck against the fence?

You can measure events, but you infer cost.

Figure 6 is the cleanest visual contrast here. The critical power and W′ model gives endurance coaches a useful map: there is a sustainable power and a finite work capacity above that line. In MMA, the metaphor becomes a box of matches. Every hard flurry, failed takedown, scramble, defensive burst, and panic reaction burns a match. Skillful fighters do not avoid spending matches entirely. They spend them on purpose.

Coach Jake: So “avoid the red zone” is not the same as never going hard.

Coach Ryan: Correct. Avoid wasting the red zone. Save matches when you can. Spend them when the exchange is worth it. That is why conditioning is not only about having more matches. It is also about pacing, economy, composure, and tactical judgment.

This is where Substance ~ Form helps. The substance might be repeat power, local endurance, aerobic support, and recovery between bursts. The form is how the fighter uses that budget: when to press, when to stall, when to shoot, when to disengage, when to accept bottom, when to stand up, and when to burn the house down.

Precision vs significance

Coach Jake: This brings us to the prescription. Where does precision fit?

Coach Ryan: Precision is useful when it is meaningful. But in messy environments, coaches often confuse precision with wisdom.

A precise prescription might say: back squat 4 × 6 at 78.3% 1RM, 2:37 rest, target velocity 0.62 m/s. That might be useful in the right context. For a powerlifter or a controlled strength block, maybe that level of detail makes sense.

But for an MMA fighter in camp, after two hard grappling sessions, a poor night of sleep, and a thumb that feels like a sausage, that prescription might be fake precision. The numbers look scientific, but the meaning is unstable.

A more significant prescription might be: ramp up to a heavy five around 120-130 kg. Take enough rest. Then do two or three back-off sets, 10-20% lighter. Stop if speed, shape, or intent collapses.

Figure 7 is the coaching version of the same idea: exact numbers are not always the most meaningful message.

Coach Jake: Less precise, but maybe more useful.

Coach Ryan: Exactly. The second prescription gives the coach and athlete room to steer. It respects the fact that readiness changes, especially when the sport practice creates unpredictable costs.

Precision answers: “What exactly did we prescribe?” Significance answers: “What are we trying to accomplish, and how do we know when to adjust?” In complex environments, significance often beats false precision.

Place of Things vs Forum for Action

Coach Jake: This connects to your Place of Things and Forum for Action distinction.

Coach Ryan: Yes. The Place of Things view describes the world in categories: exercises, muscles, joints, intensities, heart rates, velocity zones, force outputs, and test scores.

That view matters. I am not against it. I like tests. I like force plates. I like physiology. I like models. I like numbers, probably more than I should.

But the Forum for Action view asks: what is this athlete trying to do, in this context, under these constraints, and what should we do next?

The Is/Ought Gap lives between those two worlds. The “is” might be: the athlete’s squat is up 10 kg, grip test is down, submax heart rate is elevated, and sparring video shows him fading after failed takedowns. The “ought” is the coaching decision: reduce volume, add grip top-ups, change sparring constraints, improve pacing, adjust recovery, modify the microcycle, or ignore the noise.

The data does not automatically tell you the decision. You still need theory, context, values, goals, and judgment. Sorry. Coaching still requires a coach.

The cycle in Figure 8 is the bridge between description and action.

MMA: big rocks, pebbles, and top-ups

Coach Jake: Let’s apply this to MMA, because that is where reductionism usually gets exposed.

Coach Ryan: MMA is a beautiful mess. It is striking, wrestling, grappling, pacing, skill, threat, fatigue, damage, psychology, weight cuts, opponents, tactics, and chaos wearing gloves.

The best way to prepare for MMA is to do MMA.

That is the big rock.

But that sentence is not enough, because MMA practice is costly. You cannot hard spar every day. You cannot punch partners in the head to develop punching power. You cannot squeeze someone’s neck maximally every session to develop squeeze endurance. You cannot simulate every fight demand without accumulating damage.

So physical preparation sits in a support role. It should improve the quality and quantity of sport practice. It should protect the athlete from downside. It should fill gaps that MMA practice cannot reliably or safely fill.

Figure 9 is the MMA version of the big-rocks argument. The hierarchy is not complicated: “problemming” first, support work second. Strength, conditioning, top-ups, testing, recovery work, and mobility are pebbles and sand. Important pebbles and sand, but still not the rock. If you put the pebbles first, the rock might not fit.

Leverage point is only a hypothesis

Coach Jake: Here is where I still worry. You keep saying “find the leverage point.” That sounds like we can identify the magic lever and then pull it.

Coach Ryan: Good catch. That is exactly the shadow side of the model.

A leverage point is not a fact. It is a hypothesis.

We are dealing with complex adaptive systems. The athlete is not a vending machine. You do not insert an intervention and receive the predicted adaptation. You change one thing, and the system responds through physiology, psychology, behavior, readiness, motivation, tissue tolerance, skill expression, coaching culture, and the rest of the week.

So when I say “this is the bottleneck” or “this is the lever,” what I really mean is: this is my best current explanation of what seems to be limiting action, and this intervention is my best current probe.

The phrase “protected intervention” in Figure 10 matters. If we might be wrong, we should not put all eggs in one basket. We should not destroy the athlete’s week because we fell in love with our diagnosis. This is where Nassim Taleb’s barbell strategy becomes a useful coaching metaphor, summarized in Figure 11.

Barbell strategy: protect before you attack

Coach Jake: I know the financial version of the barbell strategy. Safe on one side, speculative on the other. How does that work in training?

Coach Ryan: In coaching, the left side of the barbell protects the downside. The right side pursues upside.

The left side is via negativa: remove the stupid, reduce fragility, protect availability, cover the fundamentals, keep the athlete broad enough, and do not create unnecessary cost. This is where human-specific work and sport-specific protection live very naturally. You keep the athlete a functioning human. You protect the tissues and capacities the sport tends to punish. You avoid breaking the thing you are trying to improve.

The right side is via positiva: targeted development, event-specific preparation, opponent-specific scenarios, athlete-specific bottleneck probes, and small upside experiments. This is where we try to remove the hypothesized bottleneck.

But the right side is not guaranteed. That is the whole point. If our bottleneck diagnosis is wrong, the left side keeps us safe from catastrophic stupidity.

Coach Jake: So the barbell is protection against our own confidence.

Coach Ryan: Exactly. The barbell strategy is not just a training allocation. It is an epistemic humility device. It says: “I have a hypothesis, but I might be wrong, so I will not bet the whole organism on it.”

That is the difference between robust coaching and fragile optimization.

Human-specific, sport-specific, event-specific, athlete-specific

Figure 12 turns the barbell logic into a programming hierarchy. The Programming Pyramid is useful here because it gives the barbell strategy a practical body. At the base, we have human-specific work. These are broad patterns and capacities: squatting, hinging, pushing, pulling, carrying, bracing, landing, jumping, sprinting, throwing, trunk work, and basic tissue capacity. This layer is not sexy. It is also the layer that stops a lot of nonsense.

Above that, we have sport-specific protection. In MMA, that might mean neck robustness, shoulder support, trunk integrity, grip, adductors, hip flexors, calves, wrists, posting strength, and the ability to tolerate awkward contact. This is not fake specificity. It is not pretending that a cable exercise with gloves is suddenly MMA. It is asking what the sport tends to punish and then protecting the athlete from that downside.

Then, on the right side of the barbell, we have event-specific and athlete-specific work. Event-specific means camp, opponent, match, rules, surface, travel, weight cut, tactical demands, or tournament format. Athlete-specific means the actual person in front of us: injury history, morphology, preferences, current readiness, psychological profile, training age, and the hypothesized bottleneck.

The mistake is thinking that “individualized” means everything must be unique from day one. That is usually overfitting with a coaching whistle. A good generic/fundamental program can achieve a lot because most athletes share many human problems before they reveal their individual problems.

This is the bias-variance idea in coaching clothing. A generic program has bias. It assumes some things are probably useful for most humans and most athletes. But it usually has lower variance, which means it is more stable, more coherent, easier to deliver, and less likely to become a random menu of personal tweaks. A fully individualized program tries to reduce bias, but if you do not have enough real information, it increases variance. It becomes noisy. It starts fitting yesterday’s soreness, today’s mood, and the coach’s anxiety.

Table 3 is not a mathematical claim. It is a coaching analogy. But it is a useful one. You do not reduce bias by inventing complexity. You reduce bias by learning. And until you learn enough, a robust generic base is not laziness. It is good risk management.

Agile microcycle: how the model becomes practical

Coach Jake: So how does this become a weekly planning system and not just a nice article?

Coach Ryan: Through iteration.

Agile Periodization, at least the way I currently use the term, is not “make it up as you go.” It is not random flexibility. It is disciplined adaptation. You start with a prior, build a minimum viable plan, expose it to reality, review what happened, and update.

In microcycle terms, that means the week is not a prison. It is a short experiment. You put the big rocks in first, protect the downside, run one or two targeted probes, and then review. Did the fighter tolerate the week? Did the top-up help? Did the sport session expose a different bottleneck? Did the supposed conditioning problem turn out to be pacing? Did the grip issue disappear when the athlete stopped wasting energy in bad positions?

This is where the Substance ~ Form lens, barbell strategy, human-specific pyramid, and bottleneck thinking all connect. The human-specific base gives you a stable MVP. The sport-specific layer protects the athlete from predictable sport costs. The event- and athlete-specific layers give you targeted experiments. The review tells you whether the hypothesis survived contact with reality.

A small table for not fooling ourselves

Coach Jake: Give me the practical diagnostic table. Coaches love tables because they can pretend it is not philosophy.

Coach Ryan: Fair. Here is a useful starting point.

