ICT Kill Zones & Sessions
"Kill zones" are named time windows inside the trading day that a whole family of intraday methods treats as the only periods worth trading. The vocabulary comes from Inner Circle Trader (ICT) material, but the observation underneath it is older and uncontroversial: activity in futures and FX is not spread evenly across the clock. Volume, spreads and realized range concentrate around the cash-session opens and the London–New York overlap, and thin out during the Asian hours. A kill-zone strategy hard-codes that concentration into its rules.
The point that matters most: a kill zone is a filter, not a signal. The window says *when* to look; a separate price trigger says *what* to take. Every strategy in this family has those two halves, and most disagreements between versions are about the second one.
## How the mechanism works
The windows themselves are conventions, usually quoted in New York time: an Asian range of roughly 20:00–00:00, a London kill zone around 02:00–05:00, a New York open kill zone around 07:00–10:00, and a London close window around 10:00–12:00, with narrower one-hour sub-windows also in circulation.
The liquidity logic layered on top is what gives those windows their supposed meaning. The Asian session typically produces a compressed range on low participation; its high and low become obvious, widely watched levels, and resting stop orders accumulate just beyond them. When London opens and real order flow arrives, price frequently trades through one side of that range first, then reverses and expands the other way — the move usually labelled a Judas swing or a liquidity sweep. The same template is applied at the New York open, substituting the London range, the previous day's high and low, or the session's own opening range.
A typical trade sequence therefore reads: identify the liquidity pool relevant to the window, wait for it to be taken, require evidence that the move has failed (usually a shift in market structure on a lower timeframe), and enter on the retracement into the imbalance, order block or Fibonacci zone left behind. The stop sits beyond the sweep extreme; the target is opposing liquidity or a fixed multiple of risk.
## Main variants
Most published versions recombine a few building blocks. **Asian-range models** mark the overnight range and trade its break or its sweep-and-reverse at the London open. **Session-reversal models** assume the first directional move after the open is the trap and position against it. **Opening-candle models** ignore prior sessions and use the first candle of the session — 5, 15 or 30 minutes — as the reference range. **Fixed-hour models** shrink the window and take one setup at most. **Bias-gated models** add a higher-timeframe direction filter. And **instrument-specific builds** matter: index futures at the New York open behave differently from major FX pairs at the London open, so a rule set tuned on one rarely transfers unchanged.
## What separates one implementation from another
The concept is shared; the parameters are not. The levers that actually change results are the exact clock definition and the timezone it is anchored to; which liquidity reference is used; the entry timeframe, since a 1-minute and a 15-minute entry on the same premise are different strategies; how much confirmation is required, from a bare breakout to a full sweep plus structure shift plus imbalance retest; whether a higher-timeframe bias gates entries; stop and target logic; and the cap on attempts per session. News handling is a further, often unstated, lever.
## Common mistakes
Timezone and daylight-saving drift is the most damaging: a window defined in New York time and applied on a feed running a different server time silently shifts by an hour twice a year. Second is treating the window as the signal — trading because the clock says so rather than because the trigger fired. Third is definition creep: "order block", "fair value gap" and "valid sweep" are elastic enough that, in hindsight, every chart looks like it contained a perfect setup. Fourth is ignoring execution costs at the moment they are worst, since spreads widen and slippage increases at exactly the opens these strategies trade. Fifth is sample-size illusion: one setup per session per instrument caps you near 250 opportunities a year before any filter.
## How to evaluate and backtest a version
Encode the clock explicitly, with DST rules, and verify that the timestamps in your data mean what you think they mean. Then define the reference range and entry trigger mechanically enough that two people coding it independently would produce the same trade list; otherwise the results measure your discretion, not the strategy.
Build the crudest possible baseline first (range break, fixed stop, fixed target, no discretion) and check whether each added filter improves it on data the filter was not chosen on. Run robustness checks specific to this concept: shift the window by ±15 and ±30 minutes, and test the adjacent session. If results collapse under a small time shift, the window was fitted rather than discovered. Test across several years and more than one instrument, use realistic spread and slippage for the open rather than average-of-day figures, and compare against a time-only benchmark — entries at a fixed moment inside the same window — to see whether the edge comes from the timing or the entry model.
