- The stop marks where you are wrong, not what you can afford to lose. It belongs beyond the price that would disprove your read — past the low that swept liquidity, past the origin of the order block. If that distance is too expensive, the answer is a smaller position or no position, never a closer stop.
- The target is a place, not a multiple. “Two R” is not an ICT concept. The trade is going somewhere specific — the opposing liquidity pool, the previous day’s high, the far side of the dealing range. If you cannot name the destination before you enter, you do not yet have a trade.
- What triggers the move to breakeven decides whether it is sound. At a structural level such as T1 — the first liquidity objective — it reflects a premise that has partly completed. Done because price has moved your way, it replaces a reasoned invalidation with an arbitrary one and converts winners into scratches while leaving losers full size.
- Time invalidates a trade as surely as price does. A setup taken because the kill zone was open stops being that setup when the kill zone closes. This is the exit rule most specific to ICT and the one almost nobody writes down.
Search for how to find an ICT entry and you will drown. Fair value gaps, order blocks, breakers, optimal trade entry, consequent encroachment, the sweep, the shift, the displacement — hundreds of hours of material, all of it about the moment you get in.
Now search for what to do afterwards. The results thin out immediately, and most of what survives is generic trading advice with an ICT label stapled on: move to breakeven at one R, take partials at two, trail behind the swing. None of it is drawn from the methodology. It is the same boilerplate every trading site publishes, and it fits an ICT trade badly.
That asymmetry is not an accident. Entries are teachable as patterns — a gap looks like a gap, a sweep looks like a sweep, you can put a box on a chart and point. Exits are contextual, and context does not screenshot well. So the entry got a thousand lessons and the exit got a shrug.
It is also, for a lot of people, where the money actually goes. You can read the market correctly, enter at a sensible price, and still finish the month down, because the management was invented in the moment under pressure. This page is about the decisions that happen after the fill.
Nothing here will make a bad read profitable. Management determines how much a correct read is worth and how much an incorrect one costs. It does not manufacture an edge that was not there, and no arrangement of stops and targets rescues a trade that had no reason behind it. If you are still taking the first gap you see without a bias, start with why ICT trades fail instead — this page assumes the read is sound.
The stop marks where you are wrong
Ask most traders where their stop goes and you will get a number: twenty points, thirty pips, whatever the account tolerates. Ask an ICT trader the same question and the honest answer is a location — and the location is decided by the structure that produced the setup, not by the balance in the account.
The distinction matters more than it sounds. A stop placed at a tolerable distance answers the question how much am I willing to lose. A stop placed beyond the invalidation point answers a different question: at what price is my reason for being in this trade no longer true? Those two prices are unrelated, and when they conflict, the account has to bend around the chart rather than the other way round.
Where the invalidation actually sits
Every ICT entry has a price that disproves it. It is usually obvious once you look for it:
| Entry premise | What disproves it | Where the stop sits |
|---|---|---|
| Long from a bullish order block after a sell-side sweep | Price trading back below the low that did the sweeping | Beyond that swept low, not the order block’s edge |
| Long from a bullish fair value gap in a discount | The gap filling completely and closing through it | Beyond the far edge of the gap |
| Short after a market structure shift down | A body close back above the swing high that the shift broke | Above that swing high |
| Long from a breaker | Price closing back through the breaker in the old direction | Beyond the breaker’s origin |
| Long from OTE inside a dealing range | A close below the low that defined the range | Beyond the range low |
Each of these is worked through in detail, with the buffer question and the timeframe problem, in where to place your stop loss. Notice what is common to all five: the stop sits beyond a structural price — a low, a high, an edge — and not at a round distance from entry. Notice also that in every case the stop is further away than a trader wanting a tight risk would like. That is not a flaw in the method. That is the method telling you what the trade costs.
The tight stop is a donation
The most common self-inflicted wound in ICT trading is the stop placed inside the structure rather than beyond it — a few points under the entry candle, at the edge of the gap, just below the order block rather than below the low that preceded it.
This feels disciplined. It is the opposite. The entire premise of the methodology is that price seeks out resting orders before it delivers in the intended direction. A stop parked just under an obvious level is a resting order at an obvious level. You have placed it exactly where the model says price goes looking.