The key word in Table 4 is possible. Not proven. Not obvious. Possible. That one word keeps the coach honest.

Substance ~ Form is not reductionism

Coach Jake: Let me push back. Isn’t this still reductionistic? We are splitting the athlete into parts.

Coach Ryan: It becomes reductionistic if we forget that the split is a lens.

Substance ~ Form is not reality. It is a map. And like all maps, it simplifies. It helps us orient, but it also hides things.

The athlete is not made of separate boxes called strength, conditioning, speed, skill, psychology, and tactics. Those are abstractions we use because we need to think and act. The real athlete is a living system. The parts interact. The current state changes. The same session can have a different effect depending on sleep, stress, partner, rules, mood, injury, weight cut, and yesterday’s chaos.

So we do not use Substance ~ Form to reduce the athlete. We use it to ask better questions.

What is the athlete trying to do? What substrate does that action require? How is that substrate organized in the task? Where is the bottleneck? What is the smallest useful intervention? How will we know if we were wrong?

That is not reductionism. That is pragmatic decomposition. You break the system apart only long enough to find a useful lever. Then you put it back into action.

Final working model

Coach Ryan: Here is the cleaned-up version.

Substance and form are not fixed boxes. They are roles inside a level of analysis. At one level, VO₂max might be substance and pacing might be form. But when you zoom into pacing, it has substance and form. It has a physiological budget, perception of effort, speed reserve, and fatigue tolerance. It also has structure: even pace, negative split, surge and float, when to commit.

At one level, a back squat might be substance for an MMA fighter. But when you zoom into the squat, it has form: stance, depth, bar path, bracing, tempo, intent. And when you zoom back out, the improved squat only matters if it supports the fight problem.

At one level, MMA conditioning might look like substance: aerobic base, repeat power, local muscular endurance, grip, neck, trunk. But in the cage, that substance only matters when organized into pacing, distance, shot selection, transitions, pressure, composure, and problem-solving under threat.

So the practical line is:

Every form has a substrate. Every substrate has a form. Coaching is finding which level is limiting action.

Coach Jake: And then not getting lost in the model.

Coach Ryan: Correct. The model is a tool. Not a shrine.

Closing: the gym is not just a place of things

The gym is not just a room filled with exercises. It is a forum for action.

A squat rack means one thing to a powerlifter, another thing to a weightlifter, another thing to a bodybuilder, another thing to a football player, and another thing to an MMA fighter in camp. The equipment is the same. The substance might look similar. The form is different. The problem is different. The leverage point is different.

This is why physical preparation should not start with exercises. It starts with the problem.

Then we ask what substance supports that form of action. We ask whether that substance is already being developed by the sport. We ask whether the sport develops it safely, reliably, and at the right dose. Then we decide whether to supplement, top-up, protect, remove, or leave it alone.

That is the craft.

Not reductionism. Not mysticism. Not “everything is specific.” Not “the lab explains all.” Not “the spreadsheet knows.” Just pragmatic coaching under uncertainty.

Big rocks first. Pebbles second. Sand only if it fits.

And beer, of course, after the session.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

Every coach has had this moment. The plan is clean, the microcycle makes sense, the loads are calculated, and the week has a beautiful internal logic. Then Monday happens. The athlete slept badly, the head coach changed the session, the field is unavailable, the gym is full, and the “optimal” plan starts looking like a theoretical artifact from a more obedient universe.

That is not a failure of coaching.

That is coaching.

The failure is pretending this is an exception rather than the normal working condition. We like plans because they give us warm comfort. They make the future look controllable. But athletes do not live inside our spreadsheets. They live inside sessions, matches, relationships, emotions, fatigue, injuries, schedules, logistics, and the normal stupidity of real life.

This article is about the image above: Planning Between Structure and Reality. The diagram is my attempt to connect several mental models that, at first, may look unrelated: top-down vs. bottom-up, bias vs. variance, Place of Things vs. Forum for Action, Is vs. Ought, Apollonian vs. Dionysian, and the practical planning loop behind Agile Periodization.

The big idea is simple:

Good coaching lives in the oscillation between structure and reality.

Not top-down only. Not bottom-up only. Not science worship. Not vibes. Not rigid periodization. Not “just listen to your body” randomness. We need both lenses, but we need to know what each one is good for, where each one breaks, and how to move between them without turning either side into ideology.

The top-down lens: structure before reality

The top-down lens starts with a model of how things should work. It starts from principles, theory, categories, mechanisms, plans, progressions, and prior knowledge. This is the world of periodization models, biomotor abilities, energy systems, training zones, force-velocity profiles, critical power models, rep-max tables, tactical models, load management systems, test batteries, and all the things coaches use to make reality a little less chaotic.

This is not bad. Actually, it is necessary.

Without top-down structure, every week becomes improvisation. You walk into the gym and ask, “So, what do we feel like doing today?” Sometimes this works. Often it becomes random acts of training. You might call it flexible, but the athlete might call it confusing. The body might call it underdosed, overdosed, or just pointless.

Top-down thinking gives us orientation. It tells us what probably matters before the session starts. If you are preparing an MMA fighter, your top-down model might tell you that neck strength, grip endurance, repeat-effort ability, aerobic recovery, tissue tolerance, and skill under fatigue probably matter. If you are preparing a sprinter, your model might tell you that acceleration, maximal velocity, projection, stiffness, rhythm, hamstring tolerance, and exposure to high-speed running probably matter. If you are preparing a swimmer, your model might tell you that starts, turns, stroke mechanics, pacing, water feel, aerobic power, and race-pace tolerance probably matter.

That is useful. You do not want to start every Monday as if you have never seen sport before. You need a map. You need categories. You need some starting assumptions. You need priors.

The problem begins when we forget that these assumptions are assumptions.

Bias: why structure is not the enemy

In machine learning, bias is not simply prejudice. It refers to the error introduced by simplifying assumptions in the model. A linear model assumes that the relationship between variables is linear. That assumption is a bias. A more flexible model makes fewer simple assumptions, but it may become more sensitive to noise. This is the famous bias–variance trade-off described in statistical learning.

Coaches should steal this idea.

A training plan has bias. A periodization model has bias. A force-velocity profile has bias. A tactical model has bias. A rep-max table has bias. Even “we just coach what we see” has bias, because the coach decides what is worth seeing.

Bias is not automatically bad. Bias is what makes action possible. A subway map is biased. It does not show perfect geography. It removes buildings, trees, streets, and the exact curvature of the tracks. That is precisely why it is useful. It shows what a traveler needs: stations, lines, connections, and direction.

A good training model does the same. It removes some of reality so we can act. It says: “Out of all possible things we could pay attention to, these things probably matter most right now.”

That is useful bias.

But bias becomes dangerous when the model starts pretending to be reality. This is where coaches get trapped. We start with a model because we need orientation. Then the model becomes identity. Then we defend the model even when the athlete, the sport, and the week are clearly telling us something else.

That is when the map becomes a prison.

High-bias planning: clean, logical, and sometimes wrong

A high-bias plan has strong assumptions and low flexibility. Think about the typical weekly structure:

Monday: recovery
Tuesday: strength
Wednesday: conditioning
Thursday: speed
Friday: tactical
Saturday: match
Sunday: off

Looks good. Easy to explain. Easy to print. Easy to defend in a staff meeting. The structure gives everyone a sense that someone is in control.

But what happens when Tuesday’s tactical session already crushed the legs? What happens when the athlete is not adapting to the strength work? What happens when the problem is not lack of capacity but poor timing, poor perception, or panic under pressure? What happens when the athlete needs more exposure to the actual sport problem and less beautifully categorized “qualities”?

High-bias planning can underfit reality. In machine learning language, the model is too simple for the pattern. In coaching language, the plan is neat but the athlete is not better.

This is the danger of top-down planning. The model gives structure, but it can miss the actual problem. You end up training the category instead of the athlete.

We do not remove bias. We calibrate it.

The Place of Things: the world described

The top-down lens is closely related to what Jordan Peterson calls the Place of Things in Maps of Meaning. The Place of Things is the world described as objects, categories, mechanisms, and measurable relations. It is the world of science, anatomy, physiology, biomechanics, test scores, dashboards, variables, and models.

In sport, the Place of Things is everywhere:

VO2max. Lactate threshold. Sprint split. Jump height. Bar velocity. Tendon stiffness. Force output. Joint angle. Acute workload. Chronic workload. Heart-rate recovery. Testosterone. Cortisol. Muscle fiber type. Oxygen delivery. Torque. Impulse. Power.

Again, nothing wrong with this. The Place of Things is powerful. It lets us describe the athlete. It lets us compare. It lets us test. It lets us build models. It lets us avoid pure superstition.

If an MMA fighter gets tired in round three, the Place of Things asks useful questions. What is his aerobic capacity? What is his repeat-effort ability? What is his grip endurance? What is his neck strength? What is his weight cut like? What is his sleep? What does his training load look like? What are his blood markers? What is his body composition? What happens to his heart rate between rounds?

These are good questions.

But they are not enough.

Because describing the athlete does not automatically tell you what to do with the athlete.

The Is/Ought Gap: why data does not prescribe action

David Hume famously pointed out that we cannot move directly from what is to what ought to be done without some additional reasoning. This is usually called the Is/Ought Gap. In coaching, this gap is everywhere, and it is one of the main reasons why “evidence-based practice” can become a costume for fake certainty.

“The athlete’s groin squeeze is down 15%.”

Okay. What now?

Do we rest? Modify? Retest? Ignore it? Ask about sleep? Reduce kicking? Keep training because the fight is close? Change the warm-up? Call the physio? Panic in a group chat?

The test gives us an is. It does not automatically give us an ought.

Same with VO2max.

“The athlete’s VO2max is 58.”

