The short answer
An order block trading strategy is the set of rules that turns the ICT concept of an order block into a repeatable trade, and the entire edge hangs on filtering the few blocks worth trading from the dozens that fail. Most traders lose on this concept not because order blocks are fiction, but because they trade every one they see.
The full breakdown of what an order block is, where the idea comes from, and the supply-and-demand lineage sits on the order blocks explained page. This page assumes you know that and focuses only on how to trade the thing.
I treat the strategy as a filter problem first and an entry problem second. Get the filter right and the entry almost does not matter, but get the filter wrong and no entry method will save you.
If the wider method is new, the Smart Money Concepts hub lays out the framework this strategy sits inside.
What is actually proven, and what is ICT doctrine
One framing before the mechanics, because it will save you money. Order blocks belong to ICT methodology, the work of trader Michael Huddleston, known as The Inner Circle Trader, and they are a labelled reading of a chart rather than a measurement of any real order book.
The part with external evidence is the clustering. Carol Osler's research at the Federal Reserve Bank of New York found that currency stop-loss and take-profit orders pile up at round numbers and prior swing points, and that triggering those clusters produces the sharp, self-reinforcing moves traders call displacements and sweeps.
That documented clustering is the real mechanism behind why obvious zones get revisited and why breaks of them run. What is not proven is any specific win rate for order blocks, and the confident percentages floating around the trading web trace to a single unsourced blog, so I treat them as lore and never repeat them.
I anchor the strategy to what Osler documented and trade the pattern I can see, not the institutional story about who is doing it or why. The mechanics below work whether or not that narrative is true, and that separation is what keeps the strategy honest.
The four-part filter for a high-probability order block
Every order block I trade has to pass four checks, and a block that fails any one of them is noise I leave alone. The checks are a break of structure, displacement, an imbalance, and an unmitigated state.
| Check | What it means | How to read it |
|---|---|---|
| Break of structure | A genuine break of a prior swing in the block's direction | Price closed beyond the most recent opposing swing |
| Displacement | A strong, impulsive move away from the block | A large-body candle, not a slow drift |
| Imbalance | A fair value gap left behind by the displacement | The wicks of candle one and candle three do not overlap |
| Unmitigated | Price has not returned to fill the block since it formed | The block's range is still intact |
I run the four checks in order, because the first one disqualifies most candidates before I waste time on the rest. A level with no structural break behind it is just a candle, and that single rule removes most of the chart.
The filter is the whole edge. Traders who lose on order blocks almost always pass a block that failed one of these four and then blame the concept, when the concept told them to skip the trade.
How to identify a bullish order block
A bullish order block is the last down-candle before a strong upward move that breaks structure, and ICT traders read it as the footprint of buying that launched the rally. The single candle is the marker, and the break above the prior high is what validates it.
I mark the block only after the break, never before. Drawing a box around a down-candle and hoping price comes back is how traders invent order blocks the market never confirmed, and it is the most common way beginners fool themselves.
The cleanest bullish blocks show a long lower wick and a small body, which tells me buying absorbed the selling before the displacement higher. A block with a large red body and no lower wick is weaker, because it signals unrelieved selling rather than absorption.
How to identify a bearish order block
A bearish order block is the mirror image: the last up-candle before a strong downward move that breaks structure below the prior low. ICT traders read it as the footprint of selling that launched the decline.
Again I mark the block only after the break confirms it, because a green candle on its own is not a block. The break is the evidence that the candle actually mattered to the market.
The strongest bearish blocks show a long upper wick and a small body, the sign that selling absorbed the buying before the fall. A block with a thick green body and no upper wick is weaker, because it shows buyers were never really trapped at that level.
External and internal order blocks
ICT traders split order blocks into external and internal, and the distinction decides which ones you trade. An external order block is the block that started the move on the higher timeframe, while an internal order block is one that formed inside that move on a lower timeframe.
