Failed Patterns, and What They Tell You
Most patterns fail. That is the base rate rather than a defect, and a failure carries real information because of who it traps.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- Failure is the normal outcome of any chart pattern, not an exception.
- A failed breakout traps the participants who acted on it, and their exits move price.
- Hindsight bias makes patterns look far more reliable than they are.
- Defining failure in advance is what separates a plan from a story.
- A method has to survive most of its signals failing, or it is not a method.
MAD Academy Training Video · 0:45
A Failure Is Information, Not Bad Luck
When a pattern fails, a large group of traders is instantly wrong — and that is often a better trade than the pattern was.
This lesson is part of a Stock Alerts + Tools plan.
The base rate
Any honest treatment of chart patterns has to start here. The majority of identified patterns do not produce the move associated with them. Published studies vary widely in method and in result, but none of them find a shape that works most of the time.
This is not an argument against using patterns. It is an argument against expecting them to be right, and for structuring around the possibility that they are not. A method that survives a majority of its signals failing is a method; one that requires them to work is a hope.
The risk-reward article carries the arithmetic that makes this workable: at a three-to-one ratio a method only needs to be right a quarter of the time to break even. A high failure rate is survivable; what is not survivable is a high failure rate combined with losses the same size as the wins.
Why a failure moves price
When a widely watched level breaks, participants act on it. When price immediately returns through the level, every one of them is offside, and their exits are supply pushing in the opposite direction.
This is the mechanism behind the observation that failed moves often lead to fast moves the other way. It is not mysterious; it is a population of positions being closed at once, on top of the stop orders resting just beyond the level.
Scroll the chart sideways to see all of it.
Hindsight bias
Looking back at a chart, the patterns that worked are visible and the ones that did not are invisible, because a pattern that failed does not look like a pattern afterwards. Every educational chart in every book is selected after the fact from the set that worked.
The correction is to identify patterns on the right-hand edge, in real time, and record them before the outcome is known. Almost nobody does this, and almost everybody who does is surprised by the result.
This is also why a trade journal is the only reliable evidence about whether a pattern works for the person using it. Published base rates are averages over somebody else's identification criteria; a personal record is about yours.
Defining failure in advance
A pattern is only useful if there is a price at which it is agreed to be wrong. Without that, a failing setup becomes a story about why the level does not really count, and the position outlives the reasoning that created it.
Written before entry, that price is a fact. Decided afterwards, it is a negotiation with somebody who now has a position and an opinion, and that negotiation has a predictable outcome.
Defining failure before it happens
A pattern without a stated failure condition cannot fail, and one that cannot fail cannot be evaluated. Writing the condition down in advance is what converts a description into something testable.
| Structure | A workable failure condition |
|---|---|
| Breakout from a base | Closing back inside the base |
| Bull flag | A close below the low of the flag |
| Double bottom | A close below the lower of the two lows |
| Head and shoulders | A close back above the neckline |
| Trendline break | A close back on the original side of the line |
Each of these is a specific, observable event rather than a judgement, and each can be checked without deciding whether the pattern is still valid. That is the property that matters: a condition requiring interpretation will be interpreted favourably by whoever holds the position.
The rate at which these conditions are hit is the base rate, and it is high for every named pattern. That is not a defect in the patterns; it is what the word probability means, and a process that treats a failure as an anomaly has assumed a base rate that no evidence supports.
Hindsight, and why failures are invisible
Every chart used to illustrate a pattern is chosen because it worked. That is not dishonesty; it is what an illustration is for. The consequence is that the visual memory built up from any book, course or feed is a sample from which every failure has been removed.
The effect compounds because failures are also harder to see afterwards. A structure that failed rarely looks like the pattern in retrospect: the eye reorganises the chart around what happened, and the shape that was clear at the time dissolves into the noise around it.
This is why a personal record of every instance identified in advance is worth more than any number of examples. It is the only sample that includes the failures, because it was assembled before the outcomes were known.
The same mechanism explains why patterns feel more reliable the longer someone has been looking at charts. Accumulated experience is accumulated exposure to a filtered sample, and the filter runs continuously and invisibly.