“Liquidity sweep” gets used as a synonym for “reversal” — price pokes past a level, and the level gets credited with the reversal that follows. That’s backwards. A sweep only describes what happened to the level; what happens to price next is a separate question, and treating the two as one thing is why the same setup gets called a clean sweep-and-reverse in hindsight and a stop-out in real time.
Why equal highs and lows exist in the first place
A swing high or low isn’t just a shape on the chart — it’s where resting orders accumulate. Breakout traders place stop-entries just beyond it; traders who faded the first test place protective stops just beyond it too. When price fails at the same price twice, both groups add to the same shelf instead of moving on, because the level has now been “confirmed” by two touches in the crowd’s eyes. That’s how equal highs and equal lows form: not a coincidence of geometry, but a record of where orders have been layering up. The pattern says where interest is resting. It says nothing yet about which way that interest resolves.
What a sweep actually is — and isn’t
A sweep is a level that price trades through and then, on a confirmed basis, fails to hold beyond. The word doing the work there is confirmed: judging a sweep from an intrabar wick is the same mistake covered in what repainting actually is — a forming bar can poke through a level and pull back before the close, or it can poke through and keep going, and from inside that bar the two look identical. The only honest way to classify it is to wait for the bar that took the level to close, then watch what the next bars do. A sweep and a genuine breakout start with the exact same candle. They are told apart by what happens after it, never by the wick itself.
Swept is not resolved: armed → swept → accepted or rejected
A cleaner way to track this is as a small state machine rather than a single label. A level sits armed once it’s formed and hasn’t been touched. It becomes swept the moment a confirmed close trades through it. From there it goes one of two ways: rejected, if price closes back inside the range shortly after — the classic sweep-and-reverse — or accepted, if price keeps closing beyond it, which means the “sweep” was a real break wearing a sweep’s clothes for a candle or two. Our Session Liquidity Architecture runs exactly this machine on session ranges and prior-day highs/lows, confirmed-bar only, so the state on the chart is never a guess mid-formation — it only updates when a bar actually closes.
The pools price gets drawn toward
Equal highs and lows are the most visible liquidity pool, but not the only one. Old session extremes, unfilled order blocks and fair value gaps all leave behind price areas where the last participants to trade there are now underwater or under-filled — which is a second, quieter kind of magnet. Liquidity Map & Order Blocks clusters these into pools (ATR-normalized, so a pool means the same thing on a 5-minute chart and a daily one) and scores which pool price is more likely to path toward next from structure, proximity and whether anything sits in the way. That score is a draw, not a destination — it’s muted whenever the underlying read is low-confidence, and it describes a tendency in the current structure, not a promise about where price is going.
What we actually measure, and why it’s per-chart
“London always sweeps the Asia session high” is the kind of rule that sounds universal and isn’t. Session behavior is regime- and instrument-specific: it can hold up on one FX pair for months and do nothing consistent on a crypto pair trading 24/7 with no real session structure. That’s why SLA doesn’t hard-code a global number — it calibrates each range against its own chart’s history and prints a frequency with a sample size, or says it’s still collecting one. We walked through exactly this question on one pair, honestly, including the case where the answer was closer to a coin flip than a rule. Both outcomes are useful — a rule that only holds on some charts is information too, as long as you can see the sample it came from instead of taking someone’s word for it.
FAQ
Is a sweep the same thing as a stop hunt?
“Stop hunt” implies intent — that someone pushed price there specifically to trigger those orders. A sweep only describes the outcome: a level got traded through and then failed to hold. The mechanism doesn’t require anyone deliberately targeting the level, only that resting orders were there when price arrived.
Does a sweep guarantee a reversal?
No. That’s the entire point of the armed → swept → accepted/rejected distinction — “swept” is a fact about the level, not an outcome about price. Plenty of sweeps get accepted and keep going.
How do I tell a sweep from a breakout while it’s happening?
On the forming bar, you can’t reliably — see the repainting discussion above. What you can do is wait for the close and then judge the next few confirmed bars: holding beyond the level points to acceptance, closing back inside points to rejection.
Where can I see this measured on my own chart instead of taking it on faith?
Both linked indicators print their state and sample size directly on the chart, and are open-source, so the counting logic is checkable rather than a black box.