AlphaFlow AI › Liquidity sweeps: telling a stop-run from a breakout
Liquidity sweeps: telling a stop-run from a breakout
A sweep and a breakout look identical while they are happening. Both push through a level everyone can see; only one of them comes back. This is how to read the difference, why the distinction is the whole trade, and why waiting is not the same as hesitating.
Liquidity is other people's stops
"Liquidity" sounds abstract until you replace it with what it actually is: orders sitting at a price, waiting to be filled by someone willing to take the other side. In a retail chart the largest and most predictable cluster of those orders is stop losses.
Think about where stops go. Someone long puts theirs below a recent low. Someone short puts theirs above a recent high. Nobody chooses an arbitrary price — they choose the level the chart makes obvious. Which means everyone's stops end up in roughly the same few places, and those places are visible to anyone looking at the same chart.
A stop loss is not a passive instruction. A stop on a long is a sell order, and it executes the moment price touches it. So a cluster of stops below a low is a cluster of sell orders — and if you needed to buy a large position, that cluster is precisely where someone will sell to you in size.
The whole idea in one sentence. Price is often drawn toward obvious levels not because the level predicts anything, but because that is where the orders are — and orders are what large participants need in order to transact at all.
Where liquidity actually pools
Not every level holds meaningful liquidity. The ones that reliably do share a property: they are obvious to a lot of people at once.
- Equal highs and equal lows. Two or more touches at roughly the same price. The second touch is what makes it obvious — one high is a high, two at the same level is a line, and lines attract orders.
- Prior session highs and lows. The high and low of yesterday, of the Asian session, of the week. Widely watched, widely used as stop placement.
- Swing points. The high or low that defined the last leg. Anyone who entered on that leg is using it as their invalidation.
- Round numbers. Not technical in any sense, but people place orders at them anyway, which is what makes them real.
What these have in common is not that they are special prices. It is that a large number of independent participants arrived at the same price for the same reason. A level nobody can see holds nothing worth taking.
The anatomy of a sweep
A liquidity sweep is a specific sequence, and the order matters:
- Price approaches a level that a lot of people are watching.
- It trades through the level — far enough to trigger the orders resting beyond it.
- Those orders fill, providing the volume someone needed.
- Price closes back on the original side of the level.
- The move then travels in the opposite direction to the push that took it.
On the chart, that prints as a candle with a long wick through the level and a body that closed back inside it. The wick is the part that matters: it is the visible record of price having gone somewhere it did not stay.
The reversal afterwards is what makes it a sweep rather than a failed poke. A level can be tagged and rejected without a meaningful move following, and in that case nothing useful happened — the orders may simply not have been there.
Sweep or breakout? The distinction is the whole trade
Here is the difficulty nobody teaching this can remove: while price is beyond the level, a sweep and a breakout are the same picture. Both have pushed through. Both look decisive. Both will have people entering in the direction of the push.
The difference is what happens next, and it is a difference in where the candle closes, not in how far the wick went.
- A sweep takes the level with a wick and closes back inside. The push was the taking of orders, not the beginning of a move.
- A breakout takes the level and closes beyond it, then continues. The level genuinely broke; what was resistance is now support.
That is why so much SMC teaching insists on waiting for a close. It is not superstition and it is not caution for its own sake — the close is the only thing that separates the two cases, and until a candle closes there is genuinely no information distinguishing them.
What this costs, stated honestly. Waiting for the close means you never get the very best price, and it means some real sweeps will have moved away before you can act. That is the price of the distinction. Anyone claiming to identify a sweep before the candle closes is either using a smaller timeframe — where the close has already happened — or guessing.
Why the same candle can be either
A detail that catches people: whether something is a sweep depends on which timeframe you are looking at. A four-hour candle with a long upper wick is one sweep. Inside it, on the fifteen-minute chart, there may have been a clean break upward, several candles closing above the level, and then a reversal — which on that timeframe was a breakout that failed.
Neither reading is wrong. They are descriptions at different resolutions of the same events. But it means "was that a sweep?" is an incomplete question — the complete one is "was that a sweep on the timeframe I am trading?"
It also explains a common frustration: two traders arguing about whether a level was swept, both looking at charts, both correct, because neither has said which chart.
The four mistakes that cost the most
Treating any wick through a level as a sweep. Price wicks through levels constantly, often for no reason more interesting than a thin moment in the order book. A wick is necessary but nowhere near sufficient.
Trading the sweep of a level nobody was watching. If the level is not obvious, there is no reason to think orders were resting there — and no reason to think the push was about taking them. The obviousness of the level is part of the setup, not decoration.
Ignoring the direction of the larger move. A sweep gives a reason to expect a reversal of the push that took the level. That is not the same as a reason to expect a reversal of the trend. Fading a sweep against a strong prevailing direction is a different and much harder trade than it appears.
Placing the stop inside the sweep's own range. This one is almost self-inflicted. If the idea is that price ran beyond a level to take orders and came back, then the extreme of that run is the invalidation. A stop placed nearer than the wick will be taken out by the very move being faded — and often by the same candle.
Using it without fooling yourself
A sweep on its own is an observation, not a trade. What turns it into one is the same thing that turns any observation into a trade: a level you can state in advance, an invalidation you can live with, and a reason to think the move has somewhere to go.
The useful questions, in order:
- Was the level obvious? Could a reasonable trader have drawn it before the sweep happened, not after?
- Did price close back inside? On the timeframe you are trading, not on a smaller one chosen after the fact.
- Where is the invalidation? Beyond the extreme of the run, not inside it.
- Where is it going? A fade with no target is a hope. The opposing liquidity — the next obvious level in the other direction — is usually the honest answer.
If any of those has no answer, the observation was interesting and the trade was not there. That is a normal outcome and recognising it is most of the skill.
The short version
Liquidity sweeps are the clearest idea in Smart Money Concepts because they describe something mechanical rather than intentional: orders rest where the chart makes them obvious, and price sometimes goes to get them.
The difficulty is never spotting the wick. It is that the wick looks the same whether the level held or broke, and the only thing that separates the two is a candle closing — which by definition you cannot have while you are still deciding.
Most of what people call a sweep that failed was a breakout they read early. Knowing the difference will not make every trade work. It will stop a specific expensive mistake, repeatedly, which is worth more than it sounds.
undefined
AlphaFlow AI ·
Terms · Privacy ·
Risk Disclosure