Strip away the vocabulary and this is a failed breakout with a confirmation rule. What smart-money framing adds is a specific, checkable reason the failure happened: price went to where the stop orders were, because that is where the fills are.
Why it works
Large orders need counterparties. A fund wanting to buy size cannot simply lift every offer without moving price badly against itself. It needs a cluster of sell orders — and the most predictable cluster is the stop-losses sitting beneath an obvious low.
So price is driven down through that low, the stops trigger and become market sells, the large buyer absorbs them, and price reverses. The sweep is not manipulation in a conspiratorial sense; it is the mechanical consequence of needing liquidity to transact.
Identifying the liquidity
Only obvious pools count. If you have to hunt for it, so does everyone else, which means nobody's stop is there:
- Equal highs or equal lows — two or more touches at nearly the same price.
- Previous day high and low, and previous week high and low.
- Session extremes, particularly the Asian range in forex.
- An obvious swing point that any trader would place a stop beyond.
- Round numbers, where human orders cluster for no technical reason at all.
Market conditions required
- A clearly identifiable liquidity pool, marked before price approaches it.
- An active session — sweeps in dead hours are usually just drift.
- A higher-timeframe level or bias supporting the reversal direction.
- No major news imminent.
Higher-timeframe bias
Establish direction on the 1-hour or 4-hour before looking for the sweep. The best version of this trade is a sweep of a low inside a higher-timeframe uptrend — the sweep clears out weak longs and provides fuel for continuation. A sweep against the higher-timeframe trend is the version that produces most of the losses.
Entry rules
- Price trades through the marked liquidity level and closes back beyond it.
- On the 5-minute chart, price then breaks its most recent opposing swing point — the change of character.
- Enter on the close of the CHoCH candle, or on a retracement into the imbalance that break created.
- If no CHoCH occurs within roughly an hour, abandon the setup. A sweep without a structure break is just a trend continuing.
Stop-loss rules
Beyond the sweep extreme, plus a small buffer. That wick is the invalidation: if price returns through it, the liquidity was not being collected, it was being followed.
Take-profit rules
| Target | Where | Notes |
|---|---|---|
| T1 | The nearest opposing liquidity — equal highs, prior swing | Take partial here |
| T2 | Previous day high/low on the opposite side | The core target of the trade |
| Runner | Higher-timeframe level | Only with trend alignment |
This structure often produces 2R to 4R because the stop sits just beyond a wick while the target is a structural level some distance away. That asymmetry is the reason the strategy survives a win rate well under 50%.
Risk management
- 1% maximum. The tight stop tempts oversizing — resist it.
- One sweep trade per session per instrument.
- If the CHoCH entry is stopped, do not immediately re-enter on the next structure break; that is the same idea twice, and the market has already disagreed with it once.
Example winning trade
XAU/USD, 5-minute. Equal lows at 2,342 formed over the Asian session; the 4-hour trend is up. At 08:20 price drives to 2,337, five dollars below the equal lows, then closes back at 2,344.
At 08:50 the 5-minute chart breaks its prior swing high at 2,349 — CHoCH. Long 2,349, stop 2,335 beneath the sweep wick, fourteen dollars of risk. T1 at prior-day high 2,371 reached for +1.5R; remainder ran to 2,384 for approximately +2.5R.
Example losing trade
GBP/USD, 5-minute. Equal highs at 1.2760 swept at 09:15 with a close back below. CHoCH down at 1.2748, short entry, stop 1.2766.
Price reaches 1.2740, stalls, then reverses and grinds back through the stop. Loss of −1R.
In review: the 4-hour was in a clear uptrend, and this was a counter-trend short into it. The sequence was textbook and the context was wrong — which is the usual anatomy of a losing SMC trade.
When NOT to trade it
- No CHoCH after the sweep — no trade, no exceptions.
- The liquidity level was not obvious enough for stops to have been sitting there.
- The sweep happens during dead hours with no volume behind it.
- The setup is counter to a strong higher-timeframe trend.
- The stop distance is so large that a sensible position size becomes uneconomic.
Common mistakes
- 1Front-running the sweep
Placing orders in advance because price “should” sweep the low. Sometimes it simply keeps going.
- 2Calling every wick a sweep
The excursion must be meaningful enough to have plausibly triggered resting stops.
- 3Inventing liquidity
If you needed to squint to find the equal highs, they were not obvious, and no stops were there.
- 4Ignoring the higher timeframe
Nearly every losing example reduces to this.
Backtesting considerations
- Define the liquidity pool objectively before the fact — for example, two highs within 0.1× ATR of each other. Otherwise you will only ever find the sweeps that worked.
- Impose a time limit on the CHoCH, or you will retrospectively count structure breaks hours later as confirmation.
- Split results by higher-timeframe alignment. Expect the with-trend subset to carry the edge.