Liquidity sweep and liquidity grab

The reserve: buy side and sell side liquidity

Above every visible high, buy orders accumulate: the stops of short sellers, and the entry orders of those waiting for the breakout. That reserve is called buy side liquidity. Below every visible low, its mirror: the stops of buyers and the breakdown sells, the sell side liquidity.

The more obvious the level, the bigger the reserve. Equal highs, two tops at roughly the same price, form a line everyone sees, hence a deposit everyone feeds. Same for equal lows, and for a trendline touched three times: trendline liquidity is the same idea laid on a diagonal.

The vocabulary pushes the idea one step further with engineered liquidity: levels so clean they seem built to make people park stops. You do not need to attribute that intention to anyone to use the concept: it is enough to notice that stops get parked at the obvious places, because that is where the manuals say to put them.

⚠️ None of this shows in an order book. Stops are not displayed, the reserve is an inference, not a measurement. What can be measured is what price does when it crosses the level, and that is the whole point of what follows.

Sweep, grab, raid, stop hunt: four words, one event

The pass that triggers the reserve goes by several names, and definitions vary from one teacher to the next, which is worth knowing before comparing two videos. Liquidity sweep and liquidity grab are most often synonyms: price clears the extreme, takes the stops, then comes back inside. Liquidity raid and stop hunt say the same thing with more theatre.

When a distinction is drawn, it is usually this one: the grab names the taking itself, often a fast wick that comes straight back, while the sweep can name a wider clearance, including on a close, which may or may not come back inside. The border is blurry, and no authority fixes it.

The same picture carries yet other names outside ICT vocabulary: the swing failure pattern, shortened to SFP, among volume readers, or the smart money reversal elsewhere. Four schools, one observable event: an extreme cleared, stops filled, and a price that does not hold beyond.

The way out of the naming trap is always the same in this course: pick one definition, write it down, and apply it as written every day. The one kept here: extreme cleared, then a close back inside within three candles.

What validates a sweep: displacement

A clearance alone proves nothing: half the wicks of any day poke past some extreme. What turns the sweep into information is the move that follows the other way, and it has its own lesson: displacement, the impulse that breaks structure and leaves a fair value gap.

The complete sentence therefore reads in order: a reserve identified in advance, a pass that takes it, an impulse that uses it. Remove any of the three and there is nothing left to trade: a reserve never taken is just a line, a pass without an impulse is just a wick, an impulse without a taken reserve has no identifiable fuel.

It is also the second act of the accumulation, manipulation, distribution sequence: what that sequence calls manipulation is precisely a sweep, and what it calls distribution is the displacement that follows it.

How to identify a liquidity sweep, step by step

The written rule, extreme cleared then a close back inside within three candles, unfolds into five checks. The first is done before price gets anywhere near the level, and that is what separates identifying a sweep from recognising one in hindsight.

1. Mark the level in advance. Keep to the extremes other people can see without zooming: the previous day's high and low, the high and low of the session that just ended, equal highs or equal lows, the last swing on your decision timeframe. If you have to hunt for the level after the wick, it was not a reserve, it was a coincidence.

2. Wait for the clearance. Price has to trade beyond the level, not stop at it. A touch to the tick is a test of the level: the stops resting behind it were never triggered.

3. Read the close, not the wick. Within three candles of the clearance, one of them must close back on the original side of the level. Until that happens you are looking at a possible break, and nothing on the chart tells the two apart yet.

4. Ask for the move the other way. A close back inside followed by drift is a wick with a story attached. On a lower timeframe, the displacement that confirms the sweep shows as the break of the last minor high or low, the change of character from the market structure lesson.

5. Name the invalidation. The extreme of the sweep is the line. If price trades back beyond it, the stops were not the fuel of a reversal, they were the first orders of a break, and the reading is over.

A worked example, in pips

The numbers below are illustrative, built to show the arithmetic, not taken from a real session. EUR/USD, 15-minute chart. During the Asian session price prints two highs at 1.0849 and 1.0850, equal highs one pip apart, and a session low at 1.0802. You mark 1.0850 before London opens.

Shortly after the London open a candle trades up to 1.0857, 7 pips through the level, and closes at 1.0846 with its low at 1.0844: clearance and close back inside on the same candle. The next candle falls from 1.0846 to 1.0826, and the one after it never trades above 1.0830. Between 1.0830 and 1.0844 the move has left a fair value gap of 14 pips. Checks one to four are met.

A plan built on it could read: sell at 1.0840 if price returns into the gap, stop at 1.0859, two pips above the sweep's high, target at 1.0802, the sell side liquidity under the Asian low. The risk is 19 pips, the target 38 pips away, twice the risk before spread and commission. With 100 dollars at risk and a pip worth 10 dollars per standard lot on EUR/USD, the size is 100 / (19 × 10) = 0.52 lot, rounded down, which the position size calculator works out for any pair.

