A fair value gap is a three-candle pattern where the middle candle moves so quickly that the wicks of the candles either side never overlap. That untraded space is an imbalance — prices at which buyers and sellers never actually met.
What an imbalance actually is
In a bullish FVG, the low of the third candle sits above the high of the first. Price simply skipped that band. The market moved so aggressively in one direction that orderly two-way trade did not happen.
The trading premise is that markets tend to revisit those areas — partly because unfilled orders remain there, and partly because participants who missed a fast move wait for a better price and their orders sit inside the gap.
Which gaps are worth trading
| Factor | Weak gap | Strong gap |
|---|---|---|
| Size | A few pips, within normal noise | Meaningful relative to ATR |
| Displacement | Drifting move | Sharp, large-bodied candle |
| Context | Mid-range, no structure | Created by a break of structure |
| Confluence | Standing alone | Overlaps a prior level or order block |
| Freshness | Already tapped twice | Untouched since it formed |
The single most useful filter is displacement. A gap left by a violent, decisive move means something; a gap left by ordinary drift is a rounding artefact.
Market conditions required
- A clear directional move that created the gap — ideally one that broke structure.
- The gap is untouched since formation.
- Higher-timeframe bias aligned with the direction you intend to trade.
- Enough distance to a sensible target to make the reward worthwhile.
Entry rules
- Mark the gap: the high of candle one to the low of candle three (bullish), or the reverse.
- Wait for price to retrace into the gap. The 50% level of the gap is a common resting point for a limit order.
- Either enter on a limit at that midpoint, or wait for a rejection candle inside the gap and enter on its close.
- If price closes fully through the far side of the gap, the setup is void.
Limit entries fill more often and give better prices; confirmation entries avoid the gaps that fail outright. Beginners are better served by confirmation, at the cost of some missed trades and slightly worse prices.
Stop-loss rules
Beyond the far edge of the gap, plus a buffer. A gap that fully fills and closes through has been rejected as an area of interest — the imbalance is resolved and there is nothing left to trade.
Take-profit rules
- T1: the extreme of the move that created the gap. Take partial there.
- T2: the next liquidity pool or structural level beyond it.
- Minimum: if T1 is under 1.5× the stop distance, the gap is too large relative to the move and the trade is not worth taking.
Risk management
- 1% maximum, sized from the full gap depth rather than your entry point — assume you may be filled at the worst end.
- Do not trade several overlapping gaps on the same instrument as separate positions; they are one idea.
- A gap tapped and rejected once is materially weaker on the second visit.
Example winning trade
US 100, 15-minute. A sharp move up from 19,180 to 19,340 breaks structure and leaves a gap between 19,232 and 19,258. The 4-hour trend is up.
Price retraces the following session into the gap, and a 15-minute candle wicks to 19,238 before closing at 19,262. Long 19,262, stop 19,225 below the gap, 37 points of risk. T1 at the move extreme 19,340 reached for +2.1R.
Example losing trade
EUR/USD, 5-minute. A bullish gap at 1.0844–1.0851 formed on a modest move with no structure break. Limit long filled at 1.0847, stop 1.0839.
Price slices straight through the gap without pausing and hits the stop within twenty minutes, continuing to 1.0810.
The gap failed the displacement test — it was created by a drift, not a decisive move, and there was no structural reason for buyers to be waiting there.
When NOT to trade it
- The gap came from a slow, low-conviction move.
- It has already been tested and rejected.
- The gap is enormous relative to the move, making the stop uneconomic.
- Trading it would be counter to a strong higher-timeframe trend.
- The gap formed on a news spike, which is noise rather than positioning.
Common mistakes
- 1Trading every gap on the chart
Fast markets leave gaps constantly. Most are meaningless. The filters exist to discard them.
- 2Ignoring displacement
This is the difference between a gap that means something and a gap that is just a picture.
- 3Placing the stop inside the gap
The gap must be allowed to fill. A stop inside it will be hit by the very move you are trading.
- 4Using low-timeframe gaps as if they were significant
One-minute gaps are largely spread artefacts.
Backtesting considerations
- Encode a minimum gap size as a fraction of ATR, otherwise every rounding artefact counts as a setup.
- Encode displacement too — for example, the middle candle's body exceeding 1.5× the average body of the previous ten.
- Measure the fill rate of gaps separately from the win rate of trades taken. They answer different questions.