Liquidity Sweeps & Stop Hunts: The Trade Most Retail Gets Wrecked By

A liquidity sweep isn't market noise. It's market design. The wick that just blew through your stop wasn't a coincidence — it was the entire reason that move happened. Learn to read it, and one of the most expensive setups in crypto becomes one of the most profitable.

What a Liquidity Sweep Actually Is

A liquidity sweep — also called a stop hunt, a liquidity grab, or a sweep of liquidity — is a deliberate, fast move that pierces a key swing high or swing low, fills the resting orders sitting beyond it, and then reverses sharply in the opposite direction.

It is not a failed breakout. It is a successful liquidity raid.

To understand it, you have to understand a single fact about how markets fund themselves: large orders need counter-volume to fill. An institution that wants to buy 800 BTC cannot simply click "buy" on a $72,000 ladder — the order book at that price is far too thin. The order would walk the book, blow through twenty levels of resistance, and execute at terrible average prices. So instead of going up to find sellers, smart money sends price down first. They drive it through an obvious support level, where every retail trader has placed their stop-loss as a sell-stop. Those stops, when triggered, become market sell orders. The institution is on the other side, quietly absorbing every single one. Once the resting liquidity is consumed, the artificial selling pressure ends, the institutional bid lifts, and price reverses.

From the chart, this looks like a wick that pokes below support, then snaps back inside the range with a long lower tail. From the order book, it looks like an asset being purchased.

Why They Happen — The Institutional Incentive

Retail traders place stops in obvious places. Below the most recent swing low. Above the round number. Beyond the equal high. Beyond the previous day's high. This isn't a flaw in retail behaviour — it's a feature of how charts are taught. Every standard textbook says "place your stop just beyond the structural level."

Smart money knows where the textbook says to put the stop. They have orderbook data, exchange-level liquidation maps, and aggregated heatmaps showing exactly how many millions of dollars in stops sit at every visible level. From their seat, the chart looks like a treasure map of clustered liquidity — and clustered liquidity is fuel.

Three institutional incentives drive every sweep:

• Order absorption. A buyer who needs to fill 500 BTC cheaply wants retail to panic-sell into them. A sweep below support is the most efficient way to manufacture that panic.
• Liquidation cascades. Leveraged longs cluster their liquidations at the same support levels they cluster their stops. Sweeping support triggers both — stops and liquidation engines — producing forced selling well in excess of natural retail flow.
• Position cost basis. The sweep doesn't just fill the institution — it fills them at a better average price than the pre-sweep range. Every sweep that succeeds reduces the institution's average cost.

This is why sweeps are not random and not optional. In any market with concentrated leverage and visible technical levels — which describes BTC, ETH, SOL and LINK perpetuals exactly — sweeps are the path of least resistance for any large position. They are how size gets filled.

By April 2026, with spot BTC ETF flows running into the hundreds of millions per session and BlackRock's IBIT alone absorbing $214M in a single April day, the institutional appetite for cleanly-priced bitcoin has never been larger. That demand has to be filled somewhere. It gets filled, repeatedly, in the wicks below visible support.

Sweep vs Genuine Breakout: The Four-Tell Test

The single most expensive mistake in crypto trading is treating a liquidity sweep as a real breakout. The chart looks similar in the first 90 seconds — both involve price moving through an obvious level. The difference is everything.

Use this four-tell test on every potential break. Three or four of these confirm a sweep. Three or four reversed confirm a genuine breakout.

The sharpest single tell is CVD divergence. Price can lie, especially during low-liquidity windows. CVD — the running tally of aggressive buy volume minus aggressive sell volume — cannot lie about who initiated the move. When BTC prints a new low and CVD prints a higher low, it means the sellers got smaller while passive buyers absorbed everything they had. That is a sweep, not a break, and the reversal will print within minutes. Read the full CVD divergence guide for the mechanics.

Where Sweeps Cluster on the Chart

Sweeps are not random. They happen at the specific levels where retail liquidity is thickest. The more visible a level, the larger the resting stop pool, and the higher the probability of a sweep. Memorize these — they are the only places worth watching for sweep setups.

