Every chart on TradeIntel now shows what the options market expects a stock to do next.

Turn on Show Implied Move on the chart toolbar and a cone appears. The top edge is how far the market is pricing the stock to rise, the bottom edge how far it could fall, each priced from the actual option chain.

Until now, seeing a stock’s expected move meant leaving the chart, opening another tool, finding the right expiration, doing the math, and coming back. Now it is on the chart itself.

Each dot on the cone is a Friday expiration. Hover over one and you get that week’s numbers. The implied move panel lists all the expirations at once.

• The cone stays pinned to the last candle. Zoom in and it stretches with the bars, pan back through history and it gets out of the way.

If VIX is at 18, options are pricing about 18% annualized movement in the S&P 500 over the next month. If the index has only been moving 12%, those options look expensive versus recent history. The 6-point gap between the two numbers is the VRP.

The gap is positive most of the time. Options usually cost more than the move that actually shows up afterward. That is why selling premium works often enough.

How our VRP tool calculates it

You only need two inputs: VIX and the S&P 500 close.

First it measures how much the index actually moved in the last 6 months by taking its last 120 daily moves.

Then it takes the standard deviation of those moves and annualizes it:

RV120 = standard deviation of the daily log returns × √252 × 100

The √252 converts a daily figure into an annual one because there are about 252 trading days in a year, and volatility scales with the square root of time. Multiply by 100 so the number is a percentage you can compare directly with VIX.

Then subtracts:

VRP = VIX − RV120

A VRP of 6 means nothing on its own. The tool ranks today’s reading against every past VRP in the history. If it beats 850 out of 1,000 prior days, today sits at the 85th percentile. That percentile is what the tool uses for the match.

What the VRP data shows since February 2007

Across 4,890 sessions since February 2007, VIX typically moves in a 7.77 point range over the next month. That is the baseline everything is measured against.

A high VRP reading is followed by a range about 44% wider than normal, and about 64% wider than the range that follows a low reading.

Those are moves in either direction.

Why we check our VRP Tool against the VIX level

Almost every volatility signal rises when VIX rises.

If you want to know if a tool has any edge, the test that matters is to see if it provides any predictive power when VIX is at the same level.

We took days where VIX was in the same range, then compared high-VRP days against all days in that range.

The gap is positive in all four VIX buckets, and the gap grows as VIX rises. That is what separates a real signal from one that is just VIX wearing a different name.

Our VRP tool

The tool runs this calculation every day. It shows today’s VRP, where that reading sits in its own history, and then finds the past days that looked the same: a similar VIX level and a similar VRP percentile.

From those matched days it shows what VIX went on to do over the next 5, 10 and 20 trading days.

There is also a log of every matched day underneath, so you can study the dates and readings behind the numbers.

For the past 6 months we built our charts on top of Tradingview. That era is over. Our charts are now fully built in-house and rendered using our own engine. We fully own and control everything now.

Open Smart Charts

Live Now with New features:

• Auto-pattern recognition

• Ready-made Setups

• Full Backtest Panel

• Drawing Tools

• 60+ Indicators

New chart-related improvements and updates will accelerate from here.

Open Smart Charts

September is the worst month of the year for 8 of the 10: Consumer Staples, Consumer Discretionary, Energy, Health Care, Industrials, Technology, Materials and Real Estate.

So we backtested 80 large-cap names against their own history. Turns out 69 of the 80 names we checked underperform their own average by a median of 1.74 points.

September is the worst month for SPX

Let’s start with the index. SPX averages −0.59% in September against +0.62% for an average month, which makes September its worst month. The S&P 500 only closes higher 47% of the time in September.

Individual names diverge sharply

The difference between individual names is much larger. AEHR averages +7.43% in September versus +3.38% in a typical month. AMZN and NKE also have better than average months in September, and it’s NKE’s strongest month of the year.

On the other hand, CELH averages −7.84% in September against +3.35% in a typical month and has closed lower in each of its last four Septembers.

AEHR and AMZN rank among the 6 strongest September performers, while CELH, AAPL and AMD sit among the 6 weakest.

