IFVG trading: the inverse fair value gap

The imbalance, and the hole it leaves

When price moves too fast for both sides to trade normally, it crosses an area without anyone having had time to deal there. That area is called an imbalance. It is not a theory, it is an observation: the trace is measurable, and two people looking at the same chart find the same one. The move that digs this hole has a name and a lesson of its own, displacement.

When that imbalance reads across three consecutive candles, it takes a second name, fair value gap, shortened to FVG. The definition fits in one sentence: the low of the third candle sits above the high of the first, or the reverse on the way down. The space between the two was traded by nobody.

Everything built on that hole starts from the same idea: a place the market crossed without negotiating is a place it might come back to negotiate. That is plausible, and it is not demonstrated.

The link with what you have read elsewhere is direct. The lesson on large orders and how they execute explains why the hole exists at all: nobody dug it on purpose, it is the trace of an execution that could not wait.

⚠️ An imbalance is not a signal. It is a description of what happened, in the past tense. Everything the rest of this lesson adds is interpretation laid on top of that fact, and it is worth knowing where the line runs.

Bullish FVG, bearish FVG: BISI and SIBI

A fair value gap has a direction, and ICT vocabulary names each one. The bullish FVG is called BISI, buyside imbalance sellside inefficiency: three impulse candles going up delivered nothing but buying, and a hole remains below price, between the high of the first candle and the low of the third, where selling was never offered.

The bearish FVG is its exact mirror: SIBI, sellside imbalance buyside inefficiency. The fall delivered nothing but selling, and the hole sits above price, between the low of the first candle and the high of the third.

The way to stop mixing them up fits in one sentence: the first word of the acronym says which way the move that dug the hole went, the second says what is missing inside it, hence what the market would come back looking for. A BISI sits below price and reads as demand on the return; a SIBI sits above and reads as supply.

⚠️ These are two more names for the same three-candle object, not two new concepts. Everything this lesson says about the fair value gap, the measurement, the return, the invalidation, holds identically for both, and nothing that follows changes with the acronym used.

What an inverse fair value gap is

A fair value gap is supposed to hold: price returns into it, finds the counterparty that was missing, and leaves again. When it does not hold, when price cuts clean through instead of bouncing, the object changes status. It becomes an inverse fair value gap, written IFVG.

That is the whole content of the term. The IFVG meaning comes down to this: a gap that failed. Nothing more, nothing hidden.

The idea behind IFVG trading fits in one sentence too: a level that failed to hold in one direction should now work in the other. A bullish gap cut downward becomes a zone where rejection is expected on a return, and the other way round.

Marking one takes four steps, always in the same order. First, draw the original gap from wick to wick: the high of candle one and the low of candle three for a bullish gap, the reverse for a bearish one. Second, wait for a candle to close beyond the far edge, on the timeframe the gap was drawn on. Third, from that close on, treat the same two prices as a zone of the opposite sign. Fourth, before price comes back to it, write down what cancels the reading: a close back through the zone.

The change of status happens at a precise moment, the one where a candle closes beyond the gap. Before, it is an FVG. After, it is an IFVG. The same drawing, two opposite readings, and one candle between them.

Inversion, or the same event from the other side

The moment of the flip has its own name, and this is where the vocabulary starts costing money. It is called an inversion fair value gap.

It is not a third object. It is the same event named from the other end: the gap does not become something new, it stops being what it was. Reading an inversion FVG and trading an IFVG are therefore one and the same activity.

The confusion comes from the two words describing different instants of the same fact. The inversion fair value gap names the break, the IFVG names the zone once the break is done. Depending on whose work you read, one stands in for the other without warning.

That has a practical consequence and not merely a lexical one: two people commenting on the same chart can each believe they are describing a different setup when they are describing the same candle.

The remedy is the journal, and it is simple. Record the event, not its name: gap left at this time, cut through at that time, returned to or not. Whatever you call it afterwards stops mattering.

The vocabulary, and what it costs a beginner

From one author to the next, the same object goes by a different name, and that spread is the real cost of entry to the subject. A video about an inverse FVG and another about an inversion describe the same gap cut through, not two distinct setups.

A bullish IFVG is nothing more than a bearish gap broken upward. The adjective describes the direction it is used in afterwards, not the one it was created in, and that inversion is the mistake beginners make most often.

Whatever name the author uses, check the definition they give: a gap cut through by a close. If they describe something else, it is their method that differs, not the object.

The proliferation is not innocent. An abundant vocabulary gives the impression of a deep corpus, and above all it makes two methods hard to compare: each can claim to be about something else.

The lesson Getting started with ICT translates the rest of the acronyms the same way. If they still slow you down, go through it before coming back here.

