The setup in one sentence
Price runs an obvious level (prior-day high, swing low, equal lows, opening-range edge) — taking out stops sitting beyond it — then reverses sharply because the participants who were defending that level are now flat or adding to the reversal.
The ICT model formalizes a pattern institutional algos and order-flow traders have used informally for decades. It calls the run "liquidity sweep," the reversal "displacement," and the entry zone "Fair Value Gap" or "Order Block." The names are jargon. The mechanic is just stop-hunting followed by reversal.
This guide breaks down what each part actually means, what to watch for, and how an automated strategy can trade the pattern on NQ and ES.
What gets "swept"
A liquidity pool, in ICT terms, is a price level where a meaningful amount of stop-loss orders are clustered. The classic pools traders run:
- Prior day high (PDH) / prior day low (PDL): Stops sit just beyond yesterday's extremes. Every overnight session takes a look at one or both.
- Swing highs and swing lows: The most recent local pivot. Day traders use it as a stop reference.
- Equal highs and equal lows: Two or more highs (or lows) that printed at the same level. Stops cluster here because traders see the level as "confirmed."
- Opening-range high / low: First 15 or 30 minutes of the New York cash session. A natural reference for breakout traders, so a natural stop location for breakout-faders.
- Asia and London session high / low: Overnight extremes. Often run during the New York open's first hour.
A "sweep" happens when price moves through one of these levels — typically with a wick, sometimes with a brief close — and then reverses back inside. The wick beyond the level took out the stops. The close back inside means the breakout failed, often because the participants who pushed price through the level have been filled and are now closing or reversing.
The four-step sequence
The HunterICT and HunterKillShot strategies look for the same four-step sequence before any entry fires:
1. Liquidity sweep
Price moves beyond a tracked level (PDH, PDL, swing point, equal level, OR edge, Asia/London extreme) and either:
- Closes back inside the level (close-reclaim mode), or
- Leaves an obvious wick beyond and the next bar rejects (wick-rejection mode)
Both modes are valid; close-reclaim is stricter and produces fewer signals.
2. Opposite-direction Fair Value Gap (FVG)
A Fair Value Gap is a three-candle pattern where the first candle's high and the third candle's low don't overlap — leaving a "gap" of price that wasn't traded efficiently. After a sweep, you often see a sharp directional move that prints an FVG in the opposite direction of the swept level.
If the sweep took out a high (stops above), the FVG should be bearish (gap pointing down). If the sweep took out a low (stops below), the FVG should be bullish (gap pointing up).
The FVG isn't required for entry — but its presence signals that displacement, not just absorption, followed the sweep.
3. Market Structure Shift (MSS)
After the sweep and (often) the FVG, price prints a clean break of a recent counter-trend swing. If price was making lower lows up to the sweep, the MSS is a clear higher high after the sweep — confirming the reversal pivot is in.
MSS is the stricter requirement of the four. A sweep without an MSS is an indecision pattern, not a reversal pattern.
4. Volatility filter
Last gate. ATR or session-volatility minimum. Filters out chop days where the sequence prints visually but doesn't have the expansion needed to follow through. HunterICT and HunterKillShot both use an ATR floor; bars below the floor are ignored.
When all four conditions align, the strategy enters in the direction of the structural shift.
Why the pattern works
There are two common explanations. They're not mutually exclusive.
The "stop-hunt" framing. Big participants need liquidity to fill large orders. Price levels with stops beyond them are predictable liquidity pools. Running the level fills the institutional order in the opposite direction of the broken stops. After the fill, the participant either does nothing more (price drifts back) or actively reverses (price displaces the other way).
The "false breakout" framing. Range-bound or balance-zone trading produces edges where breakouts fail more often than they continue. A breakout that reverses within one or two bars of taking out a level was probably driven by stop runs rather than genuine directional commitment. The reversal is the market re-asserting the prevailing range.
Both framings predict the same observable: sweeps that close back inside, followed by structural reversals, are tradable.
Stops and targets
ICT-style strategies typically use structural stops rather than fixed-tick stops. The logic:
- For a long entry after a low sweep: stop sits below the swept low, with an ATR buffer to avoid getting wicked out on noise. The buffer adapts to current volatility, so the same parameter scales across instruments and regimes.
- For a short entry after a high sweep: mirror.
Targets are typically projected from the swing range using fib extensions:
- -0.236 (conservative TP1)
- -0.500 (median TP)
- -0.786 (extended TP2 / runner)
HunterKillShot uses this fib-projection model with optional three-leg partial take-profit. HunterICT supports both fixed risk-to-reward and fixed-tick targets in addition to the structural model.
What the pattern doesn't catch
The ICT sweep model is a reversal pattern. It loses on three regime types:
- Strong trend days. Price runs the level, doesn't reverse, just continues. The sweep "fails" — price keeps going. The structural stop catches this losing trade quickly.
- Choppy range days without expansion. All four conditions print visually, but neither the upside nor the downside follows through. The volatility filter catches most of these; the rest take small losses.
- News-driven moves. A CPI release sweeps the prior-day high, prints an MSS, displaces 200 ticks higher, then prints another MSS lower. The pattern fires on noise. News-window blackouts in HunterICT and HunterKillShot disable the strategy during these windows.
Knowing what the pattern doesn't catch is the difference between sustainable algorithmic trading and curve-fit overconfidence.
Where to read more
- HunterICT strategy detail — full feature list and configuration
- HunterKillShot strategy detail — the time-and-structural variant with fib-projection targets
- Methodology page — how Hunters Algo strategies are built and validated
- HunterBreakOut backtest results — the only currently-published public backtest; reproducible in NinjaTrader Strategy Analyzer
How to test it yourself
If you want to see whether the pattern fits your trading, the cheapest path is the 3-day free trial. Install HunterICT or HunterKillShot, point Strategy Analyzer at a 90-day window of NQ or ES, run the default settings, and compare your output against the strategy detail pages. Within tick-data noise, your numbers should match.
Start the 3-day trial via Whop.
Past performance is not indicative of future results. Futures trading involves substantial risk of loss. See disclosures for the full risk and hypothetical-performance language.