All 80 names we checked, sorted by gap size

September is the worst month of the year for 25 of the 80: SPX, LLY, PEP, IWM, CAT, INTC, KO, QQQ, HD, TSM, HAL, SOXX, DIS, JPM, HON, FTNT, SMH, V, BAC, MS, MRVL, AMAT, ASML, AAPL and CELH. For 48 of them it lands in the bottom 3.

Measured against its own average month, September is weaker for 69 of the 80 names we tested, by a median of 1.74 points. The S&P 500 itself averages just −0.59% in September versus +0.62% in a typical month.

The gap between individual stocks is large. AEHR runs 4.04 points above its own average in September. CELH runs 11.18 points below.

We pulled up every occurrence where SOXX RSI went below 30 since SOXS’s inception date of March 2010. There were 25 of them.

The trade wins 80% of the time. Not one of them would have blown up your account.

A median gain of 14.3% in 20 days, winning 4 times out of 5. On its own that reads like a good trade.

Does it beat shorting on any random day?

SOXS is a decaying instrument. It’s designed to go lower over time regardless of where the RSI on SOXX is currently at.

So the real question is not whether the trade wins. It is whether the trade beats shorting on a random day.

Shorting when SOXX RSI is below 30 has a better median return at every timeframe.

The win rate is also better, except at 10 days.

At the 20-day hold, the 20 winners made +453% in total while the 5 losers lost 33%.

None of the 25 trades would have blown up your account

At 20 days, not one trade went more than 100% against you. The worst drawdown was 47.3%, in February 2020, and it still closed green after 20 days.

Why it works

It works because SOXS cannot run away from you the way a volatility ETF can. For the short to lose 100%, SOXX has to fall roughly 33% while you hold it. The worst 20-day fall in the whole history is 34%, in February 2020, and by the time RSI had dropped under 30 most of that decline was already behind you.

We pulled up every occurrence where VIX RSI went above 70 since UVXY’s inception date of October 2011. There were 51 of them.

The trade wins 75% of the time. It also takes the account to zero.

A median gain of 20.7% in 20 days, winning 3 times out of 4. On its own that reads like a good trade.

Does it beat shorting on any random day?

UVXY is a decaying instrument. It’s designed to go lower over time regardless of where the RSI on VIX is currently at.

So the real question is not whether the trade wins. It is whether the trade beats shorting on a random day.

Shorting when VIX RSI is above 70 has a better median return at every timeframe.

The win rate is also slightly better.

But 8 of the 51 trades would have blown up your account

At 20 days, 8 trades went more than 100% against you at some point, which for a short position means liquidation.

Those 8 trades lost 595% while the other 43 made 943%.

Even though 2 of the trades would’ve ended up profitable, it doesn’t matter because you would never have been able to hold long enough to win, so those aren’t included in our win rate calculations.

Why it fails

It fails because a high RSI reading on the VIX doesn’t necessarily mean the VIX is elevated. The median VIX reading when its RSI is above 70 is only 24.6, which is elevated but nowhere near a panic.

Median peak for the week: calls +80%, puts +44%. The best single contract reached +1,188%.

↓ Download all the trades we scanned.

Transparency

• Every trade idea was stamped in real time on the dashboard the moment it fires.

• Targets are based on daily highs, showing how often price reached each profit level. So if you had a resting limit order at +50%, this tells you how often it would have filled.

• This page includes only contracts that expired during this week. Ideas published this week with a later expiration date will appear in the week they settle, not here.

In general, NVDA gaps up after earnings 61.8% of the time (68 times out of 110 earnings reports since May 1999) so this wasn’t unexpected.

The size of the gap was about average. It gapped up of 6.16% vs its average gap up of +5.67%.

Forward returns - After Any Gap Up

After any gap up, NVDA is higher at every time frame, ending higher by a median return of 9.18% after 3 months.

The 12 comparable gaps had a median return of 14.33% higher by the next earnings.

After a gap up in a low VIX environment NVDA typically moves higher heading into the next earnings.

Natural gas is one of the least suitable markets for this type of signal. It is the most mean-reverting of the major commodities and is known as the widow maker for the speed and violence of its reversals.

We examined every 50/200-day cross on Henry Hub natural gas futures since 2000, which included 21 golden crosses and 21 death crosses, and measured price behavior before and after each one. The signals tended to fire late and frequently pointed in the wrong direction.