FVG or IFVG: the only difference that counts

The IFVG vs FVG distinction reduces to one thing: an FVG is a hole that has not been traded through yet, an IFVG is one that has. There is no second criterion.

Here is an IFVG example you can check on any chart: price falls and leaves a hole across three candles. Later it climbs back and cuts clean through that hole. The same zone is then watched as support instead of resistance. The drawing has not moved, the reading has flipped.

What that implies is more awkward than it looks: every FVG is eventually traded through. Time therefore turns one into the other mechanically, and the question is never whether this is an IFVG but since when.

That is why the better readings add a constraint of duration or timeframe. A gap cut through within half an hour and a gap cut through three weeks later do not tell the same story, and confusing the two is the first source of disappointing results.

Market structure is exactly what settles that: a gap cut through while structure holds does not say the same thing as a gap cut through at the moment structure gives way.

IFVG vs breaker block: two failed zones, not the same one

The IFVG has a close relative, and the two get confused because they do the same job. A breaker block is an order block that gave way and is then used in the opposite direction; an IFVG is a fair value gap that gave way and is then used in the opposite direction. Same logic, different object.

What differs is what you draw. The gap is a space between two wicks, the high of candle one and the low of candle three, where nothing traded. The order block is a candle, the last opposing one before the impulse, where a great deal traded. An IFVG is therefore a void that was closed through, a breaker a traded range that was closed through, and their edges are not the same prices.

They often sit side by side, which is where the confusion starts. An impulse leaves its order block at the origin and its gap a little further along, so a reversal strong enough to break one frequently breaks the other a few candles apart. The chart then shows two flipped zones stacked on each other, and two authors can label the same reversal with either name.

The stricter definition of a breaker adds a condition the IFVG does not have: the swing built on the order block must have taken out a previous high or low, a sweep, before the block failed. An IFVG asks for nothing of the kind, only a close beyond the gap. That makes the IFVG the more frequent of the two and the breaker the more selective, and frequency is precisely what should not be mistaken for quality.

For your records the rule is short: log them as two setups, with two tags, even when they overlap. Merged into one line they cannot be compared, and you will never know which of the two zones your results came from.

The midpoint, and the CE acronym

The geometric midpoint of the gap has a name of its own, consequent encroachment. Some traders enter there rather than at the edge, on the idea that a genuine imbalance rarely needs to be filled entirely.

In ICT vocabulary, the acronym CE is short for consequent encroachment, and it names nothing more than half of the hole.

The trade-off is real and it can be counted. Entering at the midpoint halves your risk when the gap is wide, and costs you every time price turns at the edge without going further. That is not a matter of opinion, it is a matter of counting on your own history.

The only honest way to settle it is to mark both levels in advance across thirty setups, note which one would have been touched, and compare. Thirty cases prove very little, and that is already more than most people did before choosing.

⚠️ This kind of trade-off is measured trade by trade, not by feel. It is exactly what a trading journal that records the setup as well as the result is for.

FVG examples: three worked cases on the ES contract

Here are FVG examples you can redo yourself on the ES contract, where the tick size is 0.25 point and the tick value 12.50 dollars. The numbers are there to show the arithmetic, not to be copied onto your chart this morning.

First case, a BISI. The first candle has a high at 5,480.25, the third a low at 5,486.00. The hole is 5.75 points, so 23 ticks, so 287.50 dollars per contract if your stop runs from one edge to the other. Its midpoint falls at 5,483.125, which is not a tradable price on a 0.25 tick: you round to 5,483.00 or to 5,483.25, and you write down once and for all which of the two, or two records of the same gap stop being comparable.

Second case, a SIBI. The first candle has a low at 5,512.00, the third a high at 5,506.50. The hole is 5.5 points and stays above price. Same geometry, opposite direction, and nothing else changes in the measurement.

Third case, the inversion. Price comes back down and a candle closes below 5,480.25, so below the lower edge of the BISI from the first case. The 5,480.25 to 5,486.00 zone no longer reads as demand on the return, it is watched as supply. The drawing stayed identical to the one in the first case, the event that flipped its reading is a close, not a wick.

On stocks, the same measurement does not give the same object. The dominant hole there is the opening one, between yesterday's close and today's open, where nothing was traded for hours. A fair value gap read on a daily stock chart is often that overnight hole dressed up as three candles, and the rules of return then have nothing to do with those of a contract that quotes almost around the clock.

What you do with the return into the hole once it happens, the entry itself and what triggers it, belongs to the lesson on ICT entry models. This one stops at the hole and its measurement.

An IFVG example, traded: entry, stop and target

Take the third case one step further. The gap ran from 5,480.25 to 5,486.00, a candle has closed below 5,480.25, and the zone is now watched as supply. Price then climbs back into it from below. That return is the trade, and it has three prices to fix before it happens.