Okay. Should we do more intervals? More low-intensity volume? More sport-specific conditioning? Less conditioning because the real issue is pacing? More sparring? Better nutrition? Better sleep? More technical economy? More breathing work? Or maybe nothing, because the bottleneck is not there.

The number does not decide.

The coach decides. Or more precisely: the coach decides using theory, context, values, goals, risk, timing, constraints, and judgment.

This is where many models break. They describe the world in the Place of Things, and then quietly assume that description has already solved prescription. It has not. The Is/Ought Gap is still sitting there, smiling at your spreadsheet.

The Apollonian: the beauty of order

This structured side also maps onto Nietzsche’s distinction between the Apollonian and the Dionysian, introduced in The Birth of Tragedy. Nietzsche used Apollo and Dionysus as symbolic forces in Greek tragedy. Apollo represents form, order, clarity, proportion, restraint, image, and intelligibility. Dionysus represents intoxication, music, chaos, embodiment, dissolution, passion, and the deeper forces of life that do not fit neatly into rational form.

Do not worry. I am not trying to turn your warm-up into a German philosophy seminar.

The distinction is useful because coaching has both forces.

The Apollonian coach wants structure. He wants a model. He wants clean categories. He wants a progression. He wants the annual plan. He wants the table, the taxonomy, the test battery, and the PDF with neat diagrams. He wants to understand the machine.

Good. We need him.

The shadow appears when Apollo becomes sterile. Everything is categorized, but nothing is alive. The coach understands the force-velocity profile but does not see the athlete hesitating. He understands energy systems but not fear. He sees the GPS report but not the tactical meaning of the running. He sees the body as a place of things, but not the athlete as a person acting under uncertainty.

That is not science. That is reductionism wearing a lab coat.

The bottom-up lens: reality before structure

The bottom-up lens starts from what is actually happening now. Not what should happen. Not what the model expected. Not what the annual plan promised. What is happening here, with this athlete, in this session, under these constraints?

This is the world of day-to-day variability, individual response, logistics, fatigue, pain, mood, motivation, relationships, team dynamics, opponent behavior, schedule changes, and the boring details that destroy beautiful plans.

This is where the rubber meets the road.

The bottom-up coach asks different questions. What can we actually do today? What is the athlete showing me? What is the session revealing? What changed? What is the real bottleneck? What does the sport problem demand? What does the environment allow? What is the smallest useful move?

This lens is not anti-science. It is anti-delusion. It reminds us that the athlete is not an average treatment effect. The athlete is an individual response unfolding through time.

And yes, this is messier.

Welcome to coaching.

The Forum for Action: the world lived

The bottom-up lens maps onto Peterson’s Forum for Action. If the Place of Things is the world described, the Forum for Action is the world lived: intention, threat, opportunity, meaning, behavior, uncertainty, and action.

In the Place of Things, the athlete is a bundle of variables.

In the Forum for Action, the athlete is someone trying to solve a problem.

The Place of Things says:

“High-speed running exposure was 620 meters.”

The Forum for Action asks:

“Did the player actually solve the pressing problem?”

The Place of Things says:

“Punch output dropped 18% in round three.”

The Forum for Action asks:

“Did the fighter fade because of physiology, anxiety, poor pacing, inefficient clinch work, tactical stupidity, or because the opponent made him carry weight for three minutes?”

The Place of Things says:

“The athlete is strong.”

The Forum for Action asks:

“Can he use that strength at the right time, against resistance, while perceiving the opponent, under fatigue, without overcommitting and giving up position?”

This is why I keep coming back to the same line:

Athletes do not live inside metrics. They live inside problems.

The Forum for Action is where those problems appear.

Variance: when the training floor fights back

In machine learning, variance refers to how much a model changes when trained on different samples of data. A high-variance model is very sensitive to the specific data it sees. It may fit the training data beautifully, but it can also chase noise and fail to generalize.

In coaching, variance is the training floor fighting back.

The plan says one thing. Reality gives you another.

The athlete who looked fresh yesterday is cooked today. The small-sided game produced more accelerations than expected. The sparring session became much harder than planned. The athlete’s knee is fine in the gym but not fine when cutting. The tactical drill was supposed to be moderate, but the competitive psychos turned it into the Champions League final. The spreadsheet did not predict that.

Bottom-up coaching respects variance. It says: “Let me observe what is happening, then update.”

This is the beauty of bottom-up thinking. It listens.

But it also has a shadow.

Too much bottom-up becomes reactive. Everything becomes “how do you feel today?” or “what does HRV say?” or “what does the vibe say?” If every day is decided from the current state, you may preserve flexibility, but you can lose direction. You may overreact to noise. You may mistake normal fluctuation for meaningful signal.

The athlete feels bad today. Is that a problem or normal fatigue?

The jump is down 5%. Is that meaningful or measurement noise?

The fighter lost a sparring round. Is that a tactical issue, a bad matchup, or useful learning?

Bottom-up coaching without structure becomes high variance. It may fit today and fail the month.

The bias–variance trade-off for coaches

The bias–variance trade-off is usually explained with prediction models. A very simple model may have high bias: it misses important patterns because its assumptions are too rigid. A very flexible model may have high variance: it follows the data too closely and starts chasing noise. The practical goal is not to eliminate bias or variance, but to find a useful balance for the problem.

Now translate this to coaching.

A rigid annual plan is high bias. It assumes too much. It may underfit reality.

A purely reactive plan is high variance. It adapts to everything. It may overfit noise.

Good coaching lives in the calibrated zone:

Enough structure to guide action. Enough flexibility to update when action teaches you something.

This is why I dislike the fake fight between “planned” and “flexible.” The real question is not whether you plan. You always plan, even if the plan is implicit. The real question is how much bias you impose, how quickly you detect variance, and how intelligently you update.

A good plan should have bias. It should say: “Here is what we think matters.” But it should also be exposed to variance. It should ask: “What is reality telling us back?”

This is Agile Periodization in plain language.

The No Free Lunch theorem: no method wins everywhere

The No Free Lunch theorem from machine learning and optimization is one of those ideas that coaches should understand, even if they never read the math. The rough idea is this: averaged across all possible problems, no algorithm is universally superior. If one algorithm performs better on one class of problems, that advantage comes from assumptions that match that class of problems. Under another class of problems, those same assumptions may hurt you.

This matters because coaches love universal claims.

“This is the best periodization model.”

“This is the best conditioning method.”

“This is the best way to teach skill.”

“This is the best strength template.”

No. It is not. It may be the best tool for a particular problem under particular assumptions.

Linear periodization works under some assumptions. Block periodization works under some assumptions. Tactical periodization works under some assumptions. Constraints-led practice works under some assumptions. Velocity-based training works under some assumptions. High-low planning works under some assumptions. Ecological dynamics works under some assumptions. Agile Periodization also works under some assumptions.

Tools, not truths.

No Free Lunch does not mean all methods are equal in the real world. That would be lazy relativism. It means performance depends on the match between the method and the problem class. In coaching terms, you need to ask:

Best for what?

For whom?

Under what constraints?

With what risk?

At what time?

With what downside if we are wrong?

That is a much better question than “what is the best method?”

McGilchrist: not left brain vs. right brain nonsense

Iain McGilchrist’s work gives another useful lens, but we need to avoid the Instagram version of “left brain vs. right brain.” The cartoon version says left brain equals logic and right brain equals creativity. That is too simple and mostly useless.

McGilchrist’s deeper argument is about modes of attention. In broad strokes, the left hemisphere tends to narrow, categorize, grasp, manipulate, and abstract. It is very good at tools, explicit representation, models, and parts. The right hemisphere tends to attend more broadly, contextually, relationally, and embodied. It is more open to the living whole.

This maps nicely onto coaching.

The Place of Things is more left-hemisphere-like: explicit, analytical, categorical, model-based.

The Forum for Action is more right-hemisphere-like: contextual, relational, embodied, alive to the whole situation.

Again, do not turn this into brain astrology. The point is not that one hemisphere is good and the other is bad. The point is that both modes of attention are needed, and the model-making mode should serve the living situation, not replace it.

This is exactly what goes wrong when the spreadsheet becomes the coach.

Pirsig: classical and romantic understanding

Robert Pirsig, in Zen and the Art of Motorcycle Maintenance, makes a similar distinction between classical and romantic understanding. Classical understanding sees the motorcycle as a system of parts, mechanisms, relations, and functions. Romantic understanding sees the motorcycle as experience: the ride, the wind, the beauty, the feel, the immediate quality of being there.

If you only have romantic understanding, you may enjoy the ride until the engine fails and you have no idea why. If you only have classical understanding, you may understand every part and still miss why anyone would ride in the first place.

Training is the same.

The classical coach understands physiology, biomechanics, programming, testing, and mechanisms. The romantic coach understands rhythm, confidence, meaning, timing, flow, fear, energy, and the lived quality of the session.

Good coaching needs both.

The athlete is not just a machine. But the athlete is also not a magical cloud of feelings immune to physiology.

That is the tension.

The Dionysian: the session is alive

The Dionysian side is the messy side of reality. It is the part of coaching that resists clean capture. You see it when the session suddenly comes alive, when the athlete finds rhythm, when a drill turns into real learning, when a fighter stops thinking and starts reading, when a team starts solving the game together.

You also see the dark side of it when chaos takes over. The session becomes emotional, reactive, loud, and uncontrolled. The athlete feels something and assumes it is true. The coach rejects data because “I know what I see.” Everything becomes intuition, and nobody can explain or reproduce anything.

This is the shadow side of the Forum for Action: relativism, pseudoscience, mistaking feelings for truth, rejecting data, and hiding behind “context” when you simply do not have a clear model.

The Place of Things has a shadow: scientism, reductionism, coaching spreadsheets instead of athletes.