I treat external blocks as the premium setups and internal blocks as lower-probability retest entries. Trading an internal block against the direction of the external block is the most common way traders pick a bad one.
The rule I follow is simple. The external block sets the bias, and the internal block offers the timing, so if the two disagree I stand aside rather than force the lower-timeframe block.
The entry: three ways to take the trade
Once a block passes the four-part filter, there are three honest ways to enter it, and each one trades risk for confirmation. The right choice depends on your temperament and the market condition, not on which method is supposedly best.
The limit entry rests a pending order at the near edge of the block and assumes price returns to fill it. It earns the best risk-to-reward but offers no proof the block will hold, so it fails the most often.
The confirmation entry waits for price to tap the block and then watches for a lower-timeframe shift in your favour before entering. It gives up some price for evidence, and it is the method I trade most often.
The break-and-retest entry skips the first return and waits for price to break through the block, fail, and retest it from the other side. It is the most conservative and the latest, so it cleanly misses the moves that never retest.
| Entry | How it works | Best when | The cost |
|---|---|---|---|
| Limit at the block | A resting order at the block's near edge | The block is external and fresh | No proof the block holds |
| Confirmation | Tap the block, then enter on a lower-timeframe shift | You want evidence before risking money | A worse entry price |
| Break and retest | Wait for the block to break, fail, and retest | Choppy conditions where first touches fail | Misses moves that never retest |
No entry is free. The limit entry trades confirmation for price, the confirmation entry trades price for safety, and the retest entry trades both for certainty, so I pick the compromise I can actually stick to under pressure.
Stop loss, take profit and the R-multiple math
Most order block pages tell you to put a stop beyond the block and leave it there, which is useless when you actually have to size a position. I work every trade in multiples of risk, called R, because it is the only honest way to compare one setup to another.
One R is the amount you risk on the trade, measured as the distance from your entry to your stop. A three-R winner risks one to make three, and a one-R loser risks one to lose one, and those two numbers are the entire game.
I place the stop just beyond the order block, past the wick that swept into it, so a normal probe does not stop me out but a real break of the block does. The target sits at the next opposing pool of liquidity, which is where the move is said to be heading.
The risk-to-reward ratio falls out of those two distances for free. If the stop is twenty points and the target is ninety-five points, the trade is close to five R, and I know that before I ever click into the market.
A worked long trade, with the numbers
The math only clicks when you walk through a real one, so here is a bullish order block long on euro-dollar, step by step. The numbers are illustrative, but the ratios are how I actually size and manage the trade.
Suppose the four-hour chart breaks structure to the upside off a down-candle at 1.0850, leaves a fair value gap below it, and prints a bullish order block from 1.0850 down to 1.0830. The block passes the four-part filter, and the next major buy-side liquidity sits at the prior swing high of 1.0940.
I wait for price to retrace into the block, and it taps 1.0840. On the fifteen-minute chart price prints a change of character and breaks a minor lower high, which is my confirmation, and I enter long at 1.0845 with a stop at 1.0825, just below the block's low.
The risk is twenty points, which is one R. The target at 1.0940 is ninety-five points from entry, which is nearly five R, so the trade risks one to make nearly five.
If price closes back below the block low instead, the stop removes me for a one-R loss and I wait for the next setup.
I risk a fixed fraction of the account per trade, usually around one percent, so the position size is whatever that one percent buys at twenty points of risk. The block, the entry, the stop, and the target are all decided before the trade, never during it.
Confluences that filter out false order blocks
A block that passes the four-part filter still fails often enough that I want a second layer of evidence, and that layer is confluence. The best order blocks line up with at least two of these before I risk money on them.
A fair value gap inside the block shows the displacement left real inefficiency behind, and a liquidity sweep just before the block forms shows stops were cleared first. The sweep is the fuel that Osler's research describes, the cluster of triggered orders that drives the sharp move.