1. Equal highs and equal lows

When two or three swing points form at the same price, retail places identical stops just beyond. Every algo on the planet sees the same shape. Equal highs are short-stop magnets above; equal lows are long-stop magnets below. That is why “equal lows (or highs) taken” is one of the seven checks that grade a CAP setup.

2. Previous day, week, and month highs/lows

The PDH, PDL, PWH, PWL, PMH, and PML are reference levels every institutional desk watches. They are the most visible levels on the chart. They are also the most consistent sweep targets — particularly the previous day's high or low during the New York session.

3. Round numbers

$70,000 BTC. $72,000 BTC. $4,000 ETH. $200 SOL. Round numbers attract pending orders the way magnets attract iron filings. The denser the resting orders, the harder the sweep snaps back.

4. The Asian session range

Between roughly 00:00 and 07:00 UTC, BTC and ETH typically establish a tight range as Asian liquidity dominates. Once London opens, that range is the first target. The Asian high and Asian low are arguably the highest-probability sweep levels on the entire 24-hour chart.

5. Wyckoff Spring and Upthrust zones

Inside an accumulation range, the Spring is a textbook liquidity sweep — price drops below support to trap shorts and shake out longs before the markup phase begins. The Upthrust is its mirror image at distribution. The Wyckoff Accumulation guide covers both in full structural context.

6. Liquidation clusters

Aggregated liquidation heatmaps from CoinGlass, Hyblock, or Bitget's own liquidation feed show in real time where leveraged longs and shorts will be force-liquidated. In April 2026, the $64,000–$68,000 BTC range held a short-liquidation cluster roughly four times larger than any cluster above current price — a magnetic field pulling price toward a sweep before any sustainable reversal could occur. These maps are the closest thing retail has to seeing the institutional treasure map.

◈ Free Resource · Nothing to Install

The 8-Point Setup Gate Checklist

Everything below this line gets easier once you trade behind a fixed pre-trade filter. This is the exact 8-point checklist used to grade every BTC, ETH, SOL and LINK setup. Free, by email, about a minute to read.

Send Me the Checklist →

Free. No spam. Unsubscribe any time.

When Sweeps Fire — The Session Map

Sweeps are not equally distributed across the 24-hour clock. They cluster heavily at session opens, when liquidity transitions from one regional desk to another and volatility expands sharply. The two highest-probability windows are universal:

The classic 24-hour rhythm in BTC and ETH on most days follows this pattern: Asian session establishes the range. London Open sweeps one side. NY Open sweeps the other side. Then the real directional move begins, supported by both sweep liquidations.

The free Session Clock tracks every one of these windows live, with countdowns. It tells you when sweeps tend to come, which is useful for planning your screen time.

Anatomy of a Sweep, Candle by Candle

Every well-formed liquidity sweep follows the same three-act structure. Once you can see this on a chart, you can never unsee it.

Act 1 — Consolidation

Price ranges sideways between obvious swing points. Volume gradually contracts. The same horizontal level gets touched two, three, four times. Retail interpretation: "Strong support — looks like accumulation." Institutional interpretation: "Stops are stacking up below; almost ready."

Act 2 — The Sweep

One candle. Sometimes two. Price knifes through support on a fast wick. Stops fire as market sells. CVD does not follow — instead, it diverges, printing a higher low while price prints a new low. Volume spikes for 30 to 90 seconds, then dies. The candle closes back above support, leaving a long lower tail.

Act 3 — Reclaim and Reversal

Within 1–3 candles after the sweep, price reclaims the swept level and breaks the prior swing high or the most recent lower high — a classic Change of Character (CHoCH). Now the structure has flipped: the candle that was supposed to be the breakdown becomes the lowest point of the next leg up. Retail traders who shorted the breakdown are stopped out at progressively worse prices, and their forced buying becomes additional fuel for the reversal.

The CAP Sweep-and-Reclaim Rules

This is how the CAP trades a sweep. It is not a separate setup: the sweep is the heart of Gate 4, and a pullback that has not swept an earlier swing is never graded at all. But a sweep on its own is not a trade either. It has to happen in the right place, in the right direction, and be confirmed.

All five must be satisfied. If any one fails, no trade. This is the Stand Down rule — and it is exactly what separates a documented protocol from discretionary pattern-matching.

