Chart patterns and their failure rates

Chart patterns are the folklore of technical analysis — named shapes passed down with origin stories and promised outcomes. The folklore is not worthless: the classic patterns are real, recurring descriptions of how crowds behave around turning points and consolidations. The problem is the packaging. Patterns get sold as predictions with success rates implied to be high, when honest measurement puts most of them closer to a coin flip with a modest lean. The professional move is not to discard them — it is to trade the lean, plan for the flip, and extract information from both outcomes.

Underneath every named pattern is the same raw material: support, resistance, and trend structure. A pattern is just structure with branding. Read it that way and the failure cases stop being betrayals and start being data.

Head and shoulders: a trend losing its grip

The shape — a peak, a higher peak, a lower peak, all above a connectable neckline — is simply an uptrend failing in slow motion: the head is the final higher high, the right shoulder is the first lower high, and the neckline break is the structure break you already know how to read. That is why it deserves its reputation as one of the better patterns; it is a trend reversal definition wearing a costume. It still fails routinely. The neckline breaks, price reverses, and reclaims the line — and that specific failure, trapping a fresh crop of shorts whose covering fuels the rally, is one of the stronger bullish events on a chart. The pattern that fails loudest informs best.

Triangles: compression without commitment

Triangles — ascending, descending, symmetrical — are consolidations where the swings contract: the market is compressing, volume usually drying up, both sides accepting a narrower and narrower band. The folklore says ascending triangles break up and descending ones break down, and measured studies do find a lean in those directions — a lean, not a law, with the contrary outcome arriving often enough to make betting the farm on direction a donation. The genuinely reliable content of a triangle is not direction but energy: compression precedes expansion, and the eventual break tends to travel. Many experienced traders therefore refuse to predict triangles at all and simply plan both branches in advance, trading the break — or, more conservatively, the first orderly retest of the broken boundary.

Flags: the trend catching its breath

A flag is a sharp directional move — the pole — followed by a shallow, drifting countertrend channel on fading volume. The logic is continuation: a move violent enough to form a pole reflects real initiative, and a pullback too weak to retrace much of it suggests the other side has little appetite. Flags are among the more dependable patterns for exactly that structural reason, with two strict conditions: the pole must be genuinely impulsive, and the flag must stay shallow — typically holding above the lower half of the pole. A “flag” that retraces 80% of its pole on rising volume is not resting; it is reversing, and renaming it will not change the outcome.

Failure is information

Here is the reframe that separates pattern users from pattern victims: a failed pattern is not a malfunction, it is a second signal — often a better one. Every completed pattern attracts a crowd positioned for the textbook outcome. When the pattern fails, that crowd is trapped at bad prices, and their forced exits add fuel to the move against them. A failed head and shoulders becomes upside thrust; a failed breakdown from a descending triangle becomes a short squeeze. Trading these failures requires nothing exotic — only that you marked the level whose reclaim proves the failure, and that you were not so married to the original pattern that you ignored its funeral.

Measured moves, with skepticism

Tradition assigns each pattern a price target: project the height of the head above the neckline downward, the widest part of the triangle from the breakout, the length of the pole from the flag. Treat these as rough planning estimates, not promises — moves fall short of their measured targets regularly, and sometimes blow far past them, because nothing in the geometry obligates the market to anything. The targets are still worth computing, for one unglamorous reason: risk-reward arithmetic. If the measured move from a flag projects 4% of upside and your invalidation sits 3% away, the pattern can be picture-perfect and the trade is still mediocre. The math does not care how clean the chart looks.

Context multiplies or erases whatever edge a pattern carries. A bull flag inside an established uptrend, forming above prior resistance, is the pattern working with the larger structure; the identical shape counter to a strong downtrend is a shape and nothing more. Before naming any pattern, name the regime it lives in — most published failure statistics quietly mix both situations, which is one reason they look like coin flips in aggregate.

Key takeaways

• Patterns are recurring crowd behaviors built from ordinary structure — useful tendencies, never guarantees, and often near coin-flip with a lean.
• Head and shoulders is a trend-structure break in costume; that is the source of what reliability it has.
• Triangles reliably signal compression and a coming expansion; their direction folklore is only a lean — plan both branches.
• Flags depend on an impulsive pole and a shallow, low-volume pullback; violate either condition and the pattern is fiction.
• Failed patterns trap positioned crowds and often produce stronger moves than completed ones — mark the failure level in advance.
• Use measured moves for risk-reward arithmetic, not prophecy.