Higher-timeframe alignment means the block points the same way as the daily trend, and session timing means it formed inside a killzone, when the largest moves tend to launch. A block with two of these confluences is a trade, and a block with all four is the setup I actually wait for.
| Confluence | What it adds | How to check it |
|---|---|---|
| Fair value gap inside the block | Proves the displacement left inefficiency | The block contains a three-candle gap |
| Liquidity sweep first | Shows stops were cleared as fuel | An obvious high or low was taken before the block |
| Higher-timeframe alignment | Confirms the block trades with the trend | The daily and four-hour bias agree |
| Killzone timing | Aligns the block with high-flow windows | It formed in the London or New York killzone |
I score the confluences rather than count them loosely, because two strong ones beat four weak ones. The table is a checklist I run the same way every time, and the discipline of it is what separates a real setup from a hopeful one.
Multi-timeframe alignment without the confusion
Multi-timeframe analysis confuses more order block traders than it helps, because they stack too many charts and talk themselves out of every trade. The fix is to use exactly two timeframes, one for bias and one for entry.
I set the bias on the four-hour or daily chart, where I mark the external order blocks and decide whether I am looking for longs or shorts. Then I drop to the fifteen-minute or five-minute chart only to refine the entry inside that already-decided bias.
The rule is that the lower timeframe can only ever refine the higher-timeframe plan, never overrule it. A clean lower-timeframe block against a higher-timeframe trend is a trap, not an opportunity, and I skip it every single time.
If the two timeframes conflict, the honest move is to stand aside. Forcing a trade through conflicting timeframes is how a promising block turns into a losing one, and the best trade is often no trade at all.
How to backtest the strategy before you risk money
No order block page I have found teaches backtesting properly, which is why most traders never know whether their version of the strategy actually has an edge. Backtesting is how you find out, and it costs nothing but time.
I scroll back through at least one hundred occurrences of the setup on a single market and log each one the same way. For every block I record the date, the market, whether it passed the four-part filter, the entry method, the result in R, and the confluences present.
The number that matters is the expectancy, which is the average R across all the trades, and a strategy with positive expectancy over one hundred trades has at least some evidence behind it. A strategy tested on ten cherry-picked examples proves nothing, and I treat any win-rate claim from a sample that small as marketing.
I backtest on the exact market and timeframe I intend to trade, because an order block on a currency pair does not behave like one on a crypto perpetual. The behaviour is the only thing that matters, and you only see it by doing the work yourself.
The honest framing on win rates
Every few months a new order block win-rate figure circulates, and the current one claims around fifty-two percent standalone and the mid-sixties with a break-of-structure filter. I traced that figure to a single unnamed blog with no methodology and no published dataset.
No peer-reviewed study has tested ICT order blocks, and no broker or exchange has published a backtest of them, so any specific percentage is invention. The only honest statement is that no published fill-rate or win-rate study exists for this concept.
I trade the strategy because the underlying mechanism, the clustering of orders at obvious levels, is documented in Osler's work at the New York Fed, not because of a percentage somebody made up. Confidence built on a fabricated number breaks the first time the market disagrees with it, and the market always disagrees eventually.
Common mistakes that sink order block trades
The losses on this strategy cluster around the same handful of errors, and avoiding them does more for your results than finding better blocks. These five are the ones I see again and again.
Trading every break of structure is the first. Most breaks do not leave tradeable order blocks, and treating each one as a setup floods your chart with garbage that stops you out.
Entering on the block with no confirmation is the second, and it puts you in just before the move continues instead of reverses. The block is a zone of interest, not a signal to click buy.
Ignoring the higher timeframe is the third. A bullish block inside a bearish daily trend is fighting the current, and those are the blocks that stop traders out over and over.
Drawing the block before the break is the fourth, and it invents order blocks the market never confirmed. Mark the block only after the structure breaks, or you are guessing.
Trading with no risk plan is the fifth and the deadliest. A strategy with no fixed risk per trade and no daily loss limit is not a strategy, it is gambling with extra steps, and the market collects on that debt quickly.
