Fair Value Gap (SMC)
Also called: FVG · price imbalance · three-candle gap
A fair value gap is a three-candle pattern where the first candle's wick and the third candle's wick do not overlap, leaving a band crossed only by the middle candle. Smart Money Concepts calls that band an inefficiency and expects price to return to it. It is measured on price alone.
The construction
Three candles. In the bullish case the high of the first sits below the low of the third, leaving a band the neighbours never touched. Mirror it for the bearish case. Some definitions use bodies instead of wicks, which changes every gap on the chart.
The pattern is also a function of the timeframe you drew it on. Merge the three candles into one higher-timeframe bar and the gap disappears: the geometry that created it was a bar boundary, not a market event.
"Inefficiency" is a claim, not an observation
The band is not untraded. Price crossed it during the middle candle — quickly, thinly, but it traded. What the pattern marks is a range where little business was done relative to its neighbours: a measurable statement, not a metaphysical one.
The measurable version already exists. On a volume profile the same thin band appears as a low volume node — and unlike a three-candle rule, you can test whether it survives a change of row size. A node that vanishes when you widen the rows was an artefact. A wick-defined gap has no equivalent test.
What the flow adds
Two questions turn the drawing into something falsifiable. How much actually traded inside the band — hundreds of contracts, or thousands? And on the return, does aggression stall there, with absorption or a delta that turns, or does price cross as fast as it left? The second question is the one worth having: it is answered in real time, not in review.
A worked example
In a synthetic ES sequence, a bullish gap leaves a band from 5 302.75 to 5 306.25 — fourteen ticks. The session profile records 1 240 contracts across it, against a median near 9 800 for bands of the same width elsewhere in the range. Thin by a factor of eight, but not empty.
Redraw the same sequence on a fifteen-minute chart and the three candles collapse into one. No gap at all.
The trap
The fill statistic. The claim that most fair value gaps eventually get filled — the figure passed around is usually near 85% — has never been published with a dataset, a symbol, a period, or a maximum holding time. Without a time limit it means little: in an oscillating market, almost any band near price is touched eventually.
The honest version needs a control group: compare fill rates against randomly drawn bands of the same width and distance, on the same data. If the two match, the pattern added nothing. Nobody quoting the figure has published that comparison.
Frequently asked
- Is a fair value gap the same as an imbalance?
- No, and the shared vocabulary causes real confusion. A fair value gap is geometry on a price chart. A footprint imbalance compares executed volume at bid and ask on adjacent prices. One is drawn, the other counted, and they can point at different prices.
- Do fair value gaps work on any timeframe?
- They appear on every timeframe, which is the problem rather than the feature. One session produces a different set of gaps at one minute, five minutes and one hour. Fix the timeframe in advance, or you will find a gap wherever you need one.
- Why take FVGs seriously at all?
- Because a thin band is a real thing, badly named. Prices the market crossed quickly often see a reaction on the return, and volume-based tools measure that directly. The disagreement is about method, not about the phenomenon.