The Forum for Action has a shadow: vibes, gurus, and “bro, trust me” coaching.

Pick your poison.

Or better: learn to move between them.

The Is/Ought Gap in actual coaching

Let’s make this less philosophical.

Suppose you test an athlete and find that their adductor squeeze is down 15% compared with their baseline. That is the Place of Things. It is an is statement.

Now what?

A rigid top-down coach may say: “Down 15%, reduce load.” Maybe that is right. Maybe not.

A purely bottom-up coach may say: “How do you feel?” Maybe the athlete says fine. Maybe they are lying. Maybe they do not know. Maybe it matters only during kicking or cutting. Maybe it is measurement noise. Maybe the test is valid in general but not meaningful today.

The coach has to bridge the gap. That bridge includes the athlete’s history, the sport context, the session goal, competition schedule, symptoms, risk tolerance, previous responses, and the cost of being wrong.

This is why evidence does not remove judgment. Evidence improves judgment. There is a difference.

Example 1: MMA conditioning

An MMA fighter fades late in rounds. The Place of Things asks: Is VO2max low? Is repeat-power ability poor? Is grip endurance poor? Is the weight cut bad? Is the athlete under-fueled? Is there enough low-intensity work? Is the athlete exposed to enough high-intensity work?

Useful questions.

The Forum for Action asks: When does he fade? After failed takedowns? Against the fence? In clinch exchanges? When pressured? When he misses big shots? When he cannot rest in positions? When he emotionally overreacts? When the opponent forces him to carry weight?

Also useful questions.

If you only use the Place of Things, you may prescribe more intervals when the real problem is tactical economy. If you only use the Forum for Action, you may talk about composure while ignoring the fact that the athlete simply lacks the engine to repeat high-quality efforts.

The synthesis is not hard to state, but it is hard to do:

Start with the fight problem. Regulate the dose. Watch what changes.

Example 2: football pressing

A football team cannot counterpress after losing the ball. The Place of Things sees accelerations, decelerations, high-speed running, player load, aerobic power, neuromuscular fatigue, and recovery status.

The Forum for Action sees poor distances before ball loss, bad body orientation, late first defender reaction, poor compactness, fear of space behind, poor role clarity, and bad timing of support.

If you only see the Place of Things, you may prescribe conditioning. If you only see the Forum for Action, you may design tactical games that underdose sprint exposure.

Again, both lenses are needed. The game problem anchors the session. Dose regulates the exposure.

Problem as anchor. Dose as regulator.

I will probably overuse that line. It deserves it.

Example 3: swimming

Swimming looks like a closed skill. There is no opponent trying to punch you, press you, or shoot for your legs. The environment is more stable, so dose-based planning is closer to the performance problem.

But even here, dose is not the whole story.

A swimmer loses stroke length in the third 50 of a 200 freestyle. The Place of Things asks about aerobic capacity, lactate tolerance, race-pace exposure, stroke rate, distance per stroke, mobility, strength, breathing pattern, and turn speed.

The Forum for Action asks what the swimmer feels. Where does rhythm break? Does breathing disrupt body line? Does the catch disappear under fatigue? Does the swimmer panic when pace rises? Does the turn destroy the next length? Is the athlete solving the water problem or just surviving the set?

A set of 10 x 200 at threshold pace may be useful. But the real question is what problem is being solved inside those 200s.

Dose is visible. The problem is where coaching lives.

Practical wisdom: the center of the model

The center of the diagram is Practical Wisdom / Agile Planning. This is not another fancy name for “experience.” Plenty of experienced coaches repeat the same mistake for twenty years. Practical wisdom is the ability to choose, switch, and update lenses based on the problem in front of you.

It is knowing when to impose structure and when to listen. It is knowing when the data matters and when it is noise. It is knowing when the athlete needs discipline and when the plan needs humility. It is knowing when to protect the downside and when to invest in upside. It is knowing when to use the map and when to look out the window.

This is not anti-science. It is science plus judgment.

Or better:

Model-informed judgment under uncertainty.

That is the job.

The iterative planning loop

At the bottom of the diagram is the practical loop:

1. Top-down model / prior

2. Act in the Forum for Action

3. Observe variance and feedback

4. Update the model

This is Agile Periodization in one loop.

You start with a prior. A prior is your best current guess. It can come from research, experience, testing, sport demands, theory, or common sense. For example: “This fighter probably needs more aerobic support,” or “This football team probably needs better counterpressing behavior,” or “This sprinter probably needs more max velocity exposure.”

Then you act.

Not by writing a sacred twelve-week plan and worshipping it, but by deploying a minimum viable plan: enough structure to start, enough flexibility to learn.

Then you observe.

Did the athlete adapt? Did the session solve the problem? Did the dose land? Did the athlete tolerate it? Did the behavior improve? Did new constraints appear?

Then you update.

That is Bayesian thinking in plain language:

Start with what usually works. Test it against what actually happens. Update the plan.

No drama. No ideology. No pretending.

Too much structure, too much chaos

The model also gives us two failure modes.

Too much top-down and you get rigidity. You underfit reality. The plan stays clean while the athlete drifts away from it. The system becomes Apollonian in the bad sense: orderly, sterile, over-controlled, detached from the living situation.

Too much bottom-up and you get noise. You overfit the day. You react to every fluctuation, chase every feeling, and call it individualization. The system becomes Dionysian in the bad sense: chaotic, emotional, unstable, impossible to reproduce.

Good coaching is not the average of these two. It is not 50% structure and 50% flexibility. That is too simplistic.

Good coaching is oscillation. Sometimes you lean into structure because the athlete needs direction. Sometimes you lean into feedback because reality is clearly telling you that the model is wrong. Sometimes you protect the downside. Sometimes you explore upside. Sometimes you impose. Sometimes you adapt.

That is the squiggle sense.

The unifying model

Here is the model in one sentence:

Top-down gives orientation. Bottom-up gives calibration. Coaching lives in the oscillation between them.

Or, slightly longer:

Use top-down models as priors, enter the Forum for Action, observe variance, bridge the Is/Ought Gap with judgment, and update the model.

This connects the whole diagram.

Top-down is bias. It gives structure.

Bottom-up is variance. It reveals reality.

Place of Things describes what is.

Forum for Action shows what needs doing.

Is gives observation.

Ought requires judgment.

Apollonian gives form and clarity.

Dionysian gives life and embodied reality.

Bias gives orientation.

Variance gives feedback.

Practical wisdom lives in the movement between them.

This is not a model that removes uncertainty. That would be bullshit. It is a model for acting more wisely inside uncertainty.

That is the best we can do.

And honestly, it is already a lot.

What should you do with this tomorrow?

Before writing the next session, ask yourself a few questions.

What is my top-down model? What do I assume matters? What theory am I using? What categories am I imposing? What does the research or prior experience suggest? What is my bias?

Then ask the bottom-up questions. What is actually happening? What does the athlete show me? What does the session reveal? What does the sport problem demand? What does the context allow? What variability am I seeing?

Then bridge the Is/Ought Gap. What do I know? What do I not know? What are the risks? What is the cost of being wrong? What is the smallest useful intervention? What should I observe next?

Then act.

Then update.

That is the loop.

Final takeaway

Your plan should have bias. It should have structure, assumptions, categories, and theory. But your plan should also be humble enough to meet variance. It should listen to the athlete, the session, the match, the opponent, the environment, and the boring logistics that ruin beautiful charts.

The goal is not to become less scientific.

The goal is to become less stupid with science.

Use the Place of Things. Respect the Forum for Action. Bridge the Is/Ought Gap with judgment. Use Apollo for structure. Use Dionysus for life. Use bias for orientation. Use variance for calibration.

And keep updating.

Because the plan is not the territory.

The plan is your opening argument.

Reality gets the reply.

Want to go deeper?

If this article resonated with you, this is exactly the type of thinking I unpack in the Agile Periodization Philosophy Course.

This course is not about giving you another template.

Templates are easy. Thinking is harder.

The course is about the mental models behind Agile Periodization: how to think about training when the world is messy, athletes are inconsistent, data is noisy, and the “perfect plan” starts falling apart the moment it touches reality.

Inside the course, I go deeper into the ideas mentioned in this article:

• Map vs. Territory

• Place of Things vs. Forum for Action

• Top-down vs. Bottom-up

• Substance ~ Form

• Robust vs. Optimal

• Explore vs. Exploit

• The Is/Ought Gap

• Bayesian updating

• MVP planning

• Barbell strategy

• Tools, not truths

The goal is not to make you memorize more concepts. The goal is to help you build better questions, safer starting points, tighter feedback loops, and wiser decisions under uncertainty.

Because this is the real bottleneck for most coaches.

Not information.

Not exercises.

Not another periodization chart.

The bottleneck is judgment.

How do you decide what matters? How do you know when to follow the plan and when to change it? How do you use science without becoming rigid? How do you respect practice without falling into vibes? How do you build systems that survive contact with real athletes?

That is what the course is about.

You can check it out here:

https://payhip.com/b/iA3qU

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

References

Geman, S., Bienenstock, E., & Doursat, R. (1992). Neural networks and the bias/variance dilemma. Neural Computation, 4(1), 1-58. https://doi.org/10.1162/neco.1992.4.1.1

Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of statistical learning: Data mining, inference, and prediction (2nd ed.). Springer. https://doi.org/10.1007/978-0-387-84858-7

Hume, D. (1739-1740). A treatise of human nature. John Noon.

James, G., Witten, D., Hastie, T., Tibshirani, R., & Taylor, J. (2023). An introduction to statistical learning: With applications in Python. Springer. https://doi.org/10.1007/978-3-031-38747-0

McGilchrist, I. (2009). The master and his emissary: The divided brain and the making of the Western world. Yale University Press.

