Liquidity Grabs and Stop Runs in Futures: What They Are and How to Define a Trade

You place a stop just under yesterday's low, price pokes through it by a few points, your stop fills, and the market turns around and goes the other way. It feels personal. A liquidity grab, also called a stop run or sweep, is the name for that pattern: a spike through a visible level that reverses within a few bars. The useful question isn't whether someone targeted you. It's whether the pattern can be written as rules you can test.

Why stops cluster where they do

Traders place stops at the same visible places: just above a prior high, just below a prior low, near round numbers, beyond recent swing points. Stop orders become market orders when triggered, so a cluster of them means a burst of buying above a high or selling below a low. A participant that needs size can find the other side of its trade there. That's the standard explanation, and it's plausible. I can't tell you whether any particular sweep was deliberate, and you don't need to know. Triggered stops also aren't guaranteed to fill where the trader hoped, since stop orders can slip. The mechanism holds without anyone planning it.

Common ES and NQ locations are the prior session high and low, the overnight high and low, round index levels such as 5,000, and multi-day swing points. These are index points, not dollars: an ES point is worth $50 and an NQ point $20.

Writing the setup as rules

1. A visible level named before the trade, such as the prior session high.
2. A sweep: price trades beyond the level by a set number of points on a wick, and the bar closes back on the original side. A bar that closes beyond the level and holds is a breakout, not a grab.
3. A fast return: the reversal begins within a set number of bars, for example 3 on the 5 minute chart. Choose the number once and test it.

If price sweeps a high and fails, the trade is short. If it sweeps a low and fails, it's long. Scale the sweep size to volatility. A 3 to 10 point window that suits ES on a calm day is too tight for NQ on a news day.

A hypothetical short on ES

Suppose the prior session high is 5,040. A 5 minute bar trades up to 5,046 and closes at 5,038, back below the level. You enter short on that close at 5,037 and put the stop 3 points above the sweep extreme at 5,049.

With about $29 of friction on ES (a $4 round-turn commission plus a tick each way), the reward nets to $1,071 and a stopped trade costs closer to $629. That's about 1.7 to 1 after costs. The wide stop is the reason position size matters here: 12 points is $600 on one ES contract, so check it against your account's loss limit before you size up. Margin is collateral, not a cap on what you can lose.

Targets and management

Use a reference you chose in advance: the opposite side of the day's range, a prior swing, or a KPL level. A partial exit at part of the way to the target plus a trailing runner is a common choice. It's a trade-off that gives up some of the best days to smooth the rest, and you should test it against a single fixed target rather than assume it helps.

How it fails

• The sweep keeps going. Price spikes through the level and doesn't come back. That's a breakout, and the stop handles it. Your stop can slip in a fast move, so the realized loss can exceed the plan.
• The confirmation is late. If the close back inside arrives after most of the reversal has happened, the entry is poor and the stop is far.
• Scheduled news. A report release can sweep several levels in seconds and make the pattern meaningless for that window.
• Level overload. If every swing counts as a level, you will find sweeps everywhere. Limit it to a short list written before the session.

KPL levels as a reference

KPL levels are fixed reference prices anchored to the regular session opening print, with four levels on the standard ES and NQ map. They don't move during the day, and they aren't predictions. A sweep near a KPL level is a reasonable thing to tag in your log, but I wouldn't claim it makes the setup better until your own data says so. The levels also don't exist before the open, so overnight sweeps can't use them. See the KPL strategy page for how they're built.

Testing it

Log every sweep you identify, traded or skipped: level, sweep size, bars to reversal, entry, stop, result in points and net dollars. Thirty trades is not a statistically proven edge, 50 SIM trades is not live proof, and a simulated fill doesn't validate queue position or slippage. If you want to run expectancy math on your own logged trades, the free edge calculator is there for that.