A fair value gap, or FVG, is a price range the market crossed so fast that it barely traded there. You spot it on three candles: when the low of the third stays above the high of the first, the hole between the two is a fair value gap.
The hole three candles leave behind
Contents
The fair value gap comes from ICT vocabulary, where it is also called an imbalance. The idea fits in one sentence: when a move leaves very fast, it leaves behind a band of prices where buyers and sellers barely met. The market went through it, it did not really trade there.
Its usual reading follows from that. A price range crossed without transactions is treated as unfinished business, and many methods conclude that price tends to come back to it before moving on. That is an observed tendency, not a law: nothing forces a market to come back and fill anything, and this page returns to that below.
An FVG can be read on every timeframe and on every continuously quoted market, which is both its convenience and its main flaw: there are a lot of them, and most mean nothing. Our lesson on the fair value gap and the inverse FVG covers the subject in full; this page sticks to the definition and to what you can honestly expect from it.
Bullish FVG: low of the 3rd candle > high of the 1st — Size = low of the 3rd − high of the 1st
The bearish version is symmetrical: there is an FVG when the high of the 3rd candle stays below the low of the 1st, and the size is then low of the 1st − high of the 3rd.
The second candle does not enter the calculation, and yet it does all the work: it is the impulsive move, the displacement, that pushes the first candle away from the third.
The calculation uses the wicks, not the bodies. Many free indicators compare candle bodies: they then mark areas that are not fair value gaps, and miss real ones.
An FVG depends on the data feed as much as on the market. Two brokers whose wicks differ by a point do not draw the same gaps, which makes a list of levels impossible to carry from one platform to another.
| State of the gap | What it means |
|---|---|
| Never touched | The level is intact. This is the state where it works as a possible target for a coming move, not as an entry. |
| Touched at its edge | Price came back into contact without entering. The reaction is at its cleanest, and this is also the rarest case. |
| Half filled | Price reached the middle of the gap, the consequent encroachment. Many methods treat the return as complete at that level. |
| Fully filled | The hole no longer exists. The level loses its value as a marker and has no business staying on the chart. |
| Traded through, then rejected | The role flips: a bullish gap broken downwards becomes a resistance area. That is an inverse FVG, and it has its own reading. |
The table above is the only reading that holds: a fair value gap is read by its state, never by its mere existence. An intact gap and an already filled gap look the same on the chart and say entirely different things, and that is exactly what automatic indicators make hard to see, since they often keep displaying both.
The second filter is context, and it matters more than the gap itself. An imbalance left by a move that breaks market structure has nothing in common with an imbalance left in the middle of a quiet range, even though both meet the same three-candle definition. The first marks flow going through, the second marks a candle slightly bigger than its neighbours.
The fair value gap runs from 1.0960 to 1.0980. Its midpoint, 1.0970, has a name in ICT vocabulary: the consequent encroachment. That 20-pip band, and nothing else, is the marker: a move of price back inside it is what is called a retracement into an FVG.
This is the most widespread use. Rather than entering at the very start of a return, where nothing guarantees it stops, you wait for price to come back into the fair value gap left by the initial move. The gap gives two precise edges and a midpoint, so a place to put an entry and a place to put a stop, which an area drawn by eye never does.
The real benefit is not prediction, it is precision. A numeric level makes risk calculable before entering, and that is the condition for the trade to be comparable with the others afterwards. A vague marker produces trades you cannot measure.
The reverse use is less practised and often more useful. A fair value gap left higher up the chart is a place price has already crossed quickly, so a plausible objective for the rest of a move, without having to guess where to exit.
It changes the nature of the decision: looking for an entry inside a gap requires believing it will hold, aiming at a gap as a target requires nothing of the sort. It is also the reading that lends itself best to a measured check, since all it takes is comparing what the trade offered at its best moment with what was actually banked.
The first flaw is abundance. An indicator that marks every fair value gap on a chart finds dozens per session on a low timeframe, because the three-candle definition is purely geometric and requires no context. Something that appears everywhere cannot announce anything: what separates a useful gap from a noise gap is never in its shape, it is in what produced it.
