The patterns with actual math behind them
Classical TA books are full of chart-pattern taxonomies. Most of them are vibes. Thomas Bulkowski's Encyclopedia of Chart Patterns (3rd ed., 2021) is the exception — 150,000 samples across ~1,000 stocks from 1991 through the late 2010s, every pattern tested, every win rate published.
For reversal patterns — the moment a trend flips — Bulkowski is the reference. We'll lean heavily on his numbers here. Murphy's Chapter 5 provides the narrative frame; Bulkowski provides the reality check.
One housekeeping note on Bulkowski's numbers, in his own words:
The performance numbers are based on perfect trades. Are the results realistic? Not really. You likely won't be able to duplicate them in real life. — Encyclopedia of Chart Patterns
Treat his stats as upper bounds. Slippage, commissions, and imperfect entries give back meaningful chunks. The relative ranking between patterns is more useful than the absolute numbers.
The patterns — gallery
Classic head-and-shoulders top: left shoulder (LS) forms on strong volume. Head pushes to a new high but volume softens. Right shoulder (RS) is lower and on still-weaker volume — the buyers are gone. Neckline break confirms. Per Bulkowski, H&S Top is one of the most reliable bearish patterns in his database — but its performance rank varies significantly depending on how strictly you define the pattern.
Toggle the four patterns above. Each is hand-crafted to show the structure cleanly; real charts are messier. Key reference lines and landmarks are marked. Now let's walk each one with its stats.
1. Head and Shoulders Top
The prototype bearish reversal. Three peaks — the middle (head) higher than the two sides (shoulders) — with a neckline connecting the two intervening troughs. Murphy:
The head and shoulders reversal is probably the best known and most reliable of all major reversal patterns.
Structure Murphy teaches:
- Prior uptrend
- Left shoulder on heavy volume
- Rally to a higher head on lighter volume
- Decline to the neckline
- Third rally (right shoulder) on "noticeably light volume" that fails to reach the head's high
- Close below the neckline — confirmation
- Frequent return move back to the neckline before new lows
Bulkowski's stats for H&S Top (3,577 patterns tested):
| Metric | Bull market | Bear market |
|---|---|---|
| Performance rank | 9 of 36 | 4 of 19 |
| Breakeven failure rate | 19% | 5% |
| Average decline from breakout | −16% | −24% |
| Pullback rate | 68% | 67% |
H&S tops work much better in bear markets — the pattern fits the regime. In a bull market, nearly one in five confirmed H&S tops fail to move even 5% from the neckline break before reversing. That's a sobering number for the "best-known reversal pattern."
Volume: the textbook lore is wrong
Every TA textbook says volume should decline from left shoulder to head to right shoulder. Bulkowski actually tested this across 3,500+ patterns:
Volume has little to do with performance. — Bulkowski
Patterns with rising volume and patterns with falling volume produced identical average declines (−16% either way). Heavy-vs-light breakout volume: same result. The classical volume story is narratively useful for identifying patterns, but not predictive of results. Kaufman concedes this openly: "There are many examples of successful head-and-shoulders formations that do not satisfy the volume criteria."
The neckline-slope trap
An under-taught detail: when the neckline slopes downward, price may never reach it before reversing. Bulkowski's fix:
For down-sloping necklines, wait for a close below the right armpit — the low between the head and the right shoulder — not the neckline itself.
Missing this costs you trades that never fire.
2. Head and Shoulders Bottom (Inverse H&S)
The bullish mirror, but with an important asymmetry. Murphy:
Volume plays a much more critical role in the identification and completion of a head and shoulders bottom. Heavy volume is an absolutely essential ingredient in the completion of the basing pattern.
Bottom formations are born out of capitulation; they require buying pressure that the topping pattern doesn't need.
Bulkowski's H&S Bottom stats (3,832 patterns):
| Metric | Bull market | Bear market |
|---|---|---|
| Performance rank | 13 of 39 | 10 of 20 |
| Breakeven failure rate | 11% | 9% |
| Average rise from breakout | +45% | +28% |
| Throwback rate | 65% | 66% |
Notice the 45% average rise in bull markets — one of the strongest numbers in the book. And unlike H&S tops, gaps help H&S bottoms: patterns with a breakout gap average +48% / +32% vs +45% / +27% without. Direction matters.
3. Double Top
The single most misidentified pattern in all of technical analysis. Murphy:
The terms 'double top and bottom' are greatly overused in the financial markets. Most potential double tops or bottoms wind up being something else.
Bulkowski's confirmation rule is absolute and worth memorizing verbatim:
The twin-peak pattern becomes a true double top when price closes below the confirmation line (or price). Price continues rising 60% of the time without confirming the twin-peak pattern. Thus, always wait for confirmation unless you have valid reasons for trading earlier.
Read that number again: 60% of visible "double tops" never confirm. The twin-peak shape on its own means nothing. The pattern is defined by the break of the intervening trough, not by the visual similarity of the two peaks.
Adam and Eve
Bulkowski's taxonomy splits double tops by peak shape:
- Adam = narrow, pointed, inverted-V, spike-like
- Eve = wide, rounded, U-shaped with a plateau
Four combinations, all separately tested and ranked:
| Variant | Rank (bull) | Breakeven fail | Avg decline | Pullback |
|---|---|---|---|---|
| Adam & Adam | 19/36 | 25% | −15% | 64% |
| Adam & Eve | 10/36 | 21% | −16% | 64% |
| Eve & Adam | 16/36 | 21% | −15% | 64% |
| Eve & Eve | 12/36 | 20% | −16% | 65% |
The Adam & Eve variant — pointy peak followed by rounded peak — is statistically the strongest, which is the opposite of most people's intuition (most assume "two matching peaks" is the tell). Eve & Eve — two rounded peaks — is "what many refer to as the classic double top" per Bulkowski, but it ranks only 12th of 36.