The decoded versions linked from this page sit at different points on those axes: different sessions, different reference ranges, and very different amounts of confirmation demanded before entry. Reading two or three side by side is the quickest way to separate the concept from one author's preferences.
Strategies in this concept (21)
- 10 AM UTC-3 Candle Strategy — Titanes del Trading
- Breakout and Retest, Fibonacci, Asian Range — Islam Sultanov
- EURUSD, Manipulaciones de Sesiones — Matias Maderna
- First 30-Minute Candle New York Session Strategy — Casper SMC
- FX Sessions — Ali Casey | StatOasis
- Joovier Gems London Breakout Strategy — Eddy Pips Trading
- London Session Reversal Strategy — Rob Donoh
- London Session Strategy — Gorka Fx
- Nasdaq 100 New York Open Strategy — Ant Finances
- New York Session Review — fxbootcamp
- New York Session Strategy — Alexflamas
- Sesiones Operativas — Brandon Arcila
- SMC London Session Strategy — Gorka Fx
- WS Time-Based Entry Model — Will Street
- Asian Range Manipulation London Session Strategy — Institucional Trading Lab
- ICT Concepts - Orderblocks, Fair Value Gap, PD Array Matrix, Smart Money Tool, Kill Zones — TradingICT
- ICT Kill Zone, Fair Value Gap, Liquidity Inflection, Change of Character Strategy — Craig Percoco
- Manual de Bias London Strategy — Gorka Fx
- Power of Three (AMD), Silver Bullet Strategy, Venom Model — Neeraj joshi
- Session High/Low, Asian Session High/Low, Order Block Strategy — Gorka Fx
- Smart Money Concepts, Asia Session, London Kill Zone, Change in State of Delivery Strategy — The Simplified Trader
Frequently asked questions
Are kill zones and trading sessions the same thing?
No. A session is the whole period a financial centre is active; a kill zone is a much narrower window inside it, usually the first two or three hours after an open, chosen because participation and range concentrate there. Session-based strategies may trade anywhere inside the session, while kill-zone strategies deliberately restrict themselves to the sub-window and stand aside outside it.
Which timezone should kill zones be defined in?
Anchor them to the market's own clock — New York time for US index futures, London time for the European open — and handle daylight saving explicitly, since Europe and the US change clocks on different dates. Charting platforms and brokers often display a third timezone, so convert once in the code or template and verify against a known event, such as a cash open or a scheduled release. This is the most common silent error in the whole family.
Is the edge in the time window or in the entry rules?
That is what a backtest should establish, and the two halves are separable. Run the same entry model outside the window, and a neutral entry inside it, to see which half carries the result. Many published versions never make this distinction, which is part of why strategies built on the same window can behave very differently.
Do you need ICT concepts such as order blocks and fair value gaps to trade a kill zone?
No. The window is purely time-based and can sit on top of any trigger — an opening-range breakout, a moving-average cross, a Fibonacci retracement. ICT vocabulary is the most common wrapper because the terminology originates there, but it is not a requirement, and simpler triggers are usually far easier to define objectively and to test.
Do kill zones apply to every instrument?
The logic depends on participation concentrating at known clock times, so it maps most naturally to instruments tied to a cash session: index futures, major FX pairs, gold. Markets that trade continuously with globally distributed participation have weaker and less stable session structure, although FX and equity session effects do carry over to some degree. A window that works on one instrument should be re-tested rather than assumed when moved to another.
How much data is needed to judge a version of this concept?
More than is usually applied. These strategies typically produce at most one opportunity per session per instrument, so a single year on a single symbol is a small sample once filters are added. Testing over several years so the sample spans different volatility regimes, and on more than one instrument, gives a more honest picture than a longer test on one symbol alone.