A false breakout, or fakeout, is a move beyond a support level, resistance level or pattern line that does not hold: price pokes through, then closes back inside. You can't avoid fakeouts completely, but you can filter many of them. Judge the candle close, not the wick, ask for a close beyond the level by a margin scaled to volatility, and size every trade so that the fakeouts that get through cost a known amount.
To see how much each filter helps, we tested them on seven years of Bitcoin 4-hour candles. When price first poked through its recent high or low, about half of those candles closed back inside. Waiting for a close beyond the level, and then for a close beyond it by a volatility margin, cut the share of breaks that fell back within a day from 78.5% to 38.5%. That is a big improvement, but more than one break in three still fell back.
- A false breakout moves beyond a level and then closes back inside instead of following through. Bull traps and bear traps are the same thing seen from each side.
- On Bitcoin's 4-hour chart (Jan 2020 to Oct 2026), 51.4% of the candles that poked through their 20-candle high or low closed back inside: a wick, not a breakout.
- Entering on the touch, 78.5% of breaks closed back inside within 24 hours. Waiting for a close cut that to 55.8%, and a close beyond the level by half an ATR to 38.5%.
- Filters cost signals: the strictest rule kept 377 of 1,596 breaks.
- No filter removes fakeouts, so the stop and the position size have to be set before the entry.
What is a false breakout?
A breakout is price moving beyond a level the market has respected, such as a recent high, a support zone or the line of a triangle. A false breakout is one that fails. IG's trading guide puts it simply: a false breakout is a failed breakout, where price moves through the level but does not keep its momentum.
You'll see several names for the same thing:
- Fakeout or failed breakout: the general term.
- Bull trap: an upside break that fails, trapping the traders who bought it.
- Bear trap: a downside break that fails, trapping the traders who sold or shorted it.
- Busted pattern: Thomas Bulkowski's term for a chart pattern that breaks out, moves no more than 10% and then reverses. He calls the 10% "arbitrary", which is a fair warning: there is no single official definition.
Because definitions differ, statistics about fakeouts differ too. This post uses one simple, checkable rule: a break is false when a candle closes back inside the level soon after it.
Why fakeouts happen
A few forces are at work, and in crypto they often act at once:
- The candle isn't finished. Until a candle closes, its high, low and close keep changing. TradingView's Pine Script documentation notes that on a live bar these values are not fixed and can change many times before the bar closes. A break you see halfway through a 4-hour candle may not exist when the candle ends.
- Obvious levels attract orders. A widely held explanation is that many traders place stops and entry orders just beyond the same visible highs and lows. A push through the level fills those orders, and once they are filled there may be nobody left to carry the move further.
- Leverage adds forced orders. Perpetual futures liquidate positions automatically when margin runs out, and those forced orders can push price through a level for a moment. On 10 and 11 October 2025, CoinGlass counted more than $19 billion of forced closures across more than 1.6 million accounts, the largest liquidation event on record.
- Weak participation. CMC Markets notes that a breach on weak volume may lack the participation needed to sustain the move, and that false breakouts occur "with frustrating frequency".
Wick or close: what counts as a breakout
The same level can be "broken" in three very different ways:
- Wick only. The high trades above resistance, but the candle closes back below it. On a closing basis nothing broke.
- Weak close. The candle closes above the level, but only just, by less than a normal candle's noise. A small push back undoes it.
- Confirmed close. The candle closes above the level by a clear margin, scaled to how much the market usually moves.
Bulkowski uses closes for the same reason. In his notes on confirmation he points to a chart where price pierces the line but does not close above it, and he counts a pattern as confirmed only when price closes outside it.
The test: 1,596 Bitcoin breaks, three entry rules
We wanted to know how much each step actually filters. The test is deliberately simple:
- Data: Binance BTCUSDT spot, 4-hour candles, 1 January 2020 to 31 October 2026 (14,975 candles), from Binance's public market data API.
- The level: the highest high, or the lowest low, of the previous 20 candles (about 3.3 days).
- A break: the first candle whose high trades above that level, or whose low trades below it, after a candle that did not. Up and down breaks are counted alike: 1,596 in total.
- Three entry rules: enter when price touches the level (a stop order resting just beyond it), enter only on a candle that closes beyond the level, or enter only on a close beyond the level by at least half of the 14-candle ATR.
- A false break: any 4-hour candle closes back inside the level within 24 hours of the break. For the touch rule that includes the breaking candle itself.
ATR is the average true range, a volatility measure developed by J. Welles Wilder. StockCharts' ChartSchool explains it well: it tells you how much a market typically moves per candle, and it "doesn't indicate price direction, just volatility".
| Entry rule | Signals | Closed back inside within 24h | Share |
|---|---|---|---|
| Enter on the touch | 1,596 | 1,253 | 78.5% |
| Enter on a close beyond the level | 776 | 433 | 55.8% |
| Enter on a close beyond by 0.5 ATR | 377 | 145 | 38.5% |
What the numbers say
- Half of all breaks were only wicks. 820 of the 1,596 breaking candles (51.4%) closed back inside the level. A stop order at the level would have bought or sold every one of them.
