Price clears an obvious high, then immediately turns around. That behaviour goes by several names — liquidity sweep, liquidity grab, stop run — used interchangeably, sometimes wrongly.
This guide clarifies the vocabulary, breaks down the mechanism, and covers the question that matters in practice: how to tell a liquidity take from a genuine breakout — and which parts of that test only arrive once the move has developed.
Three terms, one mechanic
All three describe the same underlying phenomenon: a price move that reaches resting orders and activates them. The differences in usage are worth knowing.
Liquidity sweep — The broadest term. Price passes through a liquidity zone and activates its orders.
Liquidity grab — Emphasises the brief and targeted nature. Price reaches a specific pocket, clears it, and comes straight back.
Stop run — Emphasises the type of order activated: specifically stop-losses. The term points to a cascade: the first stops triggering set off others.
One thing to set aside immediately: this is not a single actor targeting your position. It is an aggregation effect. Many traders place stops at the same visible levels, creating pockets of counterparty that a large order has a mechanical reason to reach. That framing governs the rest of this guide: where a sentence below reads as intent, it is shorthand for the aggregation effect and for a reading of price, not a claim about anyone's purpose. Nor is any of it directly observable: a retail chart shows price, not orders, and every statement below about stops, pockets or counterparty is read from price after the fact.
Why the move exists: the need for counterparty
A large order cannot fill at the current price: there is not enough counterparty. Sent all at once, it consumes several levels of the book and pushes price against itself. That book model comes from centralised venues: in spot FX, depth is fragmented across bank feeds and is not observable from a retail platform. It works as a model; you should not expect to see it.
To sell in size, then, you need buyers. Where do you find a concentration of them? Above an obvious high, where short sellers' stops — which are buy orders — are waiting to trigger.
Hence the apparent paradox: a rise into a stop cluster is where size can be sold. The move and the selling coincide; this guide does not claim the move was made in order to sell.
Inducement: the trap before the trap
Inducement is easy to misread, and misreading it turns a normal stage of the sequence into an apparent failure.
Definition
Inducement is a move that encourages traders to position themselves, leaving a pocket of resting orders between price and the larger level.
In other words: before a large pocket is reached, a smaller pocket frequently forms between price and that level, and is cleared first.
How it plays out
- Price approaches an important zone — say a set of equal highs.
- Before reaching it, it forms a small intermediate low.
- That low attracts buyers, who place stops just beneath it.
- Price dips below that low, triggering them — that is the inducement.
- Then it rises for the real target.
Why it matters
Without this lens, step 4 looks like the setup failing. With it, it looks like a normal stage of the sequence.
A practical cue: inducement sits between your entry and the real target. If a small low sits below your entry, treat it as a level that can be cleared before the move you are waiting for.
It is also why placing your stop just below the nearest low means placing it where price is frequently traded through. Place it beyond the extreme printed by the sweep itself instead, with a buffer sized from the instrument's recent range — a fraction of ATR(14) on the timeframe where the level was identified, fixed before you trade. And a stop is a market order, not a guaranteed price: it fills at the next available price, so the loss taken can exceed the distance you measured.
Turtle Soup: the reversal of the breakout
Turtle Soup predates SMC vocabulary and describes the same thing.
The origin
The name is a jab at the “Turtle Traders”, famous for a breakout method: buy above the 20-day high, sell below the 20-day low.
Turtle Soup takes the other side of those breakouts. Fix one lookback and keep it: the 20-period high or low on the timeframe you trade, as labelled in the diagram below. If price clears that extreme and closes back inside on that same candle, you position in the direction of the return; a return arriving two or three candles later is a different sequence.
Its modern translation
The reasoning is identical to the liquidity sweep, phrased differently:
- Breaking an extreme attracts breakout orders and triggers stops.
- If price cannot hold beyond the level, the breakout has failed.
- A failed break of a level many traders watch can be followed by a fast move the other way; the usual explanation is unwinding of positions taken on the break, though that positioning is inferred, not observed.
The condition is strict: without the return inside, there is no Turtle Soup — there is a breakout that is working.
That lineage matters: the behaviour described is not a recent invention of an online community. The same setup was published under the name Turtle Soup in the 1990s.
Displacement: what separates signal from noise
A sweep without displacement is worth little.
Definition
Displacement is a fast directional move following the liquidity take. Read on closed candles, the test is one criterion:
- two to three consecutive candles in the same direction, each with a body of at least 70% of its range and a range of at least twice the median range of the previous 10 candles.
A fair value gap — an imbalance across three consecutive candles — and a break of structure in the direction of the move can accompany displacement, but neither belongs to the test: candles that meet the criterion above pass it with or without them.
The 70% and the twice-the-median multiple are illustrative for the calculation, not recommended settings: fix your own thresholds on your instrument before trading, and apply the same ones every time. Judged without a number set in advance, speed and size can only be assessed once the move is over.
What it means
Displacement is what an imbalance between buyers and sellers looks like on the chart. The order flow behind it is inferred from the price move, not observed.
The decisive comparison:
- Price clears the high, drifts back slowly, chops sideways → no displacement.
- Price clears the high, rejects, then prints candles that pass the test above → displacement.
Genuine breakout or false break: the test that settles it
This is the question that decides whether the reading is usable at all.