The golden cross often marks the top

A golden cross is intended to signal a shift from a downtrend to an uptrend.

On natural gas it has done the opposite. In the month after the 21 golden crosses since 2000, prices fell a median 5.4%and closed higher only 29% of the time.

Because golden crosses are a lagging signal, it typically confirms an uptrend only after a substantial rally has already occurred. Mean reversion then takes hold, pulling prices lower.

The death cross often marks the bottom

If the golden cross tends to mark the top, the death cross has often marked the bottom. After the 21 death crosses since 2000, natural gas was roughly flat over the following month, with a median return of −0.5%.

Actually, several of the largest rallies began immediately after a death cross. In August 2024 a death cross at $1.96 was followed by a 105% rally over the next six months. In February 2022 a death cross preceded a 99% surge in sixty days.

The crosses mark the seasons

If the crosses do not signal direction, what do they signal? The seasons. They are not scattered randomly through the year. Golden crosses cluster in October, November, and December as prices climb into winter heating demand. Death crosses cluster in July and August, the shoulder-season lows. The 50/200-day cross on natural gas is, more than anything, a slow seasonal clock.

A trend signal needs a trending market. Natural gas is the opposite of one. The golden cross marks the top, the death cross marks the bottom, and the only thing the pair reliably marks is the turn of the seasons.

Every session it resets, targeting −3× the semiconductor index’s move that day. Hold it past the close and you are no longer getting what the name implies.

The −3× applies only to a single close-to-close period. Over a week, a month, or a quarter the return you actually receive depends on the path the index took to get there, not the net move.

Path 1: the index drops 50% in one day

The math here is simple. The index returns −50%. SOXS returns −3 × −50%, so +150%. A $10,000 long position returns $25,000.

There is no compounding, because it all happened on one day.

Path 2: the index drops 5% a day for 10 days

In this situation, the index does not fall 50%. Losing 5% per day for 10 sessions compounds to 0.9510, which is −40.13%.

SOXS gains 15% on each of those days. That compounds to 1.1510, or +304.6%.

A $10,000 long position becomes $40,456. The index fell 40.13% and SOXS gained 304.6%. The realized multiple on the cumulative index move is 7.59×, not 3×.

If that same −40.13% had occurred in a single session instead, SOXS would have returned 3 × 40.13%, or +120.4%. The long position would be worth $22,038.

The index ended up at the same number, and the SOXS return differed by $18,418, purely from the path of the drop.

Same destination, completely different payoffs

To make the comparison clean, force both paths to end with the semiconductor index down exactly 50%.

A one-day crash of −50% produces a +150% day in SOXS. A steady −6.697% every day for 10 sessions also leaves the index at −50%. SOXS compounds to +523.8%.

Same net move in the index. One path pays 3.5 times more than the other.

The daily reset compounds the position

Every close, SOXS resets its exposure to −3× its current net assets. That single rule drives the entire difference.

Start with $100 of NAV. The ETF is short $300 of the index. The index drops 6.697%. The short position gains $20.09, so NAV rises to $120.09. To stay at −3×, the ETF must now be short $360.27. It adds to the short at the close.

The next day’s drop is applied to a larger book. Day three is applied to a still larger book. In a one-directional decline, profits are automatically reinvested into more short exposure every afternoon. The position grows geometrically with NAV.

The one-day crash never gets the chance. It earns 3× on the original exposure and stops.

The same rule works against you

The daily reset does not care which direction the index moves.

An index that rallies 50% in a single session would produce a −150% day in SOXS, which is terminal. An index that rallies 50% over 10 sessions at +4.138% per day takes SOXS down 73.4%. The ETF survives, but it now needs a +276%move just to get back to even.

When the index goes nowhere and chops sideways, SOXS still loses. 10 index sessions that alternate −5.00% and +5.263% leave the index exactly flat. SOXS ends those 10 sessions down 14.8%. Nothing happened to semiconductors and nearly 15% of the SOXS position is now gone.

Note:

The one-day 50% crash is a thought experiment. Circuit breakers halt the market at −7%, −13% and −20%.

The chart now does the work you used to do by hand. It draws the pattern, labels it, and puts the details in a panel on the bottom right: how many sessions the structure has been building, the trigger level, and how far price is from that trigger.

