Liquidity: buy-side, sell-side, equal highs/lows and sweeps

If one concept made the SMC/ICT approach popular, it's liquidity. The idea, simplified, is that price doesn't move randomly but tends to reach the points where many orders concentrate, because to execute large positions you need counterparties. It's a fascinating and partly plausible hypothesis. Here we study it rigorously: what it says, how it translates into verifiable rules and, just as important, how to avoid telling yourself stories.

What liquidity is (and why big operators care)

To buy or sell a large amount of currency you need someone on the other side. If a fund wants to buy 500 million, it can't do it all in an instant without moving the price against itself: it has to find counterparties willing to sell. Where are many sell orders "waiting"? In predictable places:

• Buyers' stop losses, which are sell orders placed below recent lows.
• Short sellers' stop losses, which are buy orders placed above recent highs.
• Breakout orders from traders who buy above highs and sell below lows.

These orders are "fuel": when price reaches them, they trigger at market and generate volume. Hence the hypothesis: price is attracted to areas with many waiting orders. In SMC language these areas are called liquidity pools.

Buy-side and sell-side

• Buy-side liquidity (BSL): above the highs. It contains sellers' stops and breakout buyers' entry orders.
• Sell-side liquidity (SSL): below the lows. It contains buyers' stops and breakdown sellers' entry orders.

Mind the name: "buy-side" indicates that above the highs there's liquidity that buys when price reaches it (stopped-out sellers buy to exit), not that you should buy.

Where it accumulates: objective clues

Not all highs and lows are equal. The most "loaded" areas are the most obvious, because more traders place orders there:

Equal highs/lows in practice: define a numerical tolerance, for example "two swing highs at a distance below 0.1 × ATR(14)". Without a written tolerance the concept becomes subjective: any pair of highs "looks" equal if you want it to.

The sweep: operational definition

A sweep (or stop hunt, liquidity grab) happens when price exceeds a liquidity level with the wick and then returns, closing on the opposite side. Unlike a real break, the candle (or the first following candles) closes back inside the area.

An example written rule:

Sweep of a high L (apparent rally):
  1) high of the candle > L (exceeds the level)
  2) close of the same candle (or the next) < L (returns)
  3) the wick beyond the level >= 0.2 x ATR(14)  (excludes trivial touches)

A sweep on its own is not an entry signal. It becomes interesting when followed by a shift of structure in the opposite direction (MSS with displacement): the sequence sweep → MSS → return into an area is the basis of the trading model you'll see in module 3.

Sweep or real break? How to tell them apart

Price exceeds a level: sweep (returns) or breakout (continues)? You can't know in advance. But you can reduce uncertainty with objective clues:

• The close: a clean close beyond the level with a full body is more likely a break; a close that returns is a sweep.
• Displacement: a break with strong candles and continuation = breakout.
• Trend context: a sweep against the major trend is more likely to be a simple pullback.
• Time: quick returns (within 1-3 candles) are more typical of sweeps.

Liquidity as a target ("draw on liquidity")

A very practical use: use liquidity areas as targets. If you buy after a sweep of the low, the natural target is the opposite liquidity (the previous high, the PDH, equal highs). The reason isn't "magic": they're points where activity tends to concentrate and where it makes sense to take profit. Verify, as always, how often price reaches them.

How to measure whether the concept works

The part almost no course shows you. Here's a simple experiment, repeatable in a couple of hours on an M15 or H1 chart:

1. On 60 days of data, mark for each day the Asian high and low and the PDH/PDL.
2. For each day note: did price sweep the level during London/New York? (yes/no)
3. In case of a sweep: after the sweep, did price reach the opposite level before returning to exceed the sweep? (yes/no)
4. Calculate the percentages and compare them with a control case: take random levels (for example any round number) and repeat the same test.

If price "reacts" to your liquidity levels the same way it does to random levels, the concept adds nothing. If it reacts differently and consistently, you've found something useful. In the lesson on evidence we'll return to why this test is essential.

What liquidity is not

• Every high or low on the chart. Only the obvious and significant ones.
• A guarantee of reversal. After a liquidity grab price can continue.
• A forecast. It's an area where you expect activity, not a certain direction.

Common mistakes

• Marking liquidity "by eye" without a written tolerance.
• Considering every touch beyond the level a sweep, even the ones that continue.
• Entering on the sweep without waiting for structure confirmation.
• Forgetting context: a sweep in the direction of the major trend means something different from one against it.
• Not recording failures: you remember the "perfect" sweeps and forget those that continue.

In short

• Liquidity is the set of waiting orders above highs (BSL) and below lows (SSL): a plausible hypothesis, not a law.
• The most loaded areas are equal highs/lows, PDH/PDL, session highs/lows, trendlines and round numbers; written tolerances are needed.
• A sweep exceeds the level with the wick and returns; on its own it's not a signal.
• The interesting sequence is sweep → MSS with displacement → entry in an area.
• Use liquidity as a target and check with a control test whether the concept works.

Practical exercise

1. On 60 days of EUR/USD M15 mark the Asian high/low and PDH/PDL and count sweeps during London/New York.
2. For each sweep check whether price then reached the opposite level.
3. Repeat with a random level (round number) and compare the percentages.