McGilchrist, I. (2021). The matter with things: Our brains, our delusions, and the unmaking of the world. Perspectiva Press.

Mitchell, S. D. (2009). Unsimple truths: Science, complexity, and policy. University of Chicago Press.

Nietzsche, F. (1872). The birth of tragedy out of the spirit of music. E. W. Fritzsch.

Peterson, J. B. (1999). Maps of meaning: The architecture of belief. Routledge.

Pirsig, R. M. (1974). Zen and the art of motorcycle maintenance: An inquiry into values. William Morrow.

Wolpert, D. H. (1996). The lack of a priori distinctions between learning algorithms. Neural Computation, 8(7), 1341-1390. https://doi.org/10.1162/neco.1996.8.7.1341

Wolpert, D. H. (2020). What is important about the No Free Lunch theorems? arXiv. https://arxiv.org/abs/2007.10928

Wolpert, D. H., & Macready, W. G. (1997). No Free Lunch theorems for optimization. IEEE Transactions on Evolutionary Computation, 1(1), 67-82. https://doi.org/10.1109/4235.585893

Ovo nije klasičan transkript mog gostovanja kod Slobodana Maksimovića Limpa u podcastu “Prava Priča”. Razgovor je trajao dugo, otišao je u više pravaca i, kao svaki dobar razgovor, nije se držao uredne akademske strukture. Krenuli smo od MMA pripreme i famoznog “gas outovanja”, pa završili kod metodologije treninga, constraint-led pristupa, fokusera, tehnologije, fudbala, performance managera, pisanja i toga šta čovek oseća kao svoj poziv.

Zato sam ovaj tekst uredio kao dijalog. Nije doslovan prepis, nego skraćena i sređena verzija onoga što smo pričali, da može da se pročita bez gledanja dva sata videa, ali da se zadrži duh razgovora.

Limp: Ajde da krenemo od velike teme u svetu borilačkih sportova. Imali smo meč Hamzata Chimaeva i priču da je “gas outovao” posle prve runde. Posle toga su svi krenuli da napadaju njegove metode treninga. Lanci, grinder, crvena zona, ludilo. Kako ti to gledaš? Da li te metode mogu da presude meč?

Mladen: Mogu da doprinesu, ali moramo biti oprezni sa tim zaključcima. Da je pobedio, verovatno bi isti ti treninzi bili predstavljeni kao razlog pobede. Pošto je izgubio, sada su razlog poraza. To je standardna ljudska reakcija posle meča. Svi postanu proroci unazad.

Moj bivši šef Darren Burgess imao je dobru rečenicu: “Fizička priprema ne može da ti osvoji šampionat, ali može da te košta šampionata.” Mislim da je to dobar okvir. Fizička priprema verovatno neće sama napraviti šampiona, ali ako je loše uklopljena, ako pojede oporavak, ako uništi skill work i drži borca stalno u crvenom, onda može da ga košta.

Limp: Meni je delovalo kao da je pretreniran. Imao sam taj osećaj i kod sebe, kada preteraš sa teškim treninzima, posle šezdeset sekundi kao da ti nestane snaga. Radiš na glavu, ali telo ne prati. Da li je moguće da je to bilo to?

Mladen: Moguće je, ali ne bih tvrdio bez kompletne slike. Treba videti weight cut, ishranu, oporavak, san, psihološki stres, sparinge, sve. Ali generalno, ako je borac stalno u crvenom, ako radi treninge koji izgledaju brutalno, ali ga dva dana posle toga izbace iz kvalitetnog borilačkog treninga, onda imamo problem.

Svaki trening ima benefit, ali ima i cenu. To je ono što ljudi često preskaču. Nije pitanje samo “da li ovaj trening radi”. Pitanje je i “šta ovaj trening kvari”. Ako ti trening snage i kondicije pojede dva dana kvalitetnog rada na veštini, onda možda nije vredan toga.

Limp: Znači, nije problem da se radi jako, nego gde to stoji u celom sistemu?

Mladen: Upravo to. Ja volim onu analogiju sa teglom i velikim kamenovima. Ako prvo sipaš pesak i šljunak, teško ćeš posle ubaciti veliko kamenje. Ali ako prvo staviš veliko kamenje, pesak i šljunak će nekako naći mesto između.

U MMA, veliki kamenovi su specifičan trening, rešavanje borilačkih problema, sparing, metodologija obuke, kultura sale, taktika i oporavak. Sve ostalo, tegovi, intervali, Airdyne, VO2 max, HRV, laktatni prag, sve to može da bude korisno, ali to su pesak i šljunak. Problem nastaje kada pesak počne da glumi kamen.

Limp: A zašto se onda stalno gura ta fizička priprema? Otvoriš Instagram i vidiš borce kako rade specijalne vežbe, lance, grinder, maske, neke lude intervale.

Mladen: Zato što je to vidljivo. Može da se snimi. Izgleda brutalno. Može da se meri. Dobro izgleda na storiju. A ono što je stvarno bitno često se ne vidi. To je metodologija treninga.

Kako dizajniraš trening? Kako napraviš problem? Kako doziraš otpor? Kako menjaš ograničenja? Kako razvijaš percepciju? Kako napraviš kulturu gde borac sme da pogreši? Kako vodiš refleksiju posle treninga? To je santa leda ispod površine.

Na površini vidiš gedžete, znoj i brutalne klipove. Ispod površine je metodologija. A po meni je metodologija najvažnija.

Limp: Ajde konkretno. Imaš borca, pet rundi, osam nedelja. Kako bi krenuo?

Mladen: Ne bih krenuo od pitanja da li radimo 4x4 intervale, Airdyne ili trčanje. Krenuo bih od šireg plana. Imaš period van meča, kada ne znaš protivnika i razvijaš borca. Tu širiš repertoar, radiš na slabostima, izlažeš ga situacijama u kojima nije bio i pokušavaš da ga napraviš robusnijim.

Onda imaš kamp, kada znaš protivnika. Tu polako poliraš trening prema onome što će se najverovatnije pojaviti u meču. Ali tu dolazi veliko pitanje: da li se prilagođavamo protivniku ili namećemo svoju igru? Ja sam slično pitanje postavio Greggu Berhalteru dok sam radio u Hammarbyju. On mi je rekao: “Odbranu pripremamo prema protivniku, napad je naš.” Mislim da se to može lepo prevesti i na MMA.

Limp: Znači, ne menjaš sve zbog protivnika?

Mladen: Ne. Moraš znati njegove opasnosti i moraš se zaštititi od njegovih prednosti. Ali ne smeš izgubiti ono u čemu si ti dobar. Ako sve podrediš protivniku, možeš izgubiti identitet. Ako ignorišeš protivnika, možeš ući u njegove zamke. Dakle, napad gradiš oko svojih prednosti, a odbranu prilagođavaš njegovim prednostima.

Limp: A taper? To u borilačkim sportovima deluje haotično. Skidanje kilaže, povrede, posekotine, poslednji sparinzi, nervoza.

Mladen: Taper nije “ne radimo ništa”. Taper znači da pokušaš da skineš negativne efekte treninga, da ukloniš umor i da manifestuješ spremnost. Kod boraca je to još komplikovanije jer imaš weight cut, mentalni pritisak, rizik od sudara glavama, posekotina i povreda.

Pred sam meč ne želiš da ubaciš borca u frustraciju samo zato što “treba još jednom jako”. U tom periodu često biraš zadatke koji mu dižu samopouzdanje. Ne zato što su fiziološki magični, nego zato što borac treba da uđe u meč sa osećajem: “Mogu ovo da rešim.”

Limp: To mi je zanimljivo, jer često treneri hoće da testiraju borca do kraja. Kao da mora da se raspadne da bi dokazao da je spreman.

Mladen: Zato volim onu heuristiku: bolje biti dosta nepripremljen nego malo pretreniran. Zvuči čudno, ali u borilačkom sportu ima smisla. Ako si malo manje spreman, ali si svež, zdrav, oštar i sposoban da rešavaš probleme, imaš šansu. Ako si “spreman”, ali si iscrpljen, demoliran od treninga, lošeg skidanja kilaže i umora, onda ti ta spremnost ne znači mnogo.

Limp: Dobro, ako je skill glavni kamen, šta onda radi kondicioni trener? Koja je njegova uloga?

Mladen: Fizička priprema ne treba da napravi boljeg borca direktno. Bolje je reći da treba da napravi borca koji bolje podnosi MMA trening. Da se manje povređuje, da se brže oporavlja, da može češće da ulazi u specifične situacije, da mu telo ne bude limitirajući faktor.

Neke stvari ne možeš bezbedno da dobiješ samo kroz MMA. Ako hoćeš da borac jako udara, ne možeš stalno da udara partnera maksimalno. Ili ćeš povrediti partnera ili ćeš ostati bez partnera. Zato koristiš džak, fokusere, ground and pound drillove, medicinke, izometrije, dodatni rad za vrat, šaku, rame, kukove. Ali to je suplement. Nije glavno jelo.

Limp: Znači, kondicija kao podrška specifičnom treningu, a ne kao centar sveta.

Mladen: Tako je. Cilj fizičke pripreme je da borac može bolje da toleriše ono što je najvažnije. Da može češće i kvalitetnije da rešava borilačke probleme. Ako fizička priprema ometa taj proces, onda smo promašili temu.

Limp: Pričali smo i o pritisku. To je nešto što svaki borac razume, ali teško ga je objasniti naučno.

Mladen: Pritisak je dobar primer fenomena koji ne možeš lako prevesti u klasičan kondicioni jezik. Šta je pritisak? Može biti fizički, taktički, psihološki. Može biti osećaj da ne možeš da se pomeriš, da ne možeš da dišeš, da stalno kasniš, da ti neko ne dozvoljava da se repozicioniraš.