People who do this describe a familiar experience: stopped out by two points, then the move runs without them. They usually conclude their read was right but their luck was bad. The read was right. The stop was in the pool. If this is your recurring pattern, why price reverses right after you enter covers the mechanism in full, and the liquidity sweep explains why those particular points attract price.
When the structural stop is too expensive
Sooner or later the honest stop is forty points away and your risk budget allows twelve. This is the moment that separates a plan from an improvisation, and there are exactly three legitimate responses.
- Reduce the size. Risk is distance multiplied by position size. If the distance is fixed by the chart, the size is the only variable you are allowed to touch. The position size calculator does this arithmetic, and position sizing covers it properly.
- Trade a smaller instrument. A forty-point stop on ES is $2,000. The same stop on MES is $200. Same read, same chart, a tenth of the exposure. Trading ICT with a small account works through the numbers.
- Skip it. If the position needed to make the stop affordable is smaller than your platform’s minimum, the trade is not available to you today. This is a real answer and a perfectly respectable one.
What is not on the list is moving the stop closer. That converts a trade with a defined invalidation into a coin flip with a tidy-looking risk number, which is worse than not trading at all — it costs money and teaches you nothing, because when it loses you will not know whether the read was wrong or the stop was simply in the way.
The target is a destination, not a multiple
The second habit imported wholesale from generic trading education is the R multiple as a target. Take profit at two R, or three, or whatever the rule says.
Nothing in ICT works this way. The methodology is built on the idea that price moves from one pool of liquidity to another. A target is therefore a place on the chart, and the R multiple is simply what that place happens to be worth once you measure it. The multiple is an output, not an input.
The practical consequence is a discipline that sounds obvious and is routinely skipped: you identify the destination before you enter. If you cannot point at where this trade is trying to go and say why, you have found a pattern rather than a trade.
Where the draw usually is
| Candidate destination | When it is the likely draw |
|---|---|
| Previous day’s high or low | The default intraday objective when the day opens inside yesterday’s range |
| Equal highs or equal lows | Obvious resting liquidity; often the cleanest and most cited draw |
| Buy-side or sell-side liquidity above/below session extremes | When the session has built an evident pool on one side |
| The opposite end of the dealing range | When price is working within a clearly defined range |
| An unfilled higher-timeframe fair value gap | When the HTF chart has an obvious inefficiency overhead or below |
| The midnight open or a session open level | Frequently a magnet on days without a stronger external draw |
Where to take profit works through how to choose between these when several qualify, and what to do about obstacles sitting in the path. The distinction between internal and external range liquidity is the most useful filter here. Internal objectives — a gap inside the range — are nearer and more frequently reached. External objectives — the old high, the equal lows — are further and pay more when they arrive. Deciding which one you are trading for, before entry, settles most management arguments in advance.
Before entering, finish this sentence out loud: “I am long from here because price swept that, and I expect it to trade to that.” If either blank will not fill, close the platform. A trade you cannot describe in that sentence is a trade you will manage badly under pressure, because there is nothing to manage it against.
Breakeven: the move that feels responsible
Price runs your way. You move the stop to entry. Nothing can hurt you now.
Except that you have just done something quite strange, and it is worth being precise about what. You spent real effort locating a price that would disprove your read, and you placed a stop beyond it. Then, on the basis of favourable movement alone, you replaced that reasoned price with an arbitrary one — the number you happened to get filled at — and you did it without any new information about whether the idea was still valid.
Your entry price is not a structural level. The market has never heard of it. Moving your stop there converts a trade with a defined invalidation back into the exact thing this page opened by arguing against.
The asymmetry
The deeper problem is what it does to the shape of your results. A stop at breakeven turns trades that would have been winners into scratches. It does not turn any losers into scratches — a loser goes to the stop either way. So the practice removes outcomes from the upper half of your distribution and none from the lower half.
| Scenario | Left alone | Moved to breakeven early |
|---|---|---|
| Runs straight to target | Full win | Full win |
| Retraces into entry, then runs to target | Full win | Scratch |
| Retraces into entry, then fails | Full loss | Scratch |
| Goes straight against you | Full loss | Full loss |
One row out of four improves. One gets materially worse. The other two are unchanged. Whether that trade is worth making depends entirely on how often the second row happens relative to the third — and the awkward fact about ICT setups specifically is that the second row is common. Retracing into the entry zone before delivery is not an anomaly in this methodology; it is a described behaviour. Consequent encroachment, the fifty per cent level of a gap, the second tap of an order block — the methodology explicitly expects price to come back before it goes.
