ICT entry models: the FVG ICT entry

The three possible entries, and what they cost

There is no universal entry on an imbalance, and being sold the opposite is the best way to lose money with an otherwise solid concept. Three approaches coexist, from the most aggressive to the most patient.

The first places a limit order at the edge of the zone with no confirmation. A FVG ICT entry usually means exactly this: sitting inside the gap on the assumption that price will pay the zone a visit and leave. You get filled often, and you get run over often.

The second enters at the midpoint. You let price penetrate before acting, which halves your risk when the hole is wide, and costs you every shallow return.

The third waits for a confirmed reaction: price shows it refuses to go further, and you enter after. You get filled far less often, with a wider stop and better information.

⚠️ None of the three is better in the abstract. They move the dial between frequency and quality, and the only valid arbitration is the one your own record gives you on your own market.

One move, three orders, three payoffs Exactly the same candles, the same invalidation stop at 22 and the same target at 96: only the entry price changes. The edge fills as soon as price grazes the zone and pays 1.6 times the risk; the midpoint pays double and stays empty every time price turns before reaching it; confirmation waits until price has shown its refusal, and hands back almost all of the payoff to the wider stop it demands. The choice is not an opinion, it is a costed trade-off between frequency and payoff.

The usual rules, and which ones can be checked

The rules of FVGs ICT trading teaches are few and easy to state. The gap must have been left by a decisive move rather than by drift. It must not have been filled already. And the higher timeframe should agree with the direction you are taking.

The first can be checked on the spot: the speed of a move is readable on the candles as they form. So can the second, you only have to look left.

The third is slipperier. "Higher timeframe bias" is a reading, not a data point, and two honest people can draw opposite conclusions from the same chart.

That is why it has to be written down beforehand in a form that cannot be renegotiated: a level taken out, a specific break of structure, a close above a price. Not an impression.

A full FVG ICT entry example fits in one sentence: a gap left during the London session, revisited during New York, entered at its edge with the stop beyond the far side. It is concrete, and it is measurable.

What the eye can check: the speed, and the state of the gap The first two rules are facts, and facts can be seen: on the left, a long body left a clean hole nobody came back to fill; on the right, six limp candles left two points of emptiness that a wick has already closed. No interpretation is required, and that is exactly why they can be checked. The third rule, higher timeframe bias, appears on neither panel: it is not a fact but a reading, and two honest people draw opposite conclusions from it. Write it down beforehand as a level taken out or a close above a price, or it will be renegotiated at the exact moment you need it.

A retracement into an FVG is the setup, not the warning

A retracement into an FVG is the setup itself, not a sign that something is going wrong. That deserves stating plainly, because the phrase sounds like a warning.

Plenty of beginners close a position on the very move the method was waiting for. They see price coming back towards their zone, read it as failure, and exit just before the reaction.

Every entry of this kind depends on that retracement happening at all, and that is the part nobody can promise. A gap that is never revisited produces no trade, and there are more of those than the examples show.

The practical consequence is budgeted patience. If your method requires a return, your day holds fewer opportunities than you imagine, and forcing an entry elsewhere so as not to have waited for nothing is the most common way to lose what the method would have paid you.

What you can measure is the proportion. Across thirty imbalances marked in advance, how many were revisited within two days? The number is yours, and it decides the size of your annual sample.

Two futures for the same imbalance From the high, nothing is settled yet. Price may never come back: that case produces no trade at all, and it is more common than the examples show. If it does come back, the move down into the zone is not the setup failing, it IS the setup; those who close at that moment close on the very move they were waiting for. The only thing you can measure here is not the reaction, it is the share of your pre-marked imbalances that actually get revisited, because that share decides how many opportunities your year holds.

The midpoint as an entry

The midpoint of the imbalance has a name, consequent encroachment, and a whole school uses it as the entry. The idea is that a genuine imbalance rarely needs filling entirely.

The arithmetic is simple: entering at the midpoint halves your risk when the gap is wide, at an identical stop. On a twenty point zone you go from twenty points of risk to ten, and your reward-to-risk doubles.

What it costs you is just as simple: every time price turns at the edge without going further. Those trades you do not take, and some of them would have been your best.

The trade-off is therefore arithmetic, not philosophical. Mark both levels in advance across thirty setups, note which would have been touched and what happened next, and compare the two columns.

The full treatment of this point is in the lesson on the inverse fair value gap, which defines the object this lesson merely puts to work.

The stop, and why it does not go under the zone

The stop is not placed at random under the imbalance. It is placed where your reading becomes false, which is to say beyond the point that would cancel the very reason for your entry.

If you enter because the gap marks a continuation after a break of structure, your stop goes below the point that would undo that break. Not three points under the edge because that feels better.

The distinction sounds theoretical and it is not at all. A stop placed for psychological comfort gets hunted, then the trade leaves without you, and you have paid the loss without collecting the gain.

A stop placed at invalidation costs more when it goes, and it only goes when you were wrong. That is the difference between a risk you choose and a risk the market chooses for you.

The detail is in our guide on where to place your stop loss, and it holds for every zone in this category, not only for imbalances.

Two stops on the same entry, only one survives Same entry at 62, same target at 94, two stops. The comfort stop sits three points under the edge because that is where it hurts least: it risks only 19 points, promises 1.7 times the risk, and it does not survive the first wick, which reaches down to 38 before the move leaves without you. The invalidation stop costs 32 points instead of 19 and pays only one times the risk, but it would only have been hit if the low that created the break had given way, which is to say only if your reading was wrong. A stop is chosen on what would prove you wrong, never on what you are willing to lose.

Practising: thirty entries, two more columns

Take the record of imbalances marked blind. If it does not exist yet, build it first: without it, what follows measures nothing.

For each of the thirty cases, add two columns: did price touch the edge, did it touch the midpoint. You immediately get the relative frequency of the two entries.

Then add the result each one would have produced, with the same stop and the same target. That is the only calculation that genuinely compares the two approaches, and it fits in a spreadsheet.

Repeat with the third approach, the entry after confirmation, noting the price you would have got. You will see the stop widen and the number of cases shrink, which is exactly the trade-off announced.

Sort the whole thing by entry type in a trading journal. After thirty cases you will have an answer, after a hundred you will have a decision.

Key takeaways

Three entries coexist: the edge, the midpoint, confirmation. They move the dial between frequency and quality, none is better in the abstract.
Of the three usual rules, two can be checked on the spot. The third, higher timeframe bias, must be written beforehand in a form that cannot be renegotiated.
A retracement into the zone IS the setup, not a warning. Many close on the very move the method was waiting for.
Entering at the midpoint halves the risk when the gap is wide, and costs every turn at the edge. That is arithmetic, and it can be counted.
The stop goes at invalidation, never at comfort. A comfort stop gets hunted, then the trade leaves without you.
A gap never revisited produces no trade. Measure the return rate before building a day around this method.

Going further

These blog articles dig into this lesson's ideas, one subject per article.

IFVG vs FVG: how to read a fair value gap and its inverse Stop loss strategy: where to place it right Supply and demand zones explained Trading journal: the complete guide to improving (2026)