The twin-Adam pointy variant has the worst failure rate: one in four confirmed breakdowns never reach even 5% decline.
4. Double Bottom
The bullish mirror, same four-variant taxonomy. The headline stat:
The Eve & Eve double bottom is what many refer to as the classic "double bottom"… The performance rank is more than three times better than the other three varieties of Adam and Eve double bottoms. — Bulkowski
| Variant | Rank (bull) | Breakeven fail | Avg rise | Throwback |
|---|---|---|---|---|
| Adam & Adam | 26/39 | 16% | +39% | 67% |
| Adam & Eve | 17/39 | 12% | +43% | 67% |
| Eve & Adam | 20/39 | 12% | +42% | 67% |
| Eve & Eve | 5/39 | 12% | +50% | 65% |
The Eve & Eve double bottom — two rounded bottoms — ranks 5 of 39 across all bullish patterns Bulkowski catalogs. It is the single highest-ranked reversal pattern in his database for bullish trades. 50% average rise. 12% failure rate. Confirmed by breaking the intervening peak.
If you can identify exactly one chart pattern, this is the one.
5. Triple Top / Triple Bottom
Less common, often confused with H&S. Murphy:
Chartists often disagree as to whether a reversal pattern is a head and shoulders or a triple top. The argument is academic, because both patterns imply the exact same thing.
Triple Top — rank 24/36 bull, 10/19 bear. Failure rate 25% in bull markets. Bulkowski is blunt:
Failures are three times higher in bull markets. Ouch. You should trade triple tops only in bear markets.
Triple Bottom — rank 12/39 bull, failure rate 13%, average rise +46%. Less exceptional than H&S or Eve & Eve double bottoms, but solid. 74% reach their measure-rule target.
6. Rounding Patterns (Saucers)
Slow, gradual reversals. Hard to spot in real time — usually only obvious in retrospect.
Rounding Bottom is an outlier in the database:
In bear markets, rounding bottoms are the best performing chart pattern. Rounding bottoms blow the doors off your Datsun. — Bulkowski
Stats:
- Rank 1 of 20 in bear markets — literally the best.
- Breakeven failure rate: 4% bull / 6% bear — the lowest in the catalog.
- Average rise: +48% bull / +37% bear.
The catch Bulkowski flags:
67% of rounding bottoms are actually continuations, not reversals.
Two-thirds of "rounding bottoms" happen in an existing uptrend and extend it, rather than reversing a downtrend. Same shape, different meaning depending on the prior trend — the same lesson we've been drilling since Dow.
Rounding Top is the stranger cousin: 58% of the time price breaks upward out of a rounding top — the opposite of what the name suggests. Only the downward breakouts behave as reversals; those rank well (3/36 bull), but you're already filtering 58% of cases out.
The measure-rule disappointment
Every textbook tells you to project the pattern's height to estimate the target. H&S top: measure from head to neckline, subtract from the neckline to get the target.
Bulkowski tested this on H&S tops in bull markets. The measure-rule target is reached only 51% of the time. Worse: 71% fail to reach a 20% decline. A generation of traders have been using a rule that works slightly better than a coin flip.
The false-signal warning
Murphy names the pattern every H&S trader will eventually meet:
Any decisive close back above the neckline is a serious warning that the initial breakdown was probably a bad signal.
The "failed H&S" — price breaks the neckline, then reclaims it and runs to new highs — is so common it's a recognized sub-pattern. Which is why position sizing and stops matter more than pattern identification. You'll be wrong; survive it.
Quick check
You see two roughly-equal peaks on a chart, a few weeks apart, with a dip between them. Is this a double top?
What you now know
- Bulkowski's data is the reality check that most chart-pattern instruction skips. Pattern rank and breakeven failure rate beat vibes.
- Head and Shoulders Top: 19% failure in bull markets, −16% average decline; works much better in bear markets (5% failure, −24% decline). Volume lore is largely wrong — his data shows volume trend doesn't predict performance.
- Head and Shoulders Bottom: +45% average rise in bull markets. Asymmetric with the top — bottoms require heavy volume, gaps help, and capitulation is essential.
- Double Top: 60% of visible twin-peak shapes never confirm. The pattern is defined by break of the intervening trough, not by the visual similarity of peaks. Adam & Eve taxonomy matters.
- Double Bottom — Eve & Eve is the flagship pattern of the entire catalog. Rank 5 of 39, +50% average rise. Two rounded bottoms + confirmation break = best bullish reversal in Bulkowski's database.
- Triple Top: 25% failure in bulls — avoid. Trade only in bears. Triple Bottom is solid (rank 12, +46%).
- Rounding Bottom: rank 1/20 in bear markets, 4–6% failure — but 67% of them are continuations, not reversals. Context is the classifier.
- Measure-rule targets are not guarantees — H&S Top hits its target only 51% of the time.
Next: Continuation Patterns (coming soon) — triangles, flags, pennants, wedges. Shapes that pause rather than reverse a trend. Different stats, different rules.