1. Rule 1 — With the bigger move. Gate 1: the larger wave count and the Wyckoff phase set the direction. A sweep of lows is only a buy when the bigger move is up; a sweep of highs is only a sell when it is down. A sweep against the bigger move is a counter-trend trap, however clean it looks.
2. Rule 2 — After a strong break, inside the zone. Gates 2 and 3: a BOS of at least 2% on Bitcoin (scaled up for more volatile coins) has set the inverted Fibonacci, and the pullback has reached the OTE zone (0.382 to 0.236 on the inverted Fibonacci, sweet spot 0.295) without breaking the 0. A sweep anywhere else is not a CAP sweep.
3. Rule 3 — An earlier swing is swept, and the candle closes back. Gate 4’s requirement: the pullback runs a visible swing — equal lows, a prior swing low, a Wyckoff Spring level — and the candle closes back above it. A wick beyond with a body that stays beyond is a real break, not a sweep.
4. Rule 4 — Graded S or A. The seven checks: early in the count, the 4-hour discount, equal lows taken, an overlapping corrective pullback, a 4-hour order block or daily level, a tight zone, an untapped fair-value gap. Six or more, early, is S; five is A. Anything lower is skipped. CVD divergence, volume and the liquidation feed help you judge the sweep, but they are not among the seven.
5. Rule 5 — The close beyond the sweep candle. Gate 5: the first candle on the entry timeframe that closes back above the sweep candle’s high (below its low, for shorts) is the Change of Character. That close is the trigger. Without it, you are guessing.

A BTC Sweep, Step by Step: A Worked Example

Theory without application is just noise. Here is how this setup develops in BTC perpetuals, step by step. The numbers are illustrative of the rule set, not a logged trade or a tip.

Scenario: the weekly trend is up and the larger count has a third wave under way, so the bigger move is up (Rule 1). On the 4H chart, BTC breaks out of a range with a close at $72,400 — 5.5% up from the $68,600 low, well past the 2% Bitcoin needs. The inverted Fibonacci runs from 0 at $68,600 to 1 at $72,400, so the 0.382 line sits at $70,052. Over the next day price eases back through it (Rule 2). On the 1H chart, two equal lows have formed at $69,820 and $69,815, inside the zone. Every momentum scalper on retail Twitter is calling it “key support”.

03:08 EST: a 15-minute candle on BTCUSDT.P opens at $69,910, runs to $69,720 — through the equal lows by about $100 — and closes back at $69,880, with a high of $69,960. Long lower wick, body back above the lows. The wick took out every visible long stop in the area (Rule 3).

Three supporting reads print at the same time. They are not rules, but they say the sweep is real:

• CVD divergence: price made a new low at $69,720; CVD made a higher low — the selling was absorbed, not chased.
• Volume: a spike of about three times the recent average on the wick candle, then immediate contraction. Classic stop-hunt signature.
• Liquidation feed: a burst of long liquidations between $69,820 and $69,720 inside a couple of minutes.

Now the grade (Rule 4). The call was made the day before, so it is early; price is in the 4-hour discount; equal lows were taken; the pullback overlaps and looks corrective; and the sweep tags the 4-hour order block at the base of the breakout. The zone is wide and there is no untapped fair-value gap, so those two checks fail. Five of seven: an A. Tradeable.

03:23 EST: the next 15-minute candle closes at $69,940, still below the sweep candle’s $69,960 high. Wait. 03:38 EST: the one after closes at $70,010 — the first close above it. Rule 5 satisfied. CHoCH printed.

Entry executed: long at $70,010, at market. Stop at $69,510: beyond the $69,720 sweep low, the last 4-hour swing, minus Bitcoin’s 0.3% buffer. Risk = $500 per BTC. At a $1,000 risk allocation per trade (1% of a $100k account), that is 2 BTC. There is no target. The whole position is held until the bigger move ends.

What happens next: price works higher through New York and the days that follow, with two sharp pullbacks that would have shaken out anyone who had dragged a stop to break-even. The count completes near $73,500, and the position is closed there: about +7R on the whole position, one decision at entry and one at exit.