Which means a breakeven stop sits directly in the path of a move the model predicts.
That table is a logical decomposition, not a performance claim. It shows which outcomes change under each policy — it does not tell you how often each row occurs, because that depends on your model, your market and your timeframe. Establishing those frequencies is your own work, and backtesting is how it gets done. Where this site does quote figures — in risk management and backtesting — they are illustrations of how expectancy arithmetic works, not results anyone achieved. No page here publishes a track record, and you should be wary of any that does.
When breakeven is genuinely right
None of which makes the move always wrong. It makes the trigger wrong. “Price moved in my favour” is not information about validity. These are:
- The draw has been reached. You entered expecting a run to the previous day’s high, and it got there. The reason for the trade is complete. Anything after this is a new trade, and it should be sized and justified as one.
- Structure has shifted against you. You are long, and the lower timeframe puts in a clean shift downward. The conditions that produced the entry no longer hold.
- The window has closed. Covered in the next section, and the most underrated of the three.
- The news environment changed. A high-impact release you had not accounted for is minutes away and your premise did not include it.
Where this leaves the T1/T2 split
ICT teaching contains a widely taught management rule: take half off at T1 and move the remaining half to breakeven. Risk management sets it out on this site, and it is worth being clear that nothing above contradicts it — provided T1 is what it is supposed to be.
T1 is the first internal range liquidity objective. It is a place. So taking something there and moving the remainder to breakeven is not an arbitrary act triggered by favourable movement — it is the first item on the list above, the draw being reached. The premise has partly completed, and the management reflects that.
What the rule cannot survive is being detached from the level and applied as a distance. “Breakeven at T1” is structural. “Breakeven at 1R” wearing T1’s name is not, and the two get conflated constantly, because one is easy to mark on a chart and the other is easy to type into a platform. If the level you are calling T1 is a multiple of your stop rather than a pool of liquidity you identified before entering, the objection in this section applies to it in full.
The test is simple and it is not about price: has something happened that changes whether my reason for being here is still true? If yes, tightening or exiting is justified. If the only thing that happened is that you are now up and would rather not give it back, that is discomfort, and discomfort is not a signal. Trading psychology deals with the difference at length, because it is the difference that costs most people the most money.
Partials wear the same costume
Taking half off at a fixed multiple has the identical structure to the breakeven problem, and for the identical reason. Your losers close in full. Your winners close in halves. You have introduced an asymmetry that works against you, in exchange for feeling better while the trade is open.
There is a version that survives scrutiny. It is the one where the partial is taken at a place rather than at a multiple:
- There is a genuine intermediate pool of liquidity between entry and final objective — an internal gap, a minor session high — and you take something there because price has a documented tendency to react at it.
- You are holding for an external draw but the internal objective has been met, and you are reducing exposure to the part of the move you cannot justify.
- The position is larger than your normal size because two setups aligned, and you are returning to normal size once the first premise completes.
Each is a decision about the chart. “Half at 1R” is a decision about your nerves. The former is management; the latter is a tax on your best trades.
Time invalidates a trade too
Here is the exit rule most specific to ICT, and the one you will struggle to find written down anywhere.
This methodology is time-anchored to an unusual degree. Kill zones exist because certain windows are held to behave differently. Macros are twenty-minute windows. The Silver Bullet is defined by an hour. If you accept the premise that a window mattered enough to justify the entry, you have to accept the corollary: when the window closes, the reason you entered has expired — whether or not the trade has reached its target, and whether or not it has hit your stop.
A New York AM trade taken at 09:50 because the macro was open is a macro trade. At 11:30 it is no longer a macro trade. It is an open position in lunchtime, a period the methodology treats as low quality and best avoided. Nothing about the position changed. Everything about the environment did.
| Why you entered | When the premise expires |
|---|---|
| A macro window | At the end of that twenty-minute window |
| A kill zone setup | When the kill zone closes |
| A New York AM session trade | Entering lunch without having reached the draw |
| An intraday trade of any kind | The session close — overnight is a different market with different participants |
| A Judas swing reversal | Once the session’s directional move is clearly underway or over |
None of this obliges you to close at a fixed clock time. It obliges you to re-justify the position when the window that produced it closes. If you would not enter this trade now, on what is in front of you now, the question of why you are still in it deserves a real answer. “Because I am already in” is not one.