In this walkthrough I go over 4 tickers:

• SPY — bull flag, with session count and distance to trigger

• Gold — 63-session range, 16.9% wide, $424 top / $363 bottom, breakout candle circled

• McDonald’s — falling wedge

• Semiconductors — breakdown from a 12-session range

We put that theory to the test. We backtested buying at RSI 30 on the S&P 500 and ten of the largest individual stocks and measured whether buying at RSI 30 beat buying on a random day.

On the index, oversold is a real edge

On the S&P 500, buying the RSI-30 dip returned a median 2.9% over the next month and closed higher 73% of the time, a clear edge from the 1.4% you would have made buying on a random day.

But on single stocks the edge disappears

Three names rewarded the dip-buyer (TSLA, GOOGL, META). Four punished them (AAPL, NFLX, AMZN, NVDA). Pooling the eight together, buying at RSI 30 returned a median 1.5% over the next month, short of the roughly 2.4% earned by buying those same names on a random day.

And then there is NVDA. The single best-performing mega-cap of the era, the one a dip buyer would most want to own, is where buying at RSI 30 produced the worst results: a median loss of 6.2% over the next month, higher only 36% of the time. Those oversold readings were not bargains. They sat in the middle of its sharpest drawdowns, and buying the signal meant catching the falling knife over and over.

Why the index bounces and a stock often does not

On the S&P 500, an oversold reading aggregates 500 companies, and it takes broad, indiscriminate selling to drag the whole basket down. That kind of indiscriminate fear tends to overshoot and then reverse. On a single stock the same reading usually signals a real problem with that one company: an earnings miss, a guidance cut, or a broken thesis. The market is simply repricing the problem, which is why the decline is often justified and continues.

On an index, RSI 30 is broad fear that fades. On a single stock, it is just as often the market correctly pricing in something real.

My approach is different. The overnight move is useful because it ranks who just showed real strength and who just showed real weakness. Those names often stay that way for the next quarter. On the strong ones you can sell puts or buy calls. On the weak ones you can sell calls or buy puts.

In this short walkthrough, I open the Earnings Analyzer and show you how to find the biggest gap ups and biggest gap downs.

The reason is straightforward. Since Mar 2024 the share price rose 82 percent while trailing sales only rose 28 percent.

The earnings multiple shows the same pattern.

A stock can look inexpensive on sales yet expensive on earnings when margins are shrinking. That is not the case here. Earnings per share rose 35 percent, faster than the 28 percent in sales. The share price rose 82 percent over the same stretch, which moved the earnings multiple from 26.19× to 35.39×.

How today’s multiples rank against AAPL’s own history

The table below shows where the current price-to-sales and price-to-earnings multiples rank against AAPL’s own history over different look-back windows.

16–23% Downside to Historical Multiples

From here, AAPL would need to drop 23% to reach the middle of its own valuation range. Its median price-to-sales multiple over the past five years is 7.57 times. Applied to today’s trailing sales of $31.65 per share, that points to a price of $239.62.

The gap is narrower on the earnings side. The five-year median multiple of 29.66 times, applied to trailing earnings of $8.74 per share, implies a price of $259.21, or 16% lower than today’s level.

Note

“Expensive” in this analysis means expensive only relative to AAPL’s own historical range. It does not speak to valuation versus peers or to the intrinsic value of the business. A multiple can remain high for years if the market has simply re-rated the company.

This video walks through the Spike Pullback Analysis tool: how to set it up using your own spike and pullback levels, how to read the metrics, how to spot how many bounces typically follow a pullback, and how to filter those bounces to match the way you trade.

Between June 10 and August 20, 2026, TradeIntel published 8,275 option buying ideas on its dashboard. 4,001 have reached expiration: 3,033 calls and 968 puts.

This is how they performed:

↓ Download all the trades we scanned.

The Human Factor: The 3-Filter Verification Method

Our scanning algorithm measures contract behavior and historical probabilities. It does not know market catalysts, what is scheduled for next week or what positions your portfolio is holding.

Before acting on any TradeIntel alert, run these three mandatory sanity checks:

1. Verify Stock Trend (check the chart first) Case Study: AAPL travelled 26% between its low and high inside the window and every one of its call ideas doubled. Conversely, COST stayed locked in a tight 10% band and 89% of its calls failed to double. Check the chart before placing the trade.

