False Breakout: When the Range Wins
Michael Whitman
·
Trading since 2017, full-time since 2020
·
Published 2026-09-06
·
Updated
·
Reviewed
A false breakout is price closing beyond a range boundary and then closing back inside it. What separates it from a real breakout is the return: the move reached the level, failed to hold beyond it, and left everyone who entered on the break on the wrong side.
Covered on this page: TradingView.
A false breakout is a close beyond a range boundary followed by a close back inside it. The word “false” can only be applied afterwards, so this page fixes a definition in advance and counts how often it happened on SPY’s daily chart from 1993 to 2026, against how often ordinary days would have produced the same result.
How it works
A range has two boundaries and price spends most of its time between them. A breakout is price closing beyond one. A false breakout is price closing beyond one and then closing back inside.
The word “false” is assigned afterwards. At the moment of the break, the two are the same picture: price above the high, the move looking decisive. Nothing on the break day separates them, which is the entire difficulty.
Use closes, not wicks. Price pokes above a level on many approaches; a wick beyond the boundary is ordinary noise. A close beyond it, followed by a close back inside, is the sequence worth naming. A day that trades above an old high and closes back below it on the same day is a different setup, measured on the turtle soup page.
The definition used on this page. A breakout is a close above the highest high of the previous 20 sessions, on a day when the previous close was not already above its own 20-session high. It fails if any of the next five closes is below that broken level. The two numbers, 20 and five, are settings; the data section shows what changing them does.
Why the boundary attracts the break
Everyone can see the same high. Traders who are short often place protective stops above it, and traders waiting to buy place entry orders above it. Both are buy orders, resting in the same small band of price.
A move into that band can fill quickly, because the resting orders supply the other side. A large seller looking for buyers finds them there. That is one common explanation for why a break can reverse from just beyond an obvious level.
This does not require anyone to be manipulating anything. A cluster of orders is a natural place to transact size. The stop hunt page covers the same mechanic from the other direction.
Volume is the usual filter offered, and it is a lean at best. A break on light participation is often said to fail more than one on heavy participation. The SPY file used here carries no volume, so this page does not test that claim.
A worked example
SPY, August 2026. The highest high of the 20 sessions before 13 August 2026 was $776.85, set on 5 August. On 13 August SPY closed at $777.88, $1.03 above it, a margin of 0.13%. By the definition above, that is a breakout.
It failed the next day. On 14 August SPY closed at $776.34, below $776.85. On 18 August it closed at $767.45, and on 1 September at $761.78.
And it did not come back. Through 25 September 2026, no close went back above $776.85; after 14 August the highest was $773.50 on 21 September. That makes it one of the minority: most failed breakouts in the sample closed back above their level within a few sessions.
Read the margin before anything else. A $1.03 clearance on a $777 share price is 0.13%, and 264 of the 623 breakouts in the sample cleared their level by less than 0.25%. The data below shows that small margins like this one failed far more often than large ones, and that ordinary days would have produced almost exactly the same pattern.
The original data
The sample. SPY daily bars, 8,472 of them, from 29 January 1993 to 25 September 2026, prices only. The rule above found 623 breakouts of a 20-day high with five sessions still to come, about 18.5 a year.
How often they failed. 343 of the 623, 55.06%, closed back below the broken level within five sessions. 153 failed on the very next close.
Against what base rate? A breakout’s failure only needs a dip as large as its margin. So for each breakout the same question was asked of every day in the sample: how often did a close within the next five sessions fall below the starting close by at least that breakout’s margin? Summed over the 623 breakouts, ordinary days would have produced 354.7 failures, 56.93%. The breakouts produced 343, slightly fewer, and the gap is within chance (z = -0.97, p = 0.33).
| Margin above the level | Breakouts | Failed within 5 sessions | Expected from ordinary days |
|---|---|---|---|
| Under 0.25% | 264 | 68.94% | 65.86% |
| 0.25% to 0.75% | 277 | 46.93% | 54.51% |
| 0.75% or more | 82 | 37.80% | 36.36% |
| All | 623 | 55.06% | 56.93% |
The margin, not the breakout, set the failure rate. Small breaks failed most and large ones least, and the expected column moves the same way, because a small dip is common on any day. The middle row is the exception: breaks that cleared the level by 0.25% to 0.75% failed 130 times against 151.0 expected (z = -2.54, p = 0.011). The same group showed the same gap at 55-day and 252-day highs, but those samples share many of the same days, so they are not independent confirmation. It is one group, in one market, from three tested.