The argument for the rule is not that it catches more winners. It is that the positions which turn into outsized losses are, very often, the ones held long past the context that justified opening them — and a window that has demonstrably closed is one of the few invalidation signals available to you that does not depend on reading price correctly a second time.
Walkthrough — one trade, all four decisions
The numbers below are constructed to show the decision points cleanly. It is a teaching illustration rather than a record of a trade.
Pre-session. Bias long. Previous day’s low at 21,402 is the expected sweep; the draw is the previous day’s high at 21,566. A 15M bullish order block sits at 21,428–21,445. The plan, written before the open: long after a sweep of 21,402 with a 1M shift inside the block; stop beyond the sweep low; first objective the session high at 21,498, then 21,566; premise expires at the end of the AM session.
10:09 — the stop is decided before the entry. Price sweeps to 21,396 and turns. The invalidation is now fixed: below 21,396, the sweep did not hold and the reason for the trade is gone. Everything else follows from that number rather than from what the position is worth.
10:23 — entry at 21,443 after a 1M shift confirms inside the block. Stop at 21,392, four points beyond the sweep low. That is 51 points of risk — $1,020 on NQ, $102 on MNQ. The size comes from that figure, not from a preference.
10:41 — the first objective. Price reaches the session high at 21,498. This is not “the trade is up 55 points.” It is a named level being reached: the premise has partly completed. Half comes off and the remainder moves to 21,447, a few ticks above entry so a scratch costs nothing. That is the T1 trigger doing its job, and it is structural rather than a distance.
11:14 — nothing has happened. Price has drifted between 21,470 and 21,500 for half an hour. It has not reached 21,566 and it has not come back to the stop. There is no price-based decision available here, which is exactly the situation most management advice has nothing to say about.
11:52 — the fourth decision. The AM session is ending. The premise was a morning continuation to the previous day’s high, and that window is closing with the objective unreached. Nothing about price has invalidated the idea. The context has. The remaining half is closed at 21,486 rather than carried into lunch, where the methodology expects ranges to contract and the reason for holding no longer exists.
What the four decisions were. The stop came from structure, at 10:09, before anything was at risk. The target was a named level, chosen pre-entry. The move to breakeven was triggered by reaching an objective, not by comfort. And the exit came from time, because the window that justified the position had closed. Not one of those was decided while the trade was live and the temptation to improvise was highest.
The four ways a trade dies
Pulling the page together: a trade ends for one of four reasons, and knowing which one you are looking at is most of management.
| Type | What happened | Correct response |
|---|---|---|
| Price | The invalidation level traded | The stop does it for you. No decision required — which is the point of having one |
| Objective | The draw was reached | Close, or explicitly re-enter as a new trade with a new premise |
| Structure | The market shifted against the premise before reaching either | Exit. Waiting for the stop now is hope, not a plan |
| Time | The window that justified entry has closed | Re-justify or exit. Usually exit |
Only the first is automatic. The other three require you to be watching and honest, which is exactly why they need to be written down before the trade rather than negotiated during it.
Decide it before you need it
Every rule on this page has the same failure mode: it is easy to agree with while reading and nearly impossible to apply for the first time with money at risk. Under pressure you do not reason. You do what you decided earlier, or you improvise — and improvisation under pressure reliably produces the tight stop, the early breakeven, the partial taken out of nerves.
So management is written down before entry, as part of the setup, in the same breath as the entry criteria:
- Where the invalidation sits, as a price, and therefore where the stop goes
- The size that makes that stop affordable — not the stop that makes your preferred size affordable
- The destination, named as a level
- The window: when does this premise expire
- What, specifically, would make you exit before either the stop or the target
Five lines. If you cannot write them before clicking, the trade is not ready. Backtesting is where those five lines get calibrated for your own model, and it is the only way to learn which of these rules actually matters in the setups you take, as opposed to the ones you read about.
Nobody can tell you where your stop goes, because it depends on a setup only you are looking at. What is generalisable is the reasoning: the stop is a structural price, the target is a location, the reasons to exit early are informational rather than emotional, and time is one of those reasons. Applied consistently, that reasoning at least makes your exits a function of the chart. Applied inconsistently, they become a function of your mood, and you lose the ability to tell which of the two your record is measuring.