Where Sweeps Live in the CAP Framework

Inside the Continuation Acceleration Protocol, the liquidity sweep is the requirement at the heart of Gate 4. A pullback that has not swept an earlier swing is not graded, however clean the rest of the chart looks. The complete protocol runs as:

1. Gate 1 — Regime. Trade only with the bigger move, read from the larger wave count and the Wyckoff phase. The clock is not part of it; the Session Clock is a timing aid.
2. Gate 2 — BOS. A strong candle close beyond the last swing: at least 2% on Bitcoin, scaled up for more volatile coins. Full BOS guide here.
3. Gate 3 — The zone. A pullback to the OTE zone (0.382 to 0.236 on the inverted Fibonacci, sweet spot 0.295), a 61.8% to 76.4% retracement, with the 0 intact. Full OTE guide here.
4. Gate 4 — Confluence. The pullback must sweep an earlier swing, then grade S or A on seven checks. This article.
5. Gate 5 — CHoCH. The first lower-timeframe close back beyond the sweep candle: the entry trigger.

Notice how the gates compound. By the time you reach the sweep, you already know the direction of the bigger move, have a confirmed structural break, and have price in the deep part of the pullback. The sweep is the moment the stops under that pullback are taken; the grade decides whether this one is worth taking; the close decides when.

This is why CAP traders take far fewer trades than discretionary chartists. Most days, one or more gates fail. Stand Down is not a missed opportunity — it is the point of the protocol.

The Four Mistakes That Turn Sweeps Into Losses

Sweeps are a magnet for retail mistakes because the setup looks deceptively simple — wick down, candle reverses, buy. The simplicity is exactly what makes it dangerous. Here are the four expensive errors that separate profitable sweep traders from blown accounts.

1. Front-running the sweep before the wick prints

The single most common error. A trader sees price approaching equal lows, anticipates the sweep, and longs before the wick has actually printed. The problem: until the wick is in, the level might just break cleanly through to the next liquidity pocket below. You are betting on a sweep that may never happen — and even if it does, you are giving up the diagnostic information of seeing the wick close back inside. Wait for the wick. Wait for the close. Then evaluate.

2. Entering on the sweep candle close without waiting for CHoCH

The wick prints, the body closes back inside the range, and a trader buys immediately at the close of the sweep candle. This is better than front-running — but still premature. Plenty of well-formed sweep candles fail to produce a close back beyond them and instead chop sideways for hours before a second sweep takes price lower. The CHoCH (Rule 5) is what filters out these slow failures. Patience here is worth multiple R per year.

3. Putting the stop just past the wick

Some traders, trying to tighten risk, put their stop at the sweep candle’s body or a few ticks past the wick. That invites the next normal volatility pulse to take them out, and after a sweep the retest is the most normal pulse there is. In testing on Bitcoin, a stop just past the wick won only 20–35% of the time. The CAP stop sits beyond the last 4-hour swing, plus a buffer for the coin’s volatility, and the position is sized from that distance.

4. Treating every wick as a sweep

Not every wick beyond a level is a CAP-grade sweep. If the wick happens at a random level no one was watching, against the bigger move, or outside the zone, it is just price noise. The sweep setup requires visible liquidity being raided — equal highs, prior session highs, round numbers, Wyckoff zones. Everything else is non-actionable. Pre-mark your levels at the start of every session and only fire on those levels. Discipline here separates the protocol from the gamble.

What the Backtest Says About Sweeps

Two things about sweeps are asserted constantly and measured rarely: that a sweep is a signal on its own, and that the stop belongs just past the wick. Here is what the CAP’s published backtest can and cannot say about each.

Every tested trade had a sweep — and a sweep was never enough

Under the tested rules the sweep is required, so every one of the 10,899 trades in the backtest followed one. But the sweep was only the entry ticket: each trade also needed the bigger move, a strong BOS, the deep pullback, an S or A grade and the CHoCH close. Together, those S and A grade setups won 78.6% of the time across Bitcoin, Ethereum, Solana and Chainlink, with an average of +3.12R after all costs. Split by grade, S won 80.3% (n=4,041) and A won 77.5% (n=6,858).

The stop just past the wick did not survive

The textbook stop sits a few ticks beyond the sweep wick. Tested on Bitcoin, a stop there on every trade won only 20–35% of the time, because the first retest of a sweep often trades back through the wick before the move starts. The CAP stop sits beyond the last 4-hour swing plus a small buffer, and the size comes from that distance.