Other settings gave the same answer. Breaking a 55-day high: 254 of 465 failed within five sessions, 54.62%, against 57.88% expected (p = 0.15). Breaking a 252-day high: 196 of 362, 54.14%, against 58.39% (p = 0.10). Allowing ten sessions instead of five, failures rose to 66.77% for the 20-day rule, against 66.72% expected.
Failures were usually brief. Of the 343 failures, 301 (87.76%) closed back above the same level within 20 sessions of the failing close, a median of 2 sessions later. The August 2026 case above is one of the 42 that did not.
And the label did not sort what came next. Measured from the fifth session, when the outcome was known, to 20 sessions later: SPY was higher after 64.72% of failed breakouts and 61.43% of the ones that held, against 63.91% of all 20-session windows. Neither differs from the base rate by more than chance (p = 0.78 and p = 0.38). Every breakout, its margin and its outcome are in the SPY 20-day breakout file.
In practice: what a check costs
A genuine break tends to keep going is the usual advice, and on SPY’s daily chart the measurement is blunter: in the first five sessions a break looked like any other day of the same size.
Waiting for a retest has a price. Price breaks, moves away, returns to the boundary, and either holds or falls back through. A rule that waits gives up part of the move, and misses the breaks that never return.
A higher timeframe shows fewer of them. A five-minute range broken and reclaimed can be a single candle with a wick on the hourly chart. That reduces the count; whether it improves the outcome is a separate question this page does not test.
Each attempt pays a round trip in spread and commission. With more than half of all daily breaks closing back inside within a week, a rule that takes every break pays that cost often.
What it is not
It is not manipulation. It is size transacting where liquidity exists. Describing it as a deliberate trap aimed at you personally is a story attached to an ordinary mechanic.
It is not knowable in advance. Volume, timeframe and retest rules shift the picture, and none of them identifies a false break before it fails.
It is not always the end of the move. On SPY, almost nine in ten failures were followed by a close back above the level within 20 sessions. The label describes one attempt, not the level’s future.
And it is not a reason to fade every breakout. A rule that sells every failure sells into a market that, on this data, went on to rise about as often as on any other day.
When it fails
The characteristic failure is fading a break that then resumes. Price breaks out, closes back inside, you position for the range to hold, and it breaks again and runs. On SPY’s daily chart that second break usually came within days.
The second failure is defining the range loosely. A boundary drawn generously produces breaks and reclaims constantly, because the level was never precise enough to break in the first place.
A third is ignoring the margin. A close a few cents above the level and a close 1% above it are counted as the same event by most rules, and the table above shows they behave very differently.
A fourth is trading the label in a rising market. SPY rose over most of the sample, and failed and held breaks were followed by much the same 20-session results. A view about direction has to come from somewhere other than the failure itself.
And a fifth is sizing it like a high-conviction trade. Every read on this page is probabilistic, and the measured version of it is close to a coin toss weighted by the market’s own drift.
Related
Breakout is the event this is the failure of, and the pair belong together. Trading range is the structure both live inside. Stop hunts is the same liquidity mechanic described from the other side, and turtle soup measures the same-day version, where the high is taken and the close lands back below it.
Decide before the break how far beyond the level a close has to be, and how long it has to stay there, for you to call it real. Most breaks clear the level by a fraction of a percent, and at that size a close back inside is the ordinary outcome.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money. Some links on this page earn a commission if you buy through them. It costs you nothing and it does not decide what appears here or in what order — how these pages are made is set out in our methodology.