What the example cannot show is the other branch. Had that candle closed at 1.0855 and the next two held above 1.0850, the same 7-pip poke would have been a breakout, and the rule would have produced no trade at all. Keeping you out of that one is half of what it is for.

Internal, external: which liquidity price is aiming at

The vocabulary also files liquidity by where it sits relative to the current bracket, the dealing range, the span between the last swept high and the last swept low. The extremes of that bracket carry the external range liquidity: the stops beyond the high and the low. Inside it, the fair value gaps and order blocks left by the moves form the internal range liquidity.

The reading built on top is called the draw on liquidity: the idea that price travels from one reserve to the next, from external to internal and back, like a pendulum. It is a seductive reading grid, and it deserves the same treatment as the rest of the corpus: a hypothesis to count on your market, not a law.

⚠️ The trap of the draw on liquidity is the same as the three-act sequence's: in hindsight, price always went to 'fetch' some liquidity, since there are extremes and gaps everywhere. The testable version is written beforehand: which reserve, at what price, and what would invalidate the reading.

The liquidity void

One last word of the same family remains, often confused with the fair value gap: the liquidity void. The void names the entire one-sided traversal, the whole zone crossed without balanced trade. The fair value gap is the precise three-candle window cut inside it.

The confusion is harmless, many use the two words interchangeably. The useful nuance is elsewhere: a whole void can contain several measurable FVGs, and it is on the FVGs, not the void, that the return rules from the fair value gap lesson apply cleanly.

Equal highs and lows: tolerance decides everything

Equal highs and lows, two extremes parked at the same price, form the cleanest reserve on the chart: the line shows without a tool, so everyone deposits there. The word “same” is the weak point of the definition though. Two highs almost never land on the same tick, and no school fixes the tolerance allowed.

That tolerance alone decides how many levels you will see. On a futures contract, two highs two ticks apart are equal highs for everyone; fifteen ticks apart, they are equal highs for no one. Between the two, everyone chooses, and the choice changes the count threefold. The honest way to settle it is to fix the tolerance once, as a fraction of the session's average range, a tenth for instance, then never touch it again.

The classic mistake is widening the tolerance after the fact, once price has come back to fetch the level. The chart then fills with retroactive equal highs, the reading becomes unverifiable, and the statistic pulled from your journal measures nothing but your tolerance of the day.

The same rule serves on the other side of the trade: a stop backed onto equal lows spotted with a loose tolerance is parked inside the deposit, not behind it. The lesson on liquidity and stops covers that placement in detail.

The swing failure pattern: the same picture read without ICT

The swing failure pattern, shortened to SFP, comes from volume readers rather than the ICT corpus, and its definition is stricter than the sweep's. It fits in one sentence: a candle clears a prior swing high and closes below that high. At the bottom, the exact mirror, a low cleared and a close above it. The clearance and the return are judged on the same candle, on the timeframe where you spotted it.

The difference with this lesson's definition is sharp: the swing failure pattern is validated by the close alone, without waiting for displacement. It therefore keeps far more cases, some of them only wicks with no follow-through, where the ICT definition rules out more and misses a few along the way. Neither is truer than the other, they simply do not count the same population.

Some add a measurable condition to it: the clearance trades a lot for little ground gained, volume without progress. That condition only holds where volume is real, on futures and on stocks. On spot forex, the volume displayed is your broker's and not the market's, and the condition loses its meaning.

A whole setup rests on this picture, the turtle soup, described in the lesson on the ICT setups. Whatever the school, the rule does not change: one definition, written down, and logged in the journal, so the statistic covers one population and not two.

Four mistakes that turn a sweep into a loss

Entering on the wick. Selling the clearance itself, before any candle has closed back inside, is trading a picture that a sweep and a breakout share. The entry is better by a few pips when it works, and when it does not you are short at the first price of a real break.

Fading every new high in a trend. In an uptrend each higher high clears the previous one by definition: that is the trend continuing, not a reserve being raided. A sweep against the trend of your decision timeframe has to earn more, a close back inside and displacement, than one that goes with it.

Putting the stop inside the wick. A stop at the level, or halfway up the wick, makes the ratio look better and sits exactly where an ordinary retest goes. The invalidation is the extreme of the sweep; anything tighter is a second bet on the retest staying shallow.

Counting every level. A swing high on the one-minute chart and the previous day's high are not the same object, and a chart marked with twenty levels will show a sweep on almost any wick. Decide which families of levels you count before the session, and keep the list short enough to be written on one line.

Count it in your journal

All this vocabulary boils down to one testable question: do entries taken after a validated sweep behave better than the others? Tag the sweep as a condition in your journal, the one from the definition written above, and let the number answer.

Fifty cases are enough for a first reading, and a properly kept journal does the sorting on its own if the condition is logged on every trade. If the edge exists on your market at your hours, it will survive the count. If it does not, you just saved yourself years of conviction.

Frequently asked questions