ICT trading is a discretionary price-action framework that reads the chart as a sequence of moves between pools of resting orders, using a specific vocabulary โ liquidity, order blocks, fair value gaps, killzones, market structure shifts โ to describe that sequence. The name comes from Inner Circle Trader, the teaching brand of Michael J. Huddleston, who popularised the material through a long series of free videos. Traders who use it are not applying a mechanical system; they are applying a set of concepts, by judgement, to whatever the chart is doing.
In practice a trader marks the highs and lows where stop orders are likely to sit, waits for price to run through one of those levels, and looks for a fast reversal back through recent structure as the signal that the run was the move rather than the start of a trend. Entries are then taken at zones left behind by that reversal, with the opposite pool of liquidity as the target.
This article describes the framework accurately: what each term means, what market logic sits underneath the terms that have one, and where the framework goes beyond established market microstructure. Some of the ideas map onto things every order-driven market genuinely does. Others are interpretation layered on top, and none of it has been independently verified โ no public dataset or audited record establishes how these setups perform.
Where the framework comes from, and what that means for you
ICT is not a standard term in finance. You will not find "order block" or "killzone" in an exchange rulebook, a microstructure textbook or a broker's documentation. The vocabulary was built by one educator to describe patterns he saw, and it spread through video and social media rather than through institutional desks. That is a statement about provenance, not a verdict about usefulness.
It matters for one practical reason. A standard concept โ the bid-ask spread, a limit order book โ can be verified. With proprietary vocabulary you are relying on the explanation being a good description of something real. Some ICT ideas describe real mechanics in unusual words; others can be fitted to almost any chart after the fact. Knowing which is which is the whole skill.
The one idea underneath everything: liquidity
The load-bearing claim in ICT is that price moves toward places where orders are resting, because large participants need those orders to fill size without moving the market against themselves.
The mechanical part is uncontroversial. Stop-loss orders cluster just beyond obvious highs and lows, because that is where traders put them. A stop on a long is a sell order; a stop on a short is a buy order. When price reaches a cluster those stops trigger, and the flow can extend the move briefly. Resting size is where a large order can be worked with less slippage.
ICT calls those clusters liquidity pools, labels them buy-side liquidity above highs and sell-side liquidity below lows, and calls a push through them a liquidity sweep. The descriptive part is sound. The interpretive part โ that a coordinated group deliberately drives price to those levels โ goes beyond what anyone can observe from a retail chart. You can see that price ran a high and reversed; you cannot see who did it or why, and the framework's language often implies more certainty about intent than the evidence supports.
The vocabulary, term by term
| Term | What the framework means by it | Underlying market logic |
|---|---|---|
| Liquidity pool | Cluster of resting stops beyond an obvious swing high or low | Real: stops genuinely cluster at obvious levels |
| Liquidity sweep / stop hunt | A push through that level followed by rejection | Observable: the wick exists; the intent behind it is inferred |
| Market structure | The sequence of higher highs/lows or lower highs/lows | Standard trend definition, renamed |
| Market structure shift (MSS) | First break of the opposing swing after a sweep | A trend-change heuristic; the threshold is chosen by the trader |
| Order block | The last opposing candle before a strong displacement move | Proxy for a zone of unfilled interest; no way to verify orders sit there |
| Fair value gap (FVG) | A three-candle gap where wicks do not overlap | Real imbalance: one side traded through without two-way auction |
| Killzone | A defined window of the session with higher activity | Real: volume and volatility genuinely concentrate around opens |
Two entries deserve separating from the rest. Fair value gaps describe a genuine imbalance: when three consecutive candles leave a gap between the first candle's wick and the third's, price moved fast enough that no two-way trade happened in that band. That is an observable feature of price data whatever you call it, and the fair value gap article covers how to mark one and when it stops mattering.
Killzones also rest on something real: volatility in FX and index futures concentrates around the London and New York opens and around scheduled data. An economic calendar does more work for a session trader than a fixed time-of-day rule, because it tells you why a given window is active today rather than assuming every day is shaped the same.
How a setup is actually constructed
Most ICT setups follow the same skeleton, whatever they are named:
- Mark the ranges โ the previous session's high and low and the obvious swing points on your working timeframe.
- Wait for a sweep. Price trades through one of those levels and closes back inside. Nothing is done yet.
- Require displacement. The move back must be fast and wide-bodied and must break the most recent opposing swing โ the market structure shift. Without it, the sweep is just a continuation.
- Find the entry zone โ the fair value gap or order block left inside the displacement leg, usually on a lower timeframe.
- Enter on the retracement into that zone, with the stop beyond the sweep's extreme.
- Target the opposite pool โ the untouched high or low on the other side of the range.