- Each filter helped, and each one cost signals. Waiting for a close cut the false share from 78.5% to 55.8%. Asking for half an ATR beyond the level cut it to 38.5%, but only 377 of the 1,596 breaks qualified, fewer than one in four.
- The best rule still failed often. More than a third of the strictest breaks closed back inside within a day. A filter changes the odds; it does not remove the risk.
How far to trust this
The exact percentages depend on the definition. With a 12-hour window instead of 24 hours, the strictest rule's false share was 24.4%; with 48 hours it was 50.1%. With a 10-candle or a 50-candle level the shares moved by only a point or two. The ranking never changed: touch worst, close better, close by a margin best.
This test also does not measure profit. It has no fees, stops or targets, and a close back inside is not always a loss: a trade with its stop further away can ride out a brief dip back through the level. Entering on a close also means a worse price than entering on the touch. Read the table as a measure of how often a level gives way, not as a strategy result.
Five ways to filter false breakouts
- Wait for the candle to close. This alone removed the 820 wick-only breaks in our test. IG gives the same advice and suggests alerts set on the candle close so you don't have to watch the chart.
- Ask for a margin scaled to volatility. A fixed amount, say $200 above resistance, means a lot on a quiet day and nothing on a wild one. A fraction of the ATR adjusts by itself.
- Check the higher timeframe. A break on a 15-minute chart that runs straight into resistance on the 4-hour chart is fragile. IG also suggests comparing a break with the longer-term charts before acting on it.
- Look at volume. A break on unusually low volume deserves more doubt. CMC Markets is careful to add that volume confirmation does not guarantee success.
- Wait for a retest, if you can accept missing some moves. Some traders enter only when price comes back to the broken level and holds there. This gives a closer stop, but strong breakouts often never come back.
No filter removes false breakouts. In our test even the strictest rule saw 38.5% of breaks close back inside within a day. Decide the stop and the position size before you enter, so that a fakeout costs exactly what you planned.
The last point matters most. A fakeout that slips through your filters should be a planned, small loss, not a surprise. If you're not sure how to size a position from your stop, our guide on how to calculate position size walks through it step by step.
How Navixa's Price Action strategy handles fakeouts
Navixa Price Action trades breakouts of wedges, triangles and channels, so fakeouts are its main enemy. It uses the same filters described above:
- It judges breakouts on candle closes only, and it draws a pattern only after its last turning point is confirmed, so the pattern never repaints.
- It needs a close beyond the pattern line by a margin scaled to the coin's volatility. A close past the line that is not far enough gets no trade. The margin is the strategy's own, not the 0.5 ATR from our test.
- The stop, two targets and the position size are set at the breakout close, with 2% of the balance at risk by default.
- It skips Sunday breakouts by default, and when a confirmed breakout fires the other way, it closes the open trade and reverses.
Filters still don't make fakeouts disappear. In the published backtest on Bitcoin perpetual futures (BTCUSDT.P, 4-hour chart, October 2019 to October 2026), 27% of the 58 positions were stopped out before the first target and another 9% were closed by a breakout in the opposite direction. What the plan controlled was the cost: the largest loss was -1.0R, the planned risk, fees included. These are past backtest results, not a promise. Every trade is on the Price Action results page.
Can a failed breakout be a signal?
Yes, and some traders look for it on purpose. When a breakout fails and price closes decisively through the other side of the pattern, the traders who bought the first break are trapped, and their exits can push price further the other way.
Bulkowski's data on stocks points the same way for some patterns. In his busted-pattern table, a busted symmetrical triangle went on to an average rise of 52%, against 34% for a normal one. These are stocks in bull markets only, and averages hide a wide range of outcomes. Treat it as a reason to watch failed breakouts, not as a crypto statistic.
FAQ
What is the difference between a fakeout and a bull trap?
A bull trap is a fakeout to the upside: price breaks above resistance, buyers jump in, and price falls back below. A bear trap is the same thing to the downside. "Fakeout" covers both.
How do you confirm a breakout?
Wait for a candle to close beyond the level, ideally by a margin scaled to volatility, and check that the higher timeframe and the volume don't argue against it. None of these makes a breakout certain, which is why the stop is planned first.
Are false breakouts more common in crypto?
We did not compare markets, so we can't say. What our test shows is that on Bitcoin's 4-hour chart, most first pokes through a 20-candle high or low did not hold for a day.
Should I enter on a stop order at the level?
A stop order at the level catches every breakout early, including every wick. In our test that rule had the highest false share, 78.5%. If you use it, accept that most fills will be fakeouts and size accordingly.
The bottom line
A false breakout is a break that doesn't hold, and on Bitcoin's 4-hour chart most first breaks don't. Judging the close instead of the wick, and asking for a margin scaled to volatility, removed a large share of the fakeouts in our test, at the cost of far fewer signals. The ones that remain are why every breakout trade needs its stop and size set before entry. To see these rules at work on real trades, look at how Navixa Price Action handles breakouts, or our step-by-step guide on how to trade Bitcoin price breakouts.
