What is a liquidity sweep?
A liquidity sweep is a quick move beyond an obvious high or low that sets off the orders resting there - stop-losses, breakout entries, liquidations - and is then followed by price coming back inside the old range. It looks identical to a breakout until it fails, so you can only name a sweep afterwards. Here is how it works, how to tell it from a breakout, and how often it actually happens in bitcoin.
In short: on 3,215 days from 2018 to October 2026, bitcoin went beyond the previous day's high on 47% of days and beyond the previous low on 43%. About half of those moves came back and closed inside the old range (49% for highs, 54% for lows); the other half closed beyond it as a breakout. Sweeps are common - and not reliably followed by a reversal.
Why price goes to these levels
- Stop-losses. Traders place protective stops just beyond obvious levels: sell stops under a recent low for longs, buy stops above a recent high for shorts. A triggered stop becomes a market order, so a cluster of stops is a pool of orders waiting to trade.
- Breakout entries. Other traders place buy orders above a high or sell orders below a low to catch a breakout. They add to the same pool.
- Liquidations. On leveraged markets, positions are forced closed at a known price; those forced orders are market orders too. The liquidation heatmap models where such levels pile up.
- Large orders need counterparties. A big buyer or seller can only fill size where enough orders sit on the other side. A pool of stops is liquidity - which is where the term comes from.
Buy-side and sell-side liquidity
Traders using this vocabulary (it comes from "smart money" and ICT-style trading) split it by direction:
- Buy-side liquidity sits above a high: buy stops of shorts and breakout buyers. A sweep above a high takes it.
- Sell-side liquidity sits below a low: sell stops of longs and breakdown sellers. A sweep below a low takes it.
- Equal highs or equal lows - several peaks or troughs at nearly the same price - are read as especially dense pools, because many traders put stops on the same obvious level.
Sweep, stop hunt, fakeout, breakout: what is the difference?
| Term | What it describes | Claims intent? |
|---|---|---|
| Liquidity sweep | Price goes beyond a level, takes the orders there, then returns inside the old range | No |
| Stop hunt | The same pattern, described as someone pushing price to trigger stops on purpose | Yes |
| Fakeout / false breakout | A breakout that fails and reverses | No |
| Breakout | Price goes beyond a level and stays beyond it | No |
| Liquidation cascade | Forced closes push price further, which forces more closes | No |
In practice sweep, fakeout and stop hunt describe the same chart. We use "sweep" because it names the pattern without claiming to know who caused it - we can count the pattern, not the motive.
How to tell a sweep from a breakout
You cannot tell at the moment price pokes through the level. After the fact, traders look at:
- The close. Closing back inside the old range within the same candle or the next few is the defining feature; closing beyond and holding is a breakout.
- Speed and size of the poke. Sweeps in our data usually go only a little past the level (below), while real breakouts run much further.
- Liquidations and volume. A burst of forced closes on the live liquidations page and a volume spike at the level are consistent with a sweep, but they also accompany real breakouts.
- Open interest. A sharp drop in open interest after the move means positions were closed or liquidated; a rise means new positions entered, which fits a breakout with follow-through. See open interest vs price.
All of these are clues read in hindsight. None of them tells you in advance which way the move will resolve.
How often does it happen? Bitcoin, 2018-2026
We counted a simple version on daily candles: did the day's high go above the previous day's high, and where did the day close? The same for lows. (Data: Binance BTCUSDT daily candles, UTC days, 1 January 2018 to 20 October 2026.)
| Level taken out | Days | Share of all days | Closed back inside (sweep) | Closed beyond (breakout) | Typical poke: sweep | Typical poke: breakout |
|---|---|---|---|---|---|---|
| Previous day's high | 1,502 | 47% | 49% | 51% | 0.65% | 2.5% |
| Previous day's low | 1,371 | 43% | 54% | 46% | 0.83% | 2.8% |
Three things stand out. Taking out yesterday's level is routine - it happens on close to half of all days - and 683 of 3,215 days (21%) stayed fully inside the previous range. Roughly half come back and half hold, so seeing a poke beyond a level tells you little about the close. And sweeps are shallow: the median sweep went about 0.6-0.8% past the level, against 2.5-2.8% for breakouts. That gap is partly built in: a day cannot close far beyond a level without first going far beyond it.
