Liquidity Sweep vs Break of Structure: Is a Wick a Sweep? 1,846 Breaks Tested

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-08 · Updated 2026-11-02 · Reviewed 2026-11-01

A liquidity sweep is a run past a level that reverses; a break of structure is a run past a swing level that keeps going. The popular shortcut, a wick through the level for a sweep and a close beyond it for a break, leans the right way but decides little: across 1,846 tested breaches, 48% of wicks and 60% of closes were still past the level 10 bars later.

Most of what ranks for this question, from forum answers to one-minute videos, teaches that shortcut, and none of it tests it. This page does, on five markets and the whole daily history of each, and then shows a recent case of each label going wrong.

What each one is

A break of structure is price closing beyond the last swing high in an uptrend, or the last swing low in a downtrend, and continuing in that direction. It is the market structure reading that the trend is still on. Break of structure covers how swings are marked.

A liquidity sweep is price running past an obvious high or low, where stop orders are assumed to sit, and then turning back through it. The reading is that the move was fuel for a reversal rather than the start of a trend. Liquidity sweep covers the stop-order argument behind it.

The two share their first bar. Both begin with price trading past the same kind of level, which is why the question of telling them apart comes up at all.

Where they differ

In what has to happen next. A break needs nothing more than for price to stay beyond the level. A sweep needs a second event, the return through it, so it cannot be confirmed on the bar that crosses.

In the trade each one implies. A break is traded with the move, a sweep against it: opposite positions at the same price on the same bar, which is why a wrong label is expensive.

In the shortcut used to tell them apart. Traders lean on the candle’s close: a wick that closes back inside reads as a sweep, a body that closes beyond reads as a break. That is a claim about the future made from one bar, and it is the claim the data section below measures.

In which levels they care about. Structure is read from the swing sequence; a sweep is read from whatever level is most obvious, such as equal highs or the previous day’s low. Often the same line, not always.

A worked example

A textbook break that failed: QQQ, October 2026. On 22 September QQQ made a swing high at $748.35, confirmed five sessions later. On 2 October it traded to $754.54 and closed at $749.58, $1.23 above the level: by the shortcut, a break of structure. Two sessions later it closed at $759.66. On 8 October, the fourth session after the break, it closed at $747.58, back below the level, and by 16 October, ten sessions on, it closed at $709.80.

That is 5.15% under the level and 5.31% under the breaking close. In hindsight it reads as a sweep; on the day it had every feature of a break.

A textbook sweep that failed: IWM, September 2026. IWM’s swing low of 16 September was $281.03. On 24 September it traded down to $279.05 and closed at $281.66, back above: a wick-only breach, the shape usually called a sweep. Two sessions later, on 28 September, it closed at $280.02, below the level, and ten sessions after the wick it closed at $277.57, 1.23% under it. The “sweep” was a break that took two days to show.

Neither case was rare. They are two of the 1,846 breaches counted below, and the counts say a close against the first-bar reading came within 10 bars in most cases of each kind.

Which one to use: Liquidity Sweep or Break of Structure?

Default to the break when price closes beyond the level, because that look held more often than it failed: 60.46% of close-beyond breaches were still beyond 10 bars later. Expect a test of the level on the way, since two in three closed back inside at least once.

Read a sweep only when the return has actually happened. A wick alone was a coin flip: 48.04% finished beyond the level anyway. The sweep reading earns its place when price closes back through the level and stays there, not when a candle merely leaves a tail.

When neither has happened yet, size for being wrong. In 70.64% of all breaches a close within the next 10 bars contradicted the first bar’s look, so the position has to survive the other outcome.

How to make either label testable

Name the bar count. “Back inside within three sessions” is a rule; “eventually returned” makes every move a sweep given enough time.

Name the close that decides it. A wick and a close are different events, and the data shows they carry different, but small, amounts of information.

Mark the level before the move. A line drawn after the excursion is history, and both concepts look perfect on lines drawn that way. The swing high and low page sets out the five-bar rule used here.

Pick your default before the chart opens. Choosing after you see what price did is how both labels end up looking reliable in a trading journal.

The original data

The test. Daily bars from Yahoo Finance for four funds and one coin: QQQ (Nasdaq-100) from 10 March 1999, SPY (S&P 500) from 29 January 1993, IWM (Russell 2000) from 26 May 2000 and GLD (gold) from 18 November 2004, all to 30 October 2026, and bitcoin in dollars from 17 September 2014 to 31 October 2026.