A worked example: EUR/USD makes a session low at 1.0820, then prints 1.0812 and closes the hour back at 1.0834 โ the low is swept. The next hour breaks the swing high at 1.0846, the structure shift. Inside that up-leg sits a fair value gap between 1.0838 and 1.0844. You place a limit at 1.0842, a stop at 1.0808 (34 pips, below the sweep low), and target the untouched high at 1.0910, 68 pips away โ a 1:2 reward-to-risk trade. Sizing it is arithmetic, not judgement: the EUR/USD lot size calculator turns a 34-pip stop and a chosen risk percentage into a position size.
Note what the structure does: the stop sits beyond the level that would prove the read wrong, not at an arbitrary distance. That is the genuinely portable idea here, independent of the vocabulary.
What invalidates the setup, and how often it fails
The setup is invalidated when price trades back beyond the swept extreme. If the low at 1.0812 trades again, the reading was wrong and the stop should already have closed the position. No interpretation is required at that point.
Three softer failure modes cause more damage than the clean one:
- The sweep was a breakout. Price runs the low and keeps going. There is no reliable way to tell a sweep from a genuine break at the moment it happens. Requiring displacement and a structure shift before entering is how the framework tries to handle this, and it does not fully solve it.
- The entry zone is never respected. Price displaces, retraces, and goes straight through the order block or gap. The zone was a guess about where interest sat, and the guess was wrong.
- The chart is relabelled until it works. Because there are many swing points, many candles that qualify as order blocks and several timeframes to choose from, a determined trader can find a valid-looking setup on almost any chart in hindsight. A framework that explains every move explains none of them in advance.
The honest position on reliability is that nobody can give you a number. No independent test of these setups exists at scale, and the published results are self-reported. Whether they work in your hands, on your instrument, in your session is an empirical question about your own trading.
Common mistakes
- Treating the vocabulary as evidence. Naming a candle an order block does not establish that institutional orders sit there. The label is a hypothesis, and deserves the same scepticism as any other zone.
- Stacking timeframes until something confirms. Dropping to the 1-minute to find a structure shift the 15-minute does not give you is confirmation-seeking. Fix your timeframes before the session.
- Ignoring the spread. Sweeps often happen in thin conditions, which is when spreads widen. A stop 2 pips beyond a wick can be taken out by the spread alone.
- Confusing the framework with a plan. Concepts are not entry, stop or sizing rules. Those you write down yourself.
How it relates to other approaches
Most of ICT's structural content overlaps with older ideas under different names. Marking zones where price moved away sharply is what supply and demand trading does. The three-phase story of a range being built, breached and then trended out of draws on the Wyckoff schematic, which the framework restates as accumulation, manipulation and distribution. Premium and discount are the halves of a retracement โ the area above or below the midpoint of a marked range โ and market structure is the textbook definition of a trend.
That overlap tells you what is actually being tested when you test the framework. If you already trade zones and trend structure, adopting the vocabulary changes what you call things more than what you do; the distinctive additions are the emphasis on a sweep preceding the real move, and the time-of-day filter.
Frequently asked questions
Is ICT trading legitimate?
It describes some real market behaviour โ stops do cluster, imbalances exist, volatility concentrates by session. But it is not an established or verified methodology, its terminology is proprietary rather than standard, and its performance claims are self-reported. Treat it as one trader's map, not a description of how markets are known to work.
What does ICT stand for?
Inner Circle Trader, the teaching brand of Michael J. Huddleston, who developed and popularised the terminology through a large body of free video material. The concepts are widely reused and renamed by other educators.
Can ICT concepts be backtested?
Partly. Mechanically defined elements โ a three-candle gap, a break of a prior high, a session window โ can be coded and tested. The discretionary parts, such as which order block matters, resist it because two traders will mark them differently. Manual backtesting over a large sample with your own written rules is the practical route.
Why do ICT setups look obvious in hindsight and hard in real time?
The finished chart shows you which sweep was followed by a reversal. In real time you see the sweep without knowing whether the reversal is coming, and the candles that would confirm it are still forming. That gap exists in every discretionary framework, and it is the reason to test on unseen data rather than by scrolling back through charts you have already read.
Making it measurable
Any framework that relies on judgement needs a record, because judgement drifts and memory edits. If you trade these concepts, define in writing what qualifies as a sweep, what counts as displacement, which timeframes you use and what you risk per trade โ before the session, not during it. Then log every trade.
After a large enough sample you will have what the framework itself cannot give you: your own numbers. A trading journal that tags each trade by setup type shows whether the sweep-and-shift entries carry their weight or whether one variant is subsidising the rest, and account analysis shows whether the equity curve is shaped by your read or by inconsistent position size. Fix sizing first โ a consistent risk fraction calculated from the stop distance with the lot size calculator โ so that when you compare setups you are comparing setups and not bet sizes.
That is the honest use of a discretionary framework: it gives you a vocabulary and a set of hypotheses, and the evidence has to come from your own record.