2. Case Study: AAPL travelled 26% between its low and high inside the window and every one of its call ideas doubled. Conversely, COST stayed locked in a tight 10% band and 89% of its calls failed to double. Check the chart before placing the trade.

3. Count Ideas, Not Contracts Buying five different strike options on semiconductor names isn’t diversification; it’s one single position sized five times larger.

4. Buying five different strike options on semiconductor names isn’t diversification; it’s one single position sized five times larger.

5. Confirm Directional Bias The system presents mathematical odds based on contract behavior. Align those odds with fundamental catalysts (macro trends) that you agree with.

6. The system presents mathematical odds based on contract behavior. Align those odds with fundamental catalysts (macro trends) that you agree with.

Transparency

• Every idea was stamped in real time on the dashboard the moment it fired.

• Peaks are calculated from the daily high so you can see how often price would have reached a given profit target. For example, a resting limit order at +50% fills once the contract trades through that level, and the peak data shows how frequently that happened.

• Roughly half the ideas published between June 10 and August 11, 2026 are still open. Only positions that have reached expiration appear in these figures. The table will be updated when the remaining contracts settle.

The reason is straightforward. From mid-2025 onward the share price and the underlying business moved in opposite directions.

The profit multiple shows the same pattern.

A stock can look inexpensive on sales yet expensive on earnings when margins are shrinking. That is not the case here. Profit per share rose in step with revenue, increasing 17 percent, while the earnings multiple itself fell 36 percent.

How today’s multiples rank against META’s own history

The table below shows where the current price-to-sales and price-to-earnings multiples rank against META’s own history over different look-back windows.

18–20% Upside to Historical Multiples

From here, META would need to rise 20% to reach the middle of its own valuation range. Its median price-to-sales multiple over the past five years is 7.40 times. Applied to today’s trailing sales of $88.95 per share, that points to a price of $658.55.

The gap is narrower on the earnings side. The five-year median multiple of 24.31 times, applied to trailing earnings of $26.54 per share, implies a price of $645.16, or 18% higher than today’s level.

Note

“Cheap” in this analysis means inexpensive only relative to META’s own historical range. It does not speak to valuation versus peers or to the intrinsic value of the business. A multiple can remain low for years if the market has simply re-rated the company.

WMT gaps down after earnings 41.8% of the time (51 times out of 122 earnings reports since May 1996).

The size of the gap was well above average. A gap down of 6.93% vs its average gap down of −2.71%.

Forward returns - After Any Gap Down

After any gap down, WMT tends to chop sideways and go nowhere until next earnings.

Forward returns - After Comparable Gap Downs

The 2 gaps this size or bigger had negative average returns until next earnings.

After a gap down in a low VIX environment WMT typically chops sideways and goes nowhere until next earnings.

The forward return calculations in this article begin at the 9:30am open that morning after the stock has already gapped down, so the gap is not part of the forward calculations.

A crowded trade happens when too many people are making the same trade at the same time.

The problem with crowded trades isn’t that the underlying idea is wrong. The problem is that when too many people are positioned the same way, any reason to exit creates a stampede for the door. And since everyone is trying to get out at the same time, prices move violently against the position.

Think of it like a crowded theater when someone yells “fire.” It doesn’t matter if there’s actually a fire or not because the crowd rushing for the exits creates its own disaster. In trading, the crowd trying to exit a position creates its own price collapse, regardless of whether the stock is good or not.

That is the risk with AI stock advice right now.

People are asking the same models the same questions and getting back the same stock recommendations. The reasoning sounds nearly identical across platforms. Growth looks durable. The competitive position looks strong.

The result is correlated buying into a relatively narrow set of ideas.

When that happens the overcrowded results show up fast.

When something eventually snaps, whether it is an earnings miss or a broader shift in liquidity, the unwinding process will not look like a standard market correction.

Real market edge usually doesn’t come from running the exact same trades as everyone else.

To be clear, we are not quite there yet. The conditions for AI-driven crowded trades are still forming, but as these AI tools become more embedded in everyone’s daily trading workflows, that reality could arrive sooner than most people think.