A swing high is a bar whose high is above the highs of the five bars before and the five after, so it is known five bars later; a swing low is the mirror. The latest known swing high and low are the live levels, and each is counted once, on the first later bar that trades past it.

The breach “closed beyond” if that bar’s close was past the level, otherwise it was “wick only”. Each case is followed for 10 bars (sessions for the funds, days for bitcoin), and cases without 10 bars after them are left out. A later close exactly at the level counts as neither beyond nor back inside.

What the table says. The close carried some information: in every market a close beyond was followed by price staying past the level more often than a wick was, by 9 points in SPY up to 21 points in bitcoin. Pooled, the 12-point gap is far larger than chance would produce at these sample sizes; on its own, the GLD gap is small enough that chance could explain it and SPY’s is borderline, and the wick-only 48.04% is within chance of a coin flip. The close did not carry much.

Of the close-beyond breaches, 721 of 1,080 (66.76%) closed back inside at least once within 10 bars, and of the wick-only ones, 583 of 766 (76.11%) closed beyond the level within 10 bars. In 1,304 of the 1,846 cases (70.64%), a close inside the next 10 bars contradicted the first bar’s look.

Direction mattered too: on QQQ, 63.03% of breaches above swing highs were above the level 10 sessions later, against 47.31% of breaches below swing lows. The rows by market and side are in the summary file, and every breach is listed in the QQQ breach file and its four siblings.

The October case, drawn. The chart below plots the real QQQ daily bars from 8 September to 16 October 2026 with the 22 September swing high, so the break, the close back below and the slide to $709.80 can be seen in order. The bars are in the QQQ price file.

Checked against second sources. On Nasdaq’s own prices for 1 November 2016 to 30 October 2026, the same rule found 150 QQQ breaches against Yahoo’s 149, with all 149 matching on date, side, first-bar look and 10-bar outcome; SPY, IWM and GLD matched case for case (145, 136 and 139).

For bitcoin, Coinbase’s daily candles gave 217 breaches against Yahoo’s 215 and 172 matched exactly, because the two feeds print different highs and lows; the gap between the two looks held on both (60.63% against 39.77% on Yahoo, 56.20% against 41.67% on Coinbase).

Euro-dollar was tried and left out: Yahoo’s currency bars had 128 days with a high or low outside the open-close range and there was no second feed of currency highs and lows to check against.

The limits. Daily bars, one swing width, one 10-bar window, and fund prices not adjusted for dividends; a different width or window would move every count, and intraday charts were not tested.

What the video record looks like. In the 24,971-video search study behind this site, liquidity sweeps appear in 69 titles across 49 channels at a median of 17,537 views, and break of structure in 110 across 95 at a median of 7,667, so the reversal reading draws more than twice the audience per video.

When it fails

The costly failure is shorting every wick above a high. It feels like buying a trap at its best price, but a wick-only breach finished beyond the level about as often as not, and above swing highs on QQQ, 76 of 96 closed above the level within 10 sessions.

The second is trusting a close beyond without a plan for the retest. Two in three close-beyond breaches came back inside at least once, so a stop placed just under the level was exposed even when the break held.

The third is labeling after the fact. A journal that tags trades “sweep” or “break” once the outcome is known will show a perfect record and teach nothing.

And the fourth is judging on a timeframe you do not trade. Daily structure and five-minute structure disagree constantly; this test is daily only.

Questions people ask

What is the difference between a liquidity sweep and a breakout?

A breakout, like a break of structure, is price leaving a level and continuing; a sweep is price leaving it and coming straight back. The difference only exists after the move, which is why a failed breakout and a sweep are often the same event under two names.

How do you tell if it’s a liquidity sweep?

By the return, not the wick. Price has to close back through the level and stay there within a limit you set in advance; a wick on its own was followed by a close beyond the level within 10 bars in 76.11% of tested cases.

Is a break of structure a good strategy?

It is a reading, not a strategy. Here, a close beyond a swing level was still beyond it 10 bars later 60.46% of the time, before costs and with no entry, stop or exit attached.

How long should you wait before calling a sweep?

Long enough to name the number in advance. Reversals of the first-bar look came early: of the 1,304 here, the median came 2 bars after the breach and 902 (69.17%) within 3 bars. Pick your own limit and keep it the same on every chart.

Related

Liquidity sweep covers the reversal reading and break of structure the continuation one. To act on either, how to trade a liquidity sweep sets out entries after the return, and false breakout is the plainer name for a break that does not hold. Liquidity sweep vs support and resistance runs the sweep question on plain horizontal levels instead of swing points.

This page is educational, not financial advice. Test every idea on your own charts before risking money.