The stop goes beyond the far edge, because the reading is only cancelled by a close back above the zone: one tick over 5,486.00 gives 5,486.25. Selling the lower edge at 5,480.25 puts the stop 6 points away, so 24 ticks, so 300 dollars per contract. Waiting for the midpoint, rounded to 5,483.00, cuts that to 3.25 points, 13 ticks, 162.50 dollars, and leaves you without a fill every time price turns at the edge.

The target is not given by the gap. It is the next place where counterparty is known to sit, here the low the fall was heading for. Put that low at 5,468.25: from the edge the trade aims at 12 points for 6 risked, two to one; from the midpoint, 14.75 points for 3.25, about four and a half to one.

Neither ratio means anything alone. Each one is to be set against how often the zone held on your own record at that entry, edge or midpoint, and the two entries do not share the same count: the midpoint is reached less often, and that missing share is the price of its better ratio.

Does the first presented FVG need a daily bias?

The first presented FVG is the first fair value gap formed after the start of the day. The parent is the same three-candle object, only the rank changes: the first of the day, not another one.

The word first only means something once the starting hour is fixed. The ICT convention opens the day at midnight, New York time. Take it or take another one, but never move it: two people starting from two different hours find two different first presented FVGs on the same day, on the same chart, and their records stop comparing.

Does that first gap require a daily bias? The doctrine that teaches it answers clearly: yes. Being first says nothing about direction. With no direction decided beforehand, a day presents a first bullish gap and a first bearish gap, and you would take both.

So it is a filter, not a signal. It cuts the day's list of gaps down to one, once the direction has been decided elsewhere. That happens in the lesson on daily bias, and some also look for it in the SMT divergence between two correlated instruments, which has a lesson of its own. Nothing inside the gap supplies that direction.

⚠️ The cost of that filter is measurable and almost never measured: it makes you give up every other gap of the day. Before sticking to it, count on your own record how often the first gap worked better than the second. If the difference sits inside the noise, you have added a constraint with nothing in return.

Four mistakes that turn the zone into noise

Reading the close on the wrong timeframe. A gap drawn on the 15-minute chart is inverted by a 15-minute close. On the 1-minute chart, price can close beyond that same edge and come back before the 15-minute candle is done, and you will have logged an inversion the gap's own timeframe never confirmed.

Putting the stop inside the zone. It makes the risk look small, and it tests something else: whether price trades back inside a zone it has just crossed, which says nothing about the reading being wrong. The reading is cancelled by a close through the far side, so that is where the stop belongs. If that distance is too large for your account, the answer is a smaller size, not a closer stop.

Counting gaps too small to matter. A two-tick hole on a 1-minute chart is a fair value gap by definition, and ordinary noise is enough to close through it. Set a minimum width before you start, in ticks or as a share of the recent average candle range, and log nothing below it: otherwise the sample fills up with events that measure the spread.

Taking the zone with nowhere to go. An IFVG says where price was rejected, not where it is heading. If the nearest opposing level is closer than your stop, the trade cannot pay for its own risk, however clean the inversion looked, and your record should show those cases as skipped, not as losses.

What an IFVG strategy has that is fragile

An IFVG strategy built on the above has one specific weakness, and it deserves to be named rather than worked around: every gap is eventually traded through, so every FVG sooner or later becomes an inverse fair value gap.

If you look at charts after the fact, you will only see the ones that worked. The others look like nothing in particular, so your eye does not keep them. That is the most expensive bias in this whole family of methods, and no quantity of video fixes it.

The honest test is to mark inverse fair value gaps live, before you know what follows, and count. Fifty occurrences are enough to tell you whether your reading is worth anything on your instrument and your timeframe.

What that count gives you is not a win rate. It is a ratio between what the cases that work pay and what the others cost, and it is the only measurement that decides whether the method holds.

Do the exercise across fifty cases and you will know more than any course will tell you, because you will have measured it on what you actually trade, not on somebody else's screenshots.

Practising: fifty gaps marked blind

Pick one instrument and one timeframe, one of each. Go back three months and move forward candle by candle, without looking at what comes next.

Every time you see a hole across three candles, mark it and note three things: the date, the direction, and the width in points. Nothing else. Do not judge, do not anticipate.

Keep going until fifty. Then, for each one, note what happened afterwards: filled and held, filled and cut through, never revisited. You end up with three columns and a count.

What you find surprises almost everyone: the proportion of gaps never revisited is larger than the ambient talk suggests, and the spectacular cases are rare. That is good news, because a sort across fifty cases costs one evening and replaces six months of intuition.

Keep that record. When you want to compare one method with another, you will have a baseline to compare it against, which is exactly what most of the people who quit after three months never had.

Frequently asked questions