Descriptive counts, not a trading test: a level taken out on the day and a close are only known at the end of the day. 341 days took out both the high and the low and are counted in each row. Daily candles are a coarse view; the pattern traders mean usually plays out within hours.
Does a sweep predict a reversal?
This is the claim that makes sweeps interesting to trade, and the one we can test. Two things we have measured:
- Is price drawn to liquidation clusters? In our pre-registered test, price touched the side with the bigger modeled cluster first in 55.2% of 105 events over 30 days (p 0.165) - inside what random sides produce. The forward test on new data is still running: liquidation heatmap test.
- Do extreme crowd positions reverse? Shorting very high funding, with or without rising open interest, and shorting new listings after day one all showed the same pattern: frequent small wins, occasional huge losses, average near zero. See all tests in one table.
The deeper problem is hindsight. A sweep is defined by the return, so every example you see on a chart is a case that already reversed. The cases where price took the level and kept going are called breakouts and do not make it into screenshots of sweeps.
Common mistakes
- Calling it a sweep while it is happening. Entering against the move at the poke is a bet that this one will fail; half of them do not.
- Treating any wick as a sweep. Without a clear level where orders plausibly sit, a wick is just a wick.
- Ignoring leverage. The traders caught in a sweep are often those with tight stops or high leverage - at 50x a bitcoin long has about 1.6% of room, and a poke of 1% beyond a level - the size of a typical sweep in the table above - uses most of it. How often it happened is on the leverage test.
- Trusting a single timeframe. A sweep of a 5-minute low can be noise on the 4-hour chart.
Where to see liquidity on live data
- Liquidation heatmap - modeled levels where leveraged positions would be forced out.
- Order book heatmap - real resting limit orders, which is the literal "liquidity" on the book.
- Live liquidations - what has actually been forced out, longs versus shorts.
- Glossary: liquidity sweep and other terms.
FAQ
What is a liquidity sweep in trading?
A quick move of price beyond an obvious level, such as a previous high or low, that triggers the orders resting there (stop-losses, breakout entries, liquidations) and is then followed by price returning inside the old range. It can only be named a sweep after the return: while it is happening it looks exactly like a breakout.
Is a liquidity sweep the same as a stop hunt?
In everyday use, yes. "Stop hunt" implies someone deliberately pushes price to trigger stops; "liquidity sweep" describes the same price pattern without claiming intent. We can measure the pattern, but not who caused it.
How often does a liquidity sweep happen in bitcoin?
Taking out the previous day's high or low is routine: on 3,215 days from 2018 to October 2026, bitcoin went beyond the previous day's high on 47% of days and beyond the previous low on 43%. About half of those moves closed back inside the previous day's range (49% for highs, 54% for lows) and half closed beyond it as a breakout.
Does a liquidity sweep predict a reversal?
Not in anything we have tested. A sweep is identified only after price returns, so it is a description of what happened. In our pre-registered test of whether price is drawn to liquidation clusters, the first 30 days gave 55.2% of 105 events (p 0.165), no better than chance, and the forward test is still running.
How do you tell a sweep from a real breakout?
Only by what price does next: a sweep closes back inside the old range, usually within the same candle or the next few; a breakout closes beyond the level and holds it. Candle close, speed, volume and liquidations are clues, but none settles it in advance.
What is buy-side and sell-side liquidity?
Buy-side liquidity is the cluster of buy orders above a high, mostly stop-losses of shorts and breakout buyers. Sell-side liquidity is the cluster of sell orders below a low, mostly stop-losses of longs and breakdown sellers. Sweeping above a high takes buy-side liquidity; sweeping below a low takes sell-side liquidity.