Ne možeš to rešiti leg pressom. Ne možeš biomehanički objasniti pritisak i onda ga programirati kroz izolacionu vežbu. Moraš napraviti borilački zadatak u kojem se taj pritisak pojavljuje dovoljno često, dovoljno jasno i dovoljno bezbedno da borac uči da ga rešava.

Limp: Tu dolazimo do constraint-led pristupa i ekološkog pristupa. Danas je to sve popularnije, ali često izgleda kao da ljudi odu u drugu krajnost. Kao, nema više tehnike, nema drilova, samo igre.

Mladen: Da, to je sektašenje. Problem nije u drilu. Problem je kada dril postane religija. Postoji “unopposed practice”, odnosno vežbanje bez protivnika, i ono može imati smisla. Ako neko ne zna da padne, nećeš ga baciti u haos i nadati se da će organizam sam pronaći rešenje. Moraš ga naučiti da padne.

Dril može služiti bezbednosti, razvoju osnovne pokretljivosti, pripremi zglobova, istraživanju rešenja koja sportista ne bi smeo da proba u sparingu, ili razbijanju šablona. Ali dril moraš što pre vratiti u živi problem. Ako godinu dana udaraš fokuser savršeno, ali bez distance, tajminga, protivnika i odluke, ne treba da se čudiš što to ne izlazi u borbi.

Limp: Fokuseri su posebna tema. Ljudi ih rade nenormalno puno, a opet mnoge kombinacije nikad ne izađu u borbi. Zašto se ljudi toliko hvataju za njih?

Mladen: Zato što dobro izgledaju. Ima ritma, zvuka, znoja. Trener izgleda kao maestro, borac izgleda tehnički čist. Fokuseri su performativni. Ali pitanje je čemu služe.

Mogu da budu korisni za zagrevanje, ritam, mehaniku, kondicioniranje, samopouzdanje, rad na udarcima koje ne želiš jako da bacaš u partnera, ili za simulaciju protivnika blizu meča. Ali ako fokuseri čine osamdeset posto treninga i nemaju veze sa distancom, odlukom, vremenom i otporom, onda imamo problem.

Tu opet dolazimo do sinteze. Prvo su fokuseri sveto pismo. Onda dođe reakcija i ljudi kažu da su fokuseri sranje. A prava stvar je: fokuseri su alat. Tools, not truths.

Limp: To je ista stvar kao i sa ekološkim pristupom. Prvo tradicionalni treneri rade samo tehniku i linijske drilove, onda nova ekipa kaže da je sve to glupost, a verovatno je istina negde u dizajnu i doziranju.

Mladen: Tako je. Ja nisam protiv tehnike. Ja sam protiv loše teorije iza tehnike. Tradicionalna ideja često polazi od toga da postoji idealna tehnika koju ti ponavljanjem usadiš sportisti, pa će se ona kasnije sama pojaviti u borbi. Ali borba nije koreografija. Borba je rešavanje perceptivno-kretnih problema protiv protivnika koji ne sarađuje.

Ako početnik uči džeb, nije džeb samo pokret ruke. To je distanca, tajming, pokrivanje sledećeg udarca, reakcija protivnika, fintiranje, ulazak i izlazak. Ti možeš deo toga izolovati, ali moraš ga što pre vratiti u situaciju gde postoji živi protivnik.

Limp: Šta je sa početnicima? Da li oni mogu odmah da rade igre i sparinge?

Mladen: Zavisi kako dizajniraš zadatak. Ne možeš početnika baciti u pun sparing i reći “snalazi se”. To je glupo. Ali možeš napraviti jednostavan zadatak, recimo shoulder tap, gde odmah ima protivnika, distancu, odluku i kretanje, ali je rizik mali. Poenta nije da početnik odmah ratuje. Poenta je da od početka razume da veština nije samo pokret, nego rešavanje problema.

Kod početnika moraš više da ograničiš zadatak. Kod naprednih sportista detalji su drugačiji, ali princip ostaje isti. I početnik i profesionalac moraju biti izloženi problemu, samo na različitom nivou složenosti.

Limp: Pitanje kulture sale mi je tu ključno. Kako napraviš da ljudi mogu da se igraju, da eksperimentišu, da ne ulazi ego u svaki sparing?

Mladen: Prvo zadacima, ali i selekcijom ljudi. Debili moraju biti odstranjeni ako neće da se prilagode. Ili će ih grupa sama odstraniti, jer niko neće hteti da radi sa njima. Ako nema poverenja, nema učenja.

Ako svaki sparing postane dokazivanje, onda ljudi ne eksperimentišu. Samo pokušavaju da pobede trening. A postoji trade-off između performansa i učenja. Ako stalno pokušavaš da pobediš dril, često ne učiš ništa novo. Samo hraniš ego.

Nekad moraš namerno da se hendikepiraš da bi radio na slaboj strani. Ako uvek koristiš svoje prednosti da sakriješ slabosti, možda pobeđuješ trening, ali ne napreduješ. Zato volim onu rečenicu: “Living in a cave doesn’t make you a geologist.” Samo zato što si dugo u sali, ne znači da stvarno učiš. Možda samo ponavljaš istu sezonu deset godina.

Limp: Znači, moraš napraviti prostor za grešku.

Mladen: Da. Bez greške nema učenja. Ali to mora biti bezbedna greška, ne greška gde te neko odmah kazni tako da se više nikad ne usudiš da probaš. Kod početnika je često defanziva ulaznica u igru. Ako zna da se zaštiti, da ne paniči, da ne zatvara oči i ne ulazi u survival mode, onda može da eksperimentiše. Ako nema nikakav štit, svaka greška je kazna i onda se učenje gasi.

Limp: Ajde da pređemo na tehnologiju. Danas svi imaju Whoop, HRV, readiness, sleep score, REM faze, GPS, skokove. Da li smo preterali?

Mladen: Jesmo, ali ne znači da je tehnologija beskorisna. Problem je što smo počeli više da verujemo uređaju nego sebi. Ako ti HRV padne ujutro, ne znači automatski da treba da otkažeš trening. Ako si loše spavao jednu noć, ne znači da je sistem propao.

Podaci imaju smisla kada ih gledaš kao trend, a ne kao horoskop. U reprezentaciji smo koristili wellness upitnike. Jedna igračica mi je rekla: “Napisala sam da sam loše spavala i opet sam radila trening. Čemu onda skupljamo podatke?” I to je dobro pitanje. Poenta nije da skočimo na svaki šum, nego da vidimo obrazac.

Da li se loš san javlja posle određenog tipa treninga? Da li se raspoloženje ruši posle putovanja? Da li se energija menja ponedeljkom? Da li je problem oporavak ili životni stil? Možda je neko pio kolu i jeo čips u dva ujutro. Podatak je početak razgovora, ne presuda.

Limp: Ali ljudi vole da imaju jasnu boju. Zeleno, žuto, crveno. Spreman, nespreman.

Mladen: Da, ali to ima psihološke posledice. U Port Adelaideu smo imali ideju da napravimo semafor rizika. Zeleno znači sve dobro, žuto umeren rizik, crveno visok rizik. Na papiru to deluje super. Ali sam pitao: “Da li će sportisti to videti?” Ako sportista dođe na trening i vidi da je crven, šta se dešava?

Možda počne da se čuva. Možda postane nervozan. Možda dobije alibi za loš trening. Možda stvarno poveća rizik jer sada trenira sa strahom. To je problem drugog reda. Merenje menja ono što merimo.

Limp: Znači, nije svaka informacija korisna za sportistu?

Mladen: Nije u svakom trenutku. Trener mora biti malo makijavelistički mudar. Ne u smislu da manipuliše sportistom zbog sebe, nego da zaštiti sportistu od pogrešne interpretacije podataka. Nekad nije pametno sve staviti na sto. Ljudi nisu mašine. Ako im kažeš da imaju sedamdeset posto šanse za povredu, oni neće ostati isti. Njihovo ponašanje se menja, a kad se ponašanje menja, menja se i predikcija.

Limp: Pričali smo i o povredama i tome kako je često najvažnije ono što se nije desilo.

Mladen: Da, to je via negativa. Najteže je ceniti stvari koje se nisu dogodile. Volim analogiju sa domarom u obdaništu. U jednom univerzumu izbije požar, domar utrči, čupa decu napolje, gasi vatru i svi ga slave kao heroja. U drugom univerzumu isti domar oseti da instalacije smrde, isključi struju na deset minuta, proveri problem i spreči požar. Svi mu psuju jer je prekinuo crtaće.

Prvi je heroj jer je rešio katastrofu. Drugi je sprečio katastrofu, ali niko ne zna da ju je sprečio. To je često posao kondicionog trenera. Najveći uspeh je da se sranje ne desi. Da borac ne uđe povređen u meč. Da se igrač ne povredi. Da se kamp ne raspadne. Da umor ne pojede trening. Ali to je teško prodati, jer nema spektakla.

Limp: U fudbalu je još gore, zar ne? Kad se dobije, glavni trener je genije. Kad se izgubi, kondicioni trener ih nije spremio.

Mladen: Tako je. U fudbalu se često kaže: kad se pobedi, glavni trener je postavio taktiku do jaja; kad se izgubi, kondicioni trener ih je loše spremio. Kondicioni trener često pojede govno. Njegov rad je prikriven. Ako sve ide dobro, niko ne pita. Ako nešto pukne, svi gledaju u njega.

Zato kondicioni treneri često traže načine da njihov rad bude vidljiv. Testiraju, mere, prave grafikone, uvode tehnologiju. Nekad je to korisno, a nekad je samo potreba da se pokaže: “I ja sam doprineo.” Ali dobar rad u pripremi često je nevidljiv. I to je psihološki teško.