The second is the fill statistic quoted everywhere. You regularly read that 70% or 80% of fair value gaps end up filled. That kind of figure is almost always measured with no time limit, and over an infinite horizon an oscillating market eventually passes back through nearly everything. Without a time window fixed in advance, the statistic says nothing usable, and it compares to nothing.
The third is the most expensive, and the only one a chart will never fix: nobody measures their own fair value gap trades. The useful question is not whether FVGs work in general, it is whether yours work, on your market, at your timeframe, with your way of exiting. That is journal work, not indicator work, and it is settled by tagging those trades and then reading their win rate and their profit factor separately from the rest.
The classic mistake. Treating every fair value gap as an entry signal. The definition being purely geometric, a chart contains dozens of them per session, and taking them all amounts to entering at random with the conviction of following a method. The gap says where to look, it never says when to enter.
| Indicator | Formula | What it measures | When to use it |
|---|---|---|---|
| Fair value gap | 3 candles, the 1st and the 3rd do not overlap | A price range that barely traded | Locating a return, or aiming at a fill |
| Order block | The last opposing candle before the displacement | The level where a large position was built | Looking for a reaction when price returns |
| OTE zone | The 0.62 to 0.79 retracement of the move | A pullback area measured on the whole move | Placing an entry inside a pullback |
| Liquidity void | A range crossed in one block, often over several candles | A wider absence of quoting than an FVG | Aiming at a distant objective |
A fair value gap left by a move that breaks structure does not carry the same value as a gap left in the middle of a range. Keeping only the former divides the list by ten and makes it readable. It is the one filter that clearly changes the signal-to-noise ratio, and it requires no tool.
Decide in advance how many candles a return has to happen within, say twenty. A gap not filled inside that window is a gap that was not filled, full stop. Without that rule written beforehand, you always notice afterwards that price did come back, and you are no longer measuring anything at all.
It is the only way to answer the one question that concerns you: do your fair value gap trades win. One tag on every trade taken that way, thirty or so occurrences, and the comparison with the rest of your activity happens by itself. It is also the reading that survives a change of market, whereas an indicator setting does not.
It is a price range the market crossed so fast that it barely traded there. You spot it on three candles: if there is a space between the high of the first and the low of the third, that space is a fair value gap, also called an imbalance.
FVG stands for fair value gap. The term comes from ICT vocabulary. The word imbalance describes the same object, and the two are used interchangeably.
Most FVG examples look the same: three candles where the middle one is clearly larger than its neighbours, and where the wicks of the first and third never cross. The diagram at the top of this page shows a worked case, a 20-pip bullish gap between 1.0960 and 1.0980. On a bearish chart the figure is the same, flipped.
A retracement into an FVG is the moment price comes back inside the band left by the imbalance, after moving away from it. Many methods treat that return as complete at the midpoint of the gap, a level called consequent encroachment in ICT vocabulary. Nothing guarantees that the retracement happens, nor how long it would take.
A fair value gap is the imbalance itself. An inverse fair value gap is that same imbalance once price has traded through it and then rejected it: its role flips, and a bullish gap broken downwards then reads as a resistance area. It is the same role swap as broken support turning into resistance. Our lesson on the inverse fair value gap covers that case in detail.
Tradoshi does not draw fair value gaps on a chart, and that is not its job: it answers the next question, whether the trades you take that way actually win. Put a tag on every trade that came from a fair value gap, and the Performance by tag block in the Journal screen gives you, for that tag alone, the trade count, the win rate, the average R and the net P&L. Clicking the row filters the journal down to those trades.
Two readings complete that one. The Reports screen shows MFE and MAE, that is how far the trade went in your favour and against you before closing: that is what answers whether you exit a retracement into an FVG too early. And backtesting replays the market candle by candle, which lets you test the rule on history before committing capital, with the same statistics at the end.
To go deeper on this: read the full guide. Every term is gathered in the glossary.