A fair value gap (FVG) is a three-candle imbalance: when candle 2 moves so violently that the wicks of candles 1 and 3 don't overlap, the untraded space between them is the gap. In a bullish FVG, the zone between candle 1's high and candle 3's low was skipped by price in its haste — one-sided business the market often comes back to finish.

The concept comes from the ICT/SMC school, but the underlying idea is older than both: fast moves leave thin spots, and thin spots tend to get revisited.

Spotting a valid FVG

Mechanical definition (bullish case):

1. Three consecutive candles, with candle 2 strongly directional
2. Candle 1's high < candle 3's low — the wicks don't touch
3. The zone between those two prices is the FVG

The mirror (candle 1's low > candle 3's high) is a bearish FVG. Every liquid chart contains dozens — which is exactly the trap. The skill is not finding FVGs; it's filtering the meaningful ones.

Which gaps matter

• Born from displacement: the gap should come from a candle that means something — a breakout from structure, the move after a liquidity sweep, a structure shift. Gaps from random volatility are noise.
• With the trend: bullish FVGs in uptrends act as support; counter-trend gaps get steamrolled.
• Fresh: untested gaps are strongest; each revisit consumes resting interest.
• Higher timeframe: a 1H or 4H FVG outranks the dozens on the 1m chart.

The strategy rules

• Timeframe: 15m–1H
• Setup (long): uptrend; a strong displacement move (ideally following a sweep or breakout) leaves a fresh bullish FVG below current price
• Entry: limit order in the gap — at its midpoint (the "consequent encroachment" in ICT terms) as a good default
• Stop-loss: below the bottom of the gap; a fully filled-and-broken gap is invalid
• Targets: the high of the displacement move (often ~2R), then the next liquidity level
• Skip: counter-trend gaps, gaps already tested twice, and gap entries straight into a major resistance level

Why fills happen (no conspiracies needed)

The mundane mechanics: fast moves overshoot because one side temporarily vanished; market makers re-quote the skipped zone; late entrants who missed the move place bids where it started; and profit-taking naturally retraces into the imbalance. The result — price frequently returning to recent one-sided zones before continuing — is a statistical tendency you can measure yourself, which beats believing anyone's video about it.

How to backtest it

FVG trading is unusually easy to test because the definition is fully mechanical:

1. In chart replay on 15m candles, mark every FVG born from displacement as it forms — future hidden.
2. Log for each: did price return? Did the gap hold as support/resistance (your entry works) or slice through (stopped)?
3. Separate your sample by the filters above — with-trend vs. counter-trend, fresh vs. retested. 50 gaps takes a few evenings via manual backtesting.
4. Typical finding: fill rates are high everywhere (gaps do get revisited), but hold rates — the part you're paid for — concentrate almost entirely in with-trend, displacement-born, fresh gaps. That's your tradeable subset.

Common mistakes

1. Trading every gap — the chart is full of them; the filter is the strategy.
2. Counter-trend gap fades — a bearish FVG in a raging uptrend is a speed bump, not a ceiling.
3. Stops inside the gap — normal filling behavior stops you out of valid trades; invalidation is the far edge.
4. 1-minute gap safaris — everything on 1m is an imbalance; start at 15m.

FAQ

Do fair value gaps always get filled?

No — and "eventually" is not a trading plan. Many gaps fill quickly, some take weeks, some never fill. The tradeable question isn't whether it fills but whether the gap zone holds on the revisit, which is what the filters predict.

Are FVGs the same as gaps on stock charts?

Same family: stock gaps are between-session imbalances; FVGs are intra-session three-candle versions. Crypto trades 24/7 so classic overnight gaps barely exist — FVGs are the gap-trading of continuous markets.

What timeframe is best for FVG trading?

15m–1H balances frequency and signal quality. Daily FVGs are rare but powerful swing levels; 1m FVGs are mostly noise.