Limp: Tu smo došli do ideje performance managera. Objasni to, jer mislim da MMA ide ka tome.

Mladen: Performance manager je osoba koja zastupa dugoročne interese kluba ili sportiste. U fudbalu bi on povezivao glavnog trenera, kondicionog trenera, medicinsko osoblje, analitiku, fizioterapeute, nutricionistu i oporavak. On nije tu da bude glavni trener, nego da drži sistem zajedno.

U MMA-u to tek treba da se razvije. Borci često imaju haos. Jedan trener ovde, kondicioni tamo, fizio negde treće, nutricionista online, kamp u drugoj zemlji, savet od lika sa Instagrama. Neko mora da drži mapu. Ne da kontroliše sve, nego da vidi da borac ne zastrani.

Limp: Znači, borac može da ide okolo, da menja okruženja, ali treba da ima kuću?

Mladen: Upravo to. Promena okruženja može biti korisna. Novi partneri, novi problemi, novi pritisak, novi pogled. Ali borac treba da ima kuću. Treba da ima nekoga ko ga poznaje, ko vidi širu sliku i ko može da kaže: “Ovo ti treba” ili “Ne moj to da radiš.”

U fudbalu to već postoji, makar u razvijenijim sistemima. U MMA-u će doći kako bude više novca, sponzora i profesionalizma. Pitanje je samo da li će se napraviti mudro ili će postati još jedan sloj birokratije.

Limp: Ali tu dolazimo i do novca. U MMA-u ljudi treniraju iz ideala, iz dokazivanja, ne samo zbog para. Ali kako novac dolazi, menja se dinamika.

Mladen: Da. MMA se profesionalizuje i moraće da se profesionalizuje. Ugovori, menadžment, klubovi, agenti, sponzori, sve to dolazi. Ali tu je pitanje motivacije. Šta treneru znači pobeda? Da li radi jer želi da pomogne čoveku da postane bolji borac, ili želi da napravi UFC borca, uzme procenat i pojavi se na podcastima?

Nije to jednostavno pitanje. Ljudi imaju različite motive. Ali ako si iskren u radu sa borcem, gubitak te pogodi. Ne može da te ne pogodi. I treneri imaju svoje demone. Sportista izgubi, a ti se pitaš šta si mogao drugačije. To je deo posla.

Limp: Ti si rekao da se ne vezuješ direktno, ali te pogodi.

Mladen: Da. Pokušavam da ne gradim identitet na rezultatu sportiste, jer to nije potpuno pod mojom kontrolom. Malo stoički zvuči, ali fokusiraš se na ono na šta možeš da utičeš. Razvoj, metodologija, trening, odluke, proces. Ali naravno da te pogodi. Ako si iskren, mora da te pogodi.

Limp: Ajde malo o tvom putu. Ljudi te znaju kao čoveka iz sporta, ali ti si krenuo iz programiranja?

Mladen: Da. Kao klinac nisam bio neki veliki sportista u klasičnom smislu. Bio sam dobar u programiranju. C, C++, assembler. Napisao sam operativni sistem, virus, išao na takmičenja. Ali borilački sportovi su me uvek privlačili. Kao i mnoge iz generacije osamdesetih i devedesetih, Van Damme, Bloodsport, Kickboxer i sve to.

U Puli sam trenirao brazilsku jiu-jitsu kod Mike Benčića. Kada sam došao u Beograd na DIF, hteo sam da probam sve što nisam mogao u manjem gradu: aikido, ninjutsu, rusku Sistemu, razne stvari. Dođeš iz malog grada i povuku te svetla velegrada.

Limp: Ruska Sistema ti je bila zanimljiva metodološki?

Mladen: Da, jer je na neki način bila constraint-led pre constraint-led pristupa. Nema fiksnih tehnika u klasičnom smislu, nego principi, problemi, mekoća, snalaženje. Stave te u problem i rešavaš. Kasnije sam čitao Bernštajna, Keith Davidsa, Matta Thorntona, Straight Blast Gym, aliveness, ecological approach, skill acquisition. Ali zanimljivo je da su neke stvari postojale intuitivno mnogo ranije.

Limp: A onda fudbal.

Mladen: Da. Fudbal me nikad nije vukao kao navijača. Jedan igrač u Hammarbyju mi je čak rekao: “Mladene, ti ne voliš fudbal. Provalio sam te.” Možda je bio u pravu. Ali fudbal mi je bio fascinantan kao metodološki problem.

Igre na skraćenom prostoru, taktička periodizacija, dizajn zadataka, ograničenja, emergence. Sama igra je prelepa. Ali sve oko igre zna da bude gadno: politika, pare, kriminal, menadžeri, roditelji, interesi. Igra je metodološki predivna, ali ekosistem oko nje može da te ubije u pojam.

Limp: Zanimljivo je što te ljudi verovatno doživljavaju kao data guy, ekran, merenje, analitika, a ti zapravo imaš mnogo iracionalniji pogled na trening.

Mladen: Da, to je pogrešna slika. Naravno da znam analitiku i živim od alata, statistike i monitoringa. Ali mislim da treba balansirati apolonijsko i dionizijsko. Ne može sve biti planirano. Život je iracionalan. Trening je iracionalan. Ljudi su iracionalni.

Da je sve poznato unapred, trenerski posao bi bio dosadan. Samo bi izvršavali zadatke. Ali uvek se nešto pojavi. Neko se povredi, neko loše spava, neko se posvađa, neko poludi, neko procveta. Trener mora biti uronjen u problem i rešavati ga na licu mesta.

Limp: I onda dolazimo do pisanja. Rekao si da je to možda tvoj poziv.

Mladen: Da. Sve više mislim da je moj “bliss” pisanje, edukacija, kreiranje alata i sadržaja. Volim da radim sa trenerima. Volim da pišem. Volim da pravim teorijske i praktične okvire. Volim da napravim nešto što će drugom treneru pomoći da bolje misli.

Mogu da radim direktno sa sportistima, ali me to dugoročno troši. Pisanje je drugačije. Kada pišem, ne pišem samo ono što znam. Dok pišem, učim. Kreneš sa jednom idejom, napadneš je, vidiš rupe, odeš u slepu ulicu, vratiš se, napraviš bolju verziju.

Limp: Znači, knjiga je rezultat slepih ulica koje se ne vide.

Mladen: Upravo to. U knjizi ne vidiš sve stranputice. Ne vidiš sve pogrešne verzije. Ne vidiš sve ideje koje su otpale. Ali one su deo procesa. Ja često pišem dijalektički. Napišem ideju, onda je napadnem. Ako preživi, možda vredi.

Tako sam i došao do zaključka da je metodologija broj jedan. Kreneš od fizičke pripreme: intervali, VO2 max, laktat, Airdyne, tegovi. Onda napadneš tu ideju. Ako borac dobija batine, da li mu stvarno fali 4x4 interval? Ako ne rešava pritisak, da li mu treba još jedan grinder? Ako ne vidi problem, da li mu treba bolji HRV?

I onda se vratiš na osnovu: kako dizajniramo trening da borac bolje rešava probleme iz borbe?

Limp: To je možda glavna poruka celog razgovora.

Mladen: Jeste. Fizička priprema je važna, ali nije centar univerzuma. Tehnologija je korisna, ali mora biti sluga, ne gospodar. Drilovi nisu mrtvi, ali ne smeju postati religija. Constraint-led pristup nije sekta, nego način razmišljanja. Sparing nije samo tuča, nego dizajniran problem.

Kondicioni trener nije tu da pokaže najbrutalniji trening, nego da borac može bolje i češće da radi ono što je stvarno važno. A ono što je važno u MMA nije da izgledaš spremno na Instagramu. Važno je da u borbi rešavaš probleme pod pritiskom, protiv čoveka koji ne sarađuje, dok si umoran, dok te boli i dok plan puca.

Limp: Znači, prvo problem, onda doza?

Mladen: Da. Nemoj prvo pitati: “Koju vežbu da dam?” Pitaj: “Koji problem pokušavam da napravim, pojačam ili rešim?” To je razlika između treninga kao liste metoda i treninga kao dizajna.

Metode su bitne, ali metode bez problema su dekoracija. Prvo problem. Onda doza. Onda alat. Onda monitoring. Tim redom.

Veliki kamenovi prvo. Pesak posle.

If you enjoyed this post, join me inside Agile Periodization Skool. Members get access to PDFs, books and book drafts, tools, courses, live Q&As, Agile Coffee, and deeper conversations about training, physiology, planning, and coaching practice with coaches, practitioners, and athletes who take this work seriously.

Agile Periodization Skool: https://www.skool.com/agileperiodization

Agile Periodization Website: https://agileperiodization.online/

Two coaches walk into a gym.

Not a joke. More like a small civil war.

On one side, we have Philip, the Dose Coach. He believes good training starts with the right exposure: sets, reps, meters, zones, intensity, volume, density, progression. He is not wrong. Without dose, training becomes vibes with cones.

On the other side, we have Duncan, the Problemming Coach. He believes good training should start with the problem the athlete needs to solve. Not “today is strength day,” but “what is the athlete failing to solve in the sport, under pressure, against constraints?”

They both care about performance.

They both hate bullshit.

They just disagree about where the plan should start.

Duncan: Let me start with the uncomfortable question. When you design training, what is the basic unit of your plan?

Philip: Easy. It depends on the phase, the sport, the athlete, and the goal. Strength. Speed. Aerobic power. Anaerobic capacity. Technical work. Tactical work. Recovery. You know, the usual stuff.

Duncan: Exactly. The usual stuff. That is my beef.

Philip: Your beef is with strength and speed now? Wonderful. The ecological guys finally went too far.

