Fair Value Gaps: An Introduction

Fill is a tendency, not a schedule

Markets return to these bands often enough that traders use them as pullback targets. The SMC framing says price rebalances inefficiency. A plainer reading is that fast moves overshoot and retrace, and the gap simply marks where the fast move happened. Both readings predict the same behaviour, which is a useful reminder that the explanation is not carrying the weight.

What neither reading gives you is timing. A gap can fill in the next candle, in three weeks, or after the trend that created it has already reversed twice. In a strong trend, price stacks fresh gaps on top of unfilled ones and never looks back. Treating an unfilled gap as an obligation is the standard way this concept turns into a series of counter-trend losses.

Timeframe compounds the problem. The same USD/JPY move that leaves a clean 13-pip gap on the 15-minute chart may leave nothing at all on the hourly, because the hourly candle that contains the impulse also contains the retracement and the wicks overlap. Neither chart is wrong. They are measuring different windows, and a gap that exists only on the 1-minute chart is a gap that will be filled by ordinary spread noise, which makes it useless as a location.

A rough staleness rule helps more than it should. If a gap has been sitting unfilled for longer than the swing that created it took to form, the context around it has usually changed enough that it is no longer describing the same market. Some people put a hard limit on it, such as ignoring intraday gaps older than the current session. The exact rule matters less than having one, because without it every old gap stays on the chart forever and the chart stops being readable.

Gaps as targets, and where that argument breaks

The reverse use is to treat an unfilled gap ahead of price as a take-profit reference: if the market is likely to travel there, that is a sensible place to be out. It works when the gap is the nearest meaningful feature. It breaks when there are four unfilled gaps stacked in the direction of travel, because then you are choosing a target by preference rather than by evidence.

A simple discipline fixes most of it. Mark only the gap produced by the impulse that broke structure, ignore the rest, and delete a gap from the chart once price has traded fully through it. Five gaps on a screen is analysis. Twenty is decoration.

Stacking a gap with something that is not a gap

An FVG sitting inside an order block, or inside a discount zone during an uptrend, is treated as a better location than either alone, and that is reasonable as long as the reasons are genuinely independent. An FVG, an imbalance and a one-candle inefficiency are three names for the same observation. Stacking them is one reason wearing three hats, and it produces false confidence rather than a better location.

As with everything in this category, define it mechanically, log it, and let your own records decide its place. Position size from the invalidation distance using the position size calculator, and remember that a gap that does not fill costs you nothing if you never entered, while a gap you forced costs you the full stop.