Liquidity Sweep and Stop Hunt on Gold: How to Read Them

A liquidity sweep (often called a stop hunt) is when price briefly trades beyond an obvious high or low, triggers the stop orders resting there, and then reverses. On XAUUSD these sweeps are common around equal highs, equal lows and session extremes. A sweep alone is not a trade signal, though. In our testing, sweep-based entries worked on one timeframe and failed on another, so we treat them as context rather than a trigger.

What is liquidity in ICT trading?

In ICT and Smart Money terms, liquidity means clusters of resting orders at predictable prices. Most of these are stop-loss orders and breakout entry orders.

Think about where retail traders put stops:

• Longs place stops just below a recent swing low.
• Shorts place stops just above a recent swing high.
• Breakout traders place buy stops above resistance and sell stops below support.

When price reaches those levels, the stops turn into market orders. A sell stop below a low becomes a sell order, which provides liquidity for someone who wants to buy in size. That is the logic behind the idea that price "seeks" liquidity.

BSL and SSL: buy-side vs sell-side liquidity

ICT splits liquidity into two types:

Equal highs and equal lows get special attention. When price makes two tops at almost the same level, many traders see "double top" resistance and put their stops just above it. That makes the level an obvious pool of liquidity.

What a stop hunt looks like on XAUUSD

On a gold chart, a classic sell-side sweep follows a recognisable pattern:

1. Price forms a clear low, or a pair of equal lows, over several hours.
2. A fast candle breaks below that low. Stops trigger and the move looks like a breakdown.
3. Price fails to continue. The candle closes back above the low, or the next few candles reclaim it.
4. Momentum shifts upward, sometimes leaving a bullish fair value gap.
5. Price then breaks a recent lower-timeframe swing high, a market structure shift.

The buy-side version is the mirror image above a high.

Gold has a few habits that make sweeps frequent:

• Session opens. The London and New York opens often run the Asian range high or low before choosing direction.
• News spikes. Major US data can push price through both sides of a range within minutes.
• Round numbers. Traders cluster stops near big round prices, and gold reacts around them.
• Wide ranges. Gold moves many dollars in an hour, so a "small" sweep can still be large in money terms. At $1 per 0.01 lot per $1 of movement, a few dollars of wick can matter.

Why confirmation matters more than the sweep

Here is the uncomfortable truth: in real time, a sweep and a genuine breakout look identical for the first few candles. Price breaks the low either way. You only know it was a sweep after price reverses.

That is why trading the wick itself is risky. If you buy the moment price pokes below a low, you are really betting on a reversal with no evidence yet. Sometimes the "sweep" is just the start of a trend leg.

Confirmation filters out many of those cases. Common forms include:

• A close back inside the range, not just a wick.
• A market structure shift on a lower timeframe in the new direction. Our guide to the market structure shift covers how to define it.
• A fair value gap created by the reversal move, which then gives a retracement entry. See fair value gaps on gold.
• Higher-timeframe context, such as the sweep happening inside a higher-timeframe imbalance zone.

Confirmation costs you price. You enter later and often with a wider stop. That trade-off is usually worth it, because the alternative is catching breakouts with a tight stop.

What our testing found: sweep entries were inconsistent

We tested liquidity-sweep entries as part of our research on XAUUSD M15 broker data from July 2022 to September 2026, about 98,000 candles, with spread and slippage on every trade. These are hypothetical backtest results.

The finding was simple but important:

• On one timeframe, requiring a sweep before entry improved results.
• On another timeframe, the same rule made results worse.

A rule that helps in one setting and hurts in a near-identical one is a warning sign. It suggests the improvement was specific to that sample, not a robust edge. Adding it to the model would have been a form of curve-fitting.

So we did not make a sweep a required condition in our final model. The model instead relies on stacked higher-timeframe fair value gaps, a rejection from that zone, and an M15 structure shift. Many good setups include a sweep naturally, since a rejection often begins with price probing past a nearby high or low. The difference is that we do not demand one.

We explain how we test ideas like this, including why we compare against random entries, in backtesting ICT concepts. The final model beat 400 random-entry runs; the sweep filter did not earn its place.

Stop placement after a sweep

Sweeps also teach a lesson about your own stops. If obvious stop placement gets hunted, then placing yours one tick below the most obvious low is asking for trouble.

In our research, tight stops usually did worse. Stops at the swing, at the FVG candle or at 1×ATR generally produced roughly half the return with two to three times the drawdown of stops placed at wider structure. One tight-stop version looked excellent only because of a single +68R outlier trade.

Our model's structural stops have a median of about $23 of gold price, with 90% under roughly $73. That is wider than many traders expect, so position size must adjust. Our lot size guide shows how.

A practical sweep checklist

Before acting on a sweep on gold, ask:

1. Which liquidity was taken, BSL or SSL, and was it an obvious level?
2. Did price close back inside, or only wick through?
3. Is there a lower-timeframe structure shift in the new direction?
4. Does the higher timeframe support the reversal?
5. Is there a news release that could explain the spike instead?
6. Where is a structural stop, and what lot size does it allow at your risk?

You can see how a confirmation-first model handles real sweeps, winners and losers, on our live results page.

FAQ

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

In practice, traders use the terms for the same thing: price trades beyond an obvious high or low, triggers resting stops, and reverses. "Stop hunt" describes the effect on traders. "Liquidity sweep" describes it in ICT terms, as price taking resting orders.

What are BSL and SSL?

BSL is buy-side liquidity, the buy stops resting above highs. SSL is sell-side liquidity, the sell stops resting below lows. A sweep of BSL can precede a bearish move, and a sweep of SSL can precede a bullish move, if confirmed.

Should I enter as soon as price sweeps a low?

Usually not. In real time, a sweep looks the same as a breakout. Waiting for a close back inside the range and a lower-timeframe structure shift reduces the chance of buying into a genuine breakdown.

Do liquidity sweeps work on gold?

They occur often on XAUUSD, especially around session opens and news. In our backtest, requiring a sweep before entry helped on one timeframe and hurt on another, so we did not use it as a required filter. Test it on your own rules before relying on it.

This article is educational and not financial advice. Trading gold and leveraged products carries a high risk of loss.

Trading involves substantial risk. This is educational content, not financial advice. See the risk disclosure.