Duncan: No, my beef is not with strength, speed, endurance, biomechanics, physiology, or load. I like these things. I use them. I test them. I have spreadsheets full of them, which probably tells you I need therapy. My beef is with the fact that these things often become THE planning unit. We start with the dose, the capacity, or the adaptation we want to target. Then we try to dress it up as sport.

Philip: That is a strong accusation. Also slightly rude. I like it. Continue.

Duncan: Take the classical periodization tradition. Tudor Bompa popularized periodization as a structured way to organize training over time, using macrocycles, mesocycles, and microcycles, with different phases targeting different qualities. In that world, we talk about biomotor abilities: strength, speed, endurance, flexibility, coordination, and their combinations. The plan is usually built around developing these qualities in a logical sequence.

Philip: And that is bad?

Duncan: Not bad. Useful. But incomplete. The hidden assumption is that sport performance can be prepared by developing the right substances first, and then applying them later. Build the engine. Build the chassis. Build the muscle. Build the aerobic base. Build power. Then, somehow, the athlete will express it in the game.

Philip: Somehow? Come on. You are caricaturing it. Good coaches always connect physical qualities to the sport.

Duncan: Good coaches do. Systems often do not. Or they say they do, but when Monday comes, the spreadsheet says “strength day.” Tuesday says “aerobic power.” Wednesday says “speed.” The game problem becomes the delivery vehicle for a preselected physiological target. We start with the substance and then look for a sport-looking wrapper.

Dose, capacity, problem

At this point, the editor interrupts.

Because before this becomes another philosophical cage fight between two coaches who both need coffee, we need a few definitions.

By dose, we mean the quantity and character of exposure we prescribe: sets, reps, distance, time, intensity, density, heart rate, lactate, accelerations, high-speed running, number of punches, number of shots, number of takedown attempts, sparring rounds, and so on. Dose is the measurable exposure.

By capacity, we mean the quality or resource we want to improve: maximal strength, aerobic power, repeat sprint ability, rate of force development, mobility, tissue tolerance, buffering capacity, and so on. Dose is what we give. Capacity is what we hope changes.

By problem, we mean the performance situation the athlete must solve. Not a vague “improve performance” slogan, but a specific action challenge under constraints.

Can this boxer exit safely after throwing the right hand when the opponent pressures forward?

Can this swimmer maintain stroke length when fatigue changes the water feel?

Can this midfielder receive under pressure with the body orientation that allows the next action?

Can this MMA fighter stand up from half guard against a heavy top player without giving the back?

A problem includes what the athlete perceives, what options are available, what action is invited, what the opponent does, what the environment allows, and what the athlete can actually do.

Now back to our two friendly enemies.

Philip: So problems are sport situations.

Duncan: Yes, but not only situations. They are situations with demands, information, consequences, and constraints. A problem includes what the athlete perceives, what options are available, what action is invited, what the opponent does, what the environment allows, and what the athlete can actually do.

Philip: That sounds elegant. But here is my first attack. Problems are not independent of capacities. A fighter cannot solve the “stand up from bottom” problem if they have no hip strength, grip endurance, trunk stiffness, or neck tolerance. A footballer cannot exploit space behind the line if they cannot sprint. A swimmer cannot maintain technique if they lack the physical capacity to resist fatigue. So your “problem-first” language risks becoming hand-wavy. The athlete solves problems with a body.

Duncan: Correct. This is the best criticism of my position. And it is why I do not want to replace dose worship with problem worship. I want a better hierarchy. My current synthesis is this: problem is the anchor, dose is the regulator.

Philip: That is better. I can work with that.

Duncan: The problem tells us what the training is for. Dose helps us manage exposure, progression, fatigue, and adaptation. Dose matters, but it should not become the king. It should serve the problem.

Philip: Give me an example.

The heavy bag is not the enemy

Duncan: Let’s use boxing. A dose coach says: “Today we need anaerobic conditioning. Let’s do 10 x 30-second all-out heavy bag punch-outs with 30 seconds rest.” That might be useful. It is simple, hard, measurable, and everyone feels like Ivan Drago by the end.

Philip: Nothing wrong with that.

Duncan: Nothing wrong if the problem is physical output tolerance under fatigue. But what if the real problem is that the boxer throws long combinations and gets countered because they admire their own work? Then the issue is not just punch volume. The problem is perception-action coupling after attack. Can they punch, read the return, exit, frame, clinch, angle, or counter the counter?

Philip: So your session changes.

Duncan: Exactly. I might still use a hard conditioning element, but now the task has a problem. For example: boxer A attacks with a two- or three-punch combination; boxer B has one live counter option; boxer A must finish with an exit, frame, or re-attack depending on what they perceive. We can dose the rounds. We can manipulate space. We can manipulate fatigue. But the anchor is not “anaerobic lactic tolerance.” The anchor is “solve the post-attack danger problem under pressure.”

Philip: But if you make everything problem-based, don’t you lose control of the physical stimulus?

Duncan: Sometimes. This is where dose fights back. Representative tasks can be messy. Small-sided games may not expose players to enough high-speed running. Live wrestling rounds may overload the neck, fingers, and nervous system before you get the technical learning you wanted. Sparring may create too much chaos for a novice. A swimmer doing only “authentic swimming problems” may not accumulate enough targeted exposure to a specific stroke constraint. So yes, the problem-first coach needs dose intelligence.

Philip: Good. Because a bad constraints-led coach can become just as ideological as a bad periodization coach.

Duncan: Exactly. That is why I want complementarity, not revolution cosplay.

Tactical Periodization: game model or dose in a tactical jacket?

Now the conversation moves toward football.

This is where things get interesting, because Tactical Periodization already tried to move beyond generic physical preparation. Instead of starting with isolated biomotor abilities, it starts with the game model - the way the team wants to play.

That is a big step forward.

But does it fully escape the dose logic?

Not completely.

Philip: Let’s talk about Tactical Periodization. You always poke that bear. Tactical Periodization was developed in football as a methodology where the game model is central. It says the tactical dimension organizes everything. Physical, technical, psychological, and strategic elements are trained through the game idea. That seems close to your problem-first view.

Duncan: It is a big step forward. I respect it. Starting from the game model is much better than starting from generic biomotor abilities. Instead of asking, “How do we train endurance?” Tactical Periodization asks, “How do we play?” That is a major shift.

Philip: So what is the problem?

Duncan: The problem is that, in practice, the microcycle often still gets described as strength day, endurance day, speed day. The exercises are game-based, yes. The spaces, rules, player numbers, and tactical intentions are manipulated, yes. But the weekly logic often keeps a residual capacity structure underneath. It becomes: today we train strength through small spaces and high accelerations; tomorrow endurance through larger tactical games; then speed with more space and recovery.

Philip: But isn’t that the whole beauty? The physical quality is trained through tactical content.

Duncan: Yes. That is the good part. My critique is more precise: Tactical Periodization is not pure problem-first planning. It is a hybrid. It uses the game model as a top-level organizer, but the practical weekly rhythm often keeps strength, endurance, and speed as regulatory categories. That is not a crime. It may be smart. But we should not pretend the capacity logic disappeared. It put on a tactical jacket.

Philip: Nice line. Maybe too nice.

Duncan: I will probably overuse it on a slide.

Closed skills: when dose is closer to the problem

Not all sports are equally messy.

This matters.

In closed or unopposed skills, the environment is more stable. Nobody is trying to deceive you. Nobody shoots for your legs when you breathe badly. Nobody presses you from your blind side. The performance problem is still real, but it is more predictable.

Think swimming, sprinting, rowing, weightlifting, archery, or a cycling time trial.

In these sports, the dose perspective often has more power because the performance demands are more stable and repeatable. But even here, dose is not the same thing as the problem.

Philip: Let’s talk about open and closed skills. Because this matters. In a closed or unopposed skill, like swimming a 100-meter freestyle, sprinting, weightlifting, or a rowing erg test, the environment is more stable. There is no opponent trying to deceive you. The performance problem is still real, but it is more predictable.

Duncan: Yes. And that is why dose-based planning has more power there. If you are preparing a swimmer, the problems include start, breakout, stroke mechanics, pacing, turns, fatigue resistance, and race execution. But the world is not changing every second because another human is trying to manipulate your perception.

Philip: So in swimming, a coach can more reasonably say: “We need to develop aerobic base, race-pace tolerance, power, stroke efficiency, and lactate tolerance.” The dose maps more cleanly to the performance problem.

Duncan: More cleanly, but not perfectly. Andrew Sheaff’s Constraints-Led Approach to Swim Coaching is useful here because it reminds us that even swimming is not just physiology plus technique correction. The swimmer is solving movement problems in water. Water is an environment with affordances. Stroke mechanics emerge from the swimmer interacting with task and environmental constraints. If you only prescribe meters and intensities, you may miss the problem the swimmer is actually solving.

Philip: Give me a swimming example.

Duncan: A dose coach might say: “We need threshold work. Do 10 x 200 meters at CSS pace.” Fine. Useful. But a problem coach asks: “What is breaking down at threshold pace?” Maybe the swimmer loses catch quality when breathing to one side. Maybe stroke rate rises while distance per stroke collapses. Maybe the turn destroys rhythm. Maybe the swimmer can hit the split but cannot manage the transition from breakout to clean swimming. The dose is 10 x 200. The problem is what happens inside those 200s.

Philip: So in closed skills, dose might be closer to the problem, but it is still not identical to the problem.

Duncan: Exactly. In sprinting, “max velocity day” is a useful category. But the problem might be: the athlete cannot project out of acceleration without popping up; or they overstride at top speed; or they lose rhythm under pressure; or the hamstring history changes their willingness to attack the ground. If we only say “speed day,” we are naming the capacity, not the action problem.