Liquidity sweep (stop hunt)

How it works

A liquidity sweep starts with orders that are easy to predict. Traders put stop losses at the places that invalidate their trade: just above resistance, just below support, past the edge of a range. Breakout traders put entry orders in the same places. The result is a cluster of resting orders at levels every chart reader can see.

A stop loss is a dormant market order. When price trades through its trigger, it becomes a market buy (for a short's stop) or a market sell (for a long's stop) and fills at whatever the book offers. So when price pushes a little past a visible high, three things fire at once:

• Buy stops from shorts who placed their invalidation above the high.
• Breakout buy orders from traders waiting for the level to break.
• Liquidations of leveraged shorts whose liquidation price sat just above the high.

All of that is buying, and it arrives in a burst. Larger sellers need exactly that burst. Selling a large position into a quiet book pushes price against you; selling into a wave of forced and eager buying fills the order with little slippage. Once the cluster has been consumed, nothing is left to keep buying, and price falls back below the level. That poke-and-return is the sweep. The diagram above shows the shape: price pierces the prior high, the resting orders trigger, and the candle closes back underneath.

Traders use three names for the same event. Liquidity sweep is the neutral, structural term. Stop hunt is the older term and implies someone targeting the stops. Liquidity grab is the term most common in ICT (Inner Circle Trader) and smart-money material, where the grab is treated as the step before the real move.

ICT splits resting orders into two pools:

• Buy-side liquidity sits above swing highs, equal highs and range tops: buy stops from shorts plus breakout buy orders. A sweep of it is a buy-side sweep, and it usually leaves a wick on the high.
• Sell-side liquidity sits below swing lows, equal lows and range bottoms: sell stops from longs plus breakdown sell orders. A sell-side sweep is the mirror image, with a wick on the low.

The mechanism needs no villain. In deep markets like BTC and ETH, no single actor can steer price to a precise level on demand. Price travels to where orders are because that is where size can be traded.

A worked example

BTC has ranged for several days on the 4-hour chart and printed two highs at almost the same price, $62,380 and $62,410. Equal highs are among the clearest stop clusters on a chart. Average 4-hour volume in the range is 1,200 BTC. (The numbers are illustrative.)

Two possible outcomes from the same approach to $62,410:

The breaching candle looks identical in both cases: a push through the level on a volume spike, because the stops fire either way. The difference is what happens after. In the sweep, volume collapses to 0.75× average and the close sits back inside the range. In the breakout, price closes above the level and the retest holds.

Now the cost side. A trader shorted at $61,800 inside the range and risks $500 per trade.

• Stop just above the equal highs at $62,450: distance $650, so position size is $500 ÷ $650 = 0.769 BTC. The sweep high of $62,780 triggers the stop. Loss: $500, and the stop itself was one of the buy orders that fed the sweep.
• Stop beyond the likely sweep zone at $62,900: distance $1,100, so size is $500 ÷ $1,100 = 0.455 BTC. The sweep does not reach it. If price then returns to a range low of $60,800, the gain is 0.455 × $1,000 = about $455.

The wider stop did not raise the risk per trade. It cut the position size by 41% so the same $500 covered a stop outside the obvious cluster. If the move had been a real breakout to $62,950, the wider stop would also have been hit, for the same $500.

What it tells you

A confirmed sweep tells you that the orders at a visible level have been used up and that the move through the level did not attract follow-through. The checks below separate a sweep from a breakout. They work best in this order.

1. Start on the higher timeframe. Mark the swing high or low on the 4-hour or daily chart first. A wick on the 15-minute chart that does not touch anything visible on the 4-hour is usually noise.
2. Read the close, not the wick. A wick can stretch on a thin book for many reasons. The close back inside the prior range is what makes the event a sweep.
3. Compare volume before and after. Both a sweep and a breakout spike volume at the breach, because stops fire either way. A breakout keeps volume elevated; a sweep sees it fall away as price retreats.
4. Watch the retest. After a breakout, the old high tends to act as support on the first pullback. After a sweep, price fails to get back above it.
5. Check the trend. A buy-side sweep inside a downtrend tends to be followed by continuation lower. A sell-side sweep at the end of a long range can mark the start of a move up. The sweep is the event; the market regime is the context that says what it means.

The common errors are the inverse of these checks: calling every wick a sweep, entering on the wick itself before the close, fading a slow grind through a level that keeps closing above it, and chasing the first impulsive candle after the reversal instead of waiting for a pullback.

In ICT terms, the sweep is the "manipulation" phase of the accumulation, manipulation, distribution cycle. The word labels a phase of price action rather than accusing anyone. ICT traders also watch for a fair value gap (FVG), a three-candle imbalance that a fast reversal often leaves behind, as a retest zone after the sweep.

When it matters most

At the most visible levels. Previous swing highs and lows, equal highs and equal lows, range edges, and daily and weekly opens collect the most stops, so they produce the clearest sweeps. If you have to argue that a level was important, the sweep probably was not.

At session handoffs. Crypto trades around the clock but still inherits rhythm from traditional markets. The Asian session often leaves a narrow range with obvious stops on both sides, and it is frequently swept when London or New York opens.

On the timeframe of the level. A sweep matters in proportion to the level it takes. A weekly high swept over several candles can mark the top of a multi-month leg. A daily high usually shows up as one wick on one candle. On the 15-minute chart sweeps happen constantly, and one that runs against the daily trend rarely delivers more than a small reversal. The 4-hour chart is a reasonable middle: it shows session-level sweeps without drowning in micro-wicks.

In thin books and crowded leverage. Altcoins with shallow order books produce violent sweeps with wicks several times the size of the candle body, because few resting orders absorb the push. When open interest is high, liquidations add to the stops, and a sweep can turn into a short liquidation cascade before it reverses.

Common misconceptions

"A stop hunt means someone is targeting my stop." Your stop sits where thousands of similar stops sit, because everyone uses the same chart logic. Price goes to the cluster because that is where large orders can be filled, not because anyone sees your individual order.

"Every wick through a level is a liquidity sweep." A sweep needs a meaningful level and a close back inside the range. A wick that tags nothing in particular, or a move that closes beyond the level and holds the retest, is something else.

"Sweeps mean you should stop using stop losses." The lesson is about placement and size. A stop at the most obvious level sits inside the cluster most likely to be swept. Placing it beyond the sweep zone and reducing size, as in the example above, keeps the risk per trade the same.

"A sweep always reverses the trend." A sweep removes the orders at one level. What price does next depends on the higher-timeframe trend; a sweep against the trend often produces only a short reversal before the trend resumes.