Volume confirmation is an assumption, not a result

The standard advice is to demand a volume expansion on the sweep candle. No filter of that kind is part of the tested rules and no table on this site measures it, so this guide does not quote a number for it. Treat it the way you would treat any filter nobody has shown you working: test it on the instrument you trade, count how many trades sit on each side of it, and adopt it only if it earns its place.

This is a backtest of S and A grade setups, after all costs, on modelled versions of Charles’s calls, not account statements. The direction of each trade is his call; the same entries with a mechanical direction showed no edge. Past performance does not guarantee future results. Every trade, the weak years and the losing stretches are published on Protocol Results.

Frequently Asked Questions

What is a liquidity sweep in trading?

A liquidity sweep is a deliberate move beyond a key swing high or swing low designed to trigger clustered stop-loss and pending orders, fill institutional size on the resulting forced flow, and then reverse. The wick spikes through the level, prints volume from forced exits, and rejects sharply — leaving a long tail and a candle body back inside the prior range. It is the structural signature of smart money harvesting retail liquidity.

What is the difference between a liquidity sweep and a stop hunt?

They describe the same event from different angles. "Liquidity sweep" is the structural description — price sweeps a pool of resting orders. "Stop hunt" is the intent description — institutions hunt clustered retail stops. In the CAP Framework, “sweep” is the chart pattern and “hunt” is the motivation.

How do you identify a real liquidity sweep versus a genuine breakout?

A liquidity sweep shows: (1) a wick beyond the level with the candle body closing back inside, (2) CVD divergence — price prints a new extreme but Cumulative Volume Delta does not, signalling institutional absorption, (3) a volume spike on the wick followed by an immediate volume contraction, and (4) reversal within 1–3 candles. A genuine breakout closes its candle body beyond the level, expands volume across the next several candles, and follows through with directional CVD agreement.

Where do liquidity sweeps usually happen in BTC and ETH?

Sweeps cluster at the most obvious technical levels: prior daily and weekly highs/lows, equal highs and equal lows, the previous session high/low (Asian, London, New York), the round number above or below current price, and the highs/lows of the last 48 hours. The more visible the level, the larger the resting stop pool, and the higher the probability of a sweep.

What is the best time to expect a liquidity sweep?

The two highest-probability windows are the London Open (08:00 UTC, 03:00 EST, 15:00 ICT) and the New York Open (13:30 UTC, 08:30 EST, 20:30 ICT). The Asian session typically establishes the range; London or NY frequently sweeps the Asian high or low before the real directional move begins. The free Session Clock tracks these windows live. In the CAP it is a timing aid, not a rule: a sweep that passes the gates counts at any hour.

Can you trade against a liquidity sweep?

Yes — and in the CAP Framework the sweep is required. Gate 4 only grades a pullback that has swept an earlier swing and closed back, and only S and A grades are traded. Entry is taken at market on the first close back beyond the sweep candle, the stop goes beyond the last 4-hour swing plus a small buffer, and the whole position is held until the bigger move ends.

Are liquidity sweeps illegal manipulation?

No. Sweeps are an emergent property of how leveraged markets fill large orders, not a coordinated conspiracy. No single institution controls them; they happen because every large participant has the same incentive to execute size at the levels with the most resting liquidity. This is true in equities, FX, commodities, and crypto. The only difference in crypto perpetuals is that the leverage is higher, the liquidation cascades are more violent, and the sweeps are correspondingly more visible on the chart.

Do liquidity sweeps work the same way in gold, forex and equities?

Yes — with one timing variation. Gold sweeps cluster heavily at the London AM Fix (10:30 UTC, 05:30 EST) and London PM Fix (15:00 UTC, 10:00 EST) in addition to the standard session opens. Equity index futures sweep around the New York Open, the European cash close, and major economic data releases. The mechanic — wick + reverse + CVD divergence — is universal. The level locations and timing windows are market-specific. The CAP itself is tested on Bitcoin, Ethereum, Solana and Chainlink.

◈ The Chart Whisperer · Free Resource

Get the 8-Point Setup Gate Checklist

The exact pre-trade checklist that filters every BTC, ETH, SOL and LINK setup before risk goes live. The sweep is one of its eight points; the other seven decide whether that sweep is a trade or a trap.

Get the Checklist →

Free. No spam. Unsubscribe any time.