The central criterion: the close
A breakout confirms with a close beyond the level, on the timeframe where the level was identified; a liquidity take breaks on the wick and closes back inside.
| Criterion | Genuine breakout | Liquidity take |
|---|---|---|
| Available at the close of the breaking candle | ||
| Close vs level | Beyond | Short of it: return inside |
| Candle shape | Large body beyond the level | Long wick, small body |
| Only confirms afterwards — cannot inform the entry | ||
| What follows | Consolidation above, then continuation | Fast rejection, move the other way |
| Retest | Level becomes support/resistance | Level is not retested from beyond |
| Tick volume / pace (tick count from your broker's feed, not traded volume) | Sustained | Brief spike, then fades |
| Structure | Break of structure confirmed | Structure preserved the other way |
Classic traps
Judging on the live candle. An unclosed candle tells you nothing. Until the candle closes, a break is provisional — the same price action can end beyond the level or back inside it. Waiting for the close removes the most common source of false readings: a break that exists only mid-candle.
Mixing timeframes. A close beyond on M1 is not a close beyond on H1.
Forgetting the wider direction. A bullish sweep against the direction of the H4 trend is not read the same way as one with it.
The full sequence, in four acts
The progression this framework describes, in order:
Act 1 — Build-up. Price ranges. The highs and lows it leaves are the levels this framework assumes stop orders sit behind. Nothing to do here but mark them.
Act 2 — Inducement (optional). An intermediate move draws participants in and leaves a smaller pocket between price and the target, cleared first.
Act 3 — Liquidity take. Price clears the target zone. The break does not survive the candle: the close returns inside.
Act 4 — Displacement. The move starts in the opposite direction, with candles passing the test above and possibly a fair value gap.
Acts 3 and 4 are inseparable. The liquidity take alone is not a signal. It is the combination of rejection and displacement that gives the pattern its informative value.
A counter-case, since no diagram above draws one: the sequence can complete — close back inside, marked rejection, displacement — and price can still trade back above the extreme printed by the sweep itself, which invalidates the reading.
Confirmation signs after a sweep
What to look for once the pocket is taken, in order of importance:
- A close back inside the level — a necessary condition.
- A marked rejection candle — a wick on the side of the take at least twice the length of the body.
- Displacement — candles that pass the test above, on the thresholds you fixed in advance.
- Break of structure in the direction of the reversal — an intermediate low broken after a bullish sweep, for instance.
- Fair value gap left by the displacement — it offers a potential return zone.
- No return beyond the extreme printed by the sweep itself — if price trades back above that extreme, the reading is invalidated. This test is deliberately tick-based while confirmation is close-based: any trade beyond that extreme ends the reading, and the position is closed there rather than left to run to the stop, which sits a buffer further away and only catches the exits that are missed.
Point six is your invalidation. A framework with no defined invalidation is not a framework: it is an opinion.
Where the sweep disappoints
It is worth knowing which conditions this reading describes worst.
In very low volatility. In a tight, low-volume range — mid-Asian session, the eve of a public holiday — price crosses and re-crosses the same levels without any imbalance forming. Every break looks like a sweep; none leads to displacement. Spreads are also at their widest in these hours.
On a major economic release. An unexpected figure produces a move driven by information, not by execution mechanics. Price can cut through several pockets in a row without ever rejecting. The framework still holds, but its explanatory power drops sharply. Spreads widen around the release as well.
Against a strong higher-timeframe trend. A bullish sweep inside a bearish daily trend can give a short-lived reversal that the underlying direction absorbs quickly.
On timeframes that are too low. On M1, small levels are broken and reclaimed continuously. The sequence is there, but drowned in invalid setups that are hard to filter by hand.
On illiquid instruments. The logic depends on a concentration of orders existing. On an exotic pair or a thinly traded stock, that concentration is weak and the behaviour becomes erratic.
Sweeps of several pockets at once
Price does not always take one pocket; it sometimes clears several in a single move.
Stacked pockets. When equal highs, a previous day high and a session high all sit within one ATR(14) of each other, measured on the timeframe where the levels were identified, a single push can take all three. The move looks disproportionate for the level involved — because it is servicing three levels, not one.
What that changes. A sweep that clears stacked levels has more room before the next opposing one. Whether more counterparty was released is inferred from the move, not observed.
What it does not change. The validation criterion is identical: a close back inside the highest level cleared. Taking three pockets and closing beyond all of them is still a breakout, not a sweep.
Practically, levels inside that one-ATR tolerance are worth marking as a single band rather than separate lines. It keeps the chart readable and reflects how price treats them.
What this model does not do
For methodological honesty:
- It does not predict the future. It describes a pattern, not a rule price is obliged to follow.
- Not all breakouts are traps. Genuine breakouts exist, and identifying them as such is part of the method.
- A perfectly formed sweep can fail. Risk management remains essential regardless of conviction.
- This model does not replace position management, sizing, or execution discipline.
Key takeaways
- Sweep, grab and stop run describe the same mechanic: activating resting orders.
- A large order needs counterparty; stop clusters supply it.
- Inducement leaves a smaller pocket between price and the target, cleared first — where a stop placed just below the nearest low would sit.
- Turtle Soup named the same behaviour long before SMC vocabulary: break an extreme, fail, reverse.
- Displacement separates a tradable sweep from a mere break.
- A breakout confirms with a close beyond the level, on the timeframe where the level was identified; a liquidity take breaks on the wick and closes back inside.
- Always define your invalidation before entering — and remember a stop is a market order, not a guaranteed price.