The four-step identification process
Works on futures, equities, FX, and crypto, the pattern is pure price geometry, not instrument-specific.
Scan for rapid price moves
Look for two or three candles where price moved hard in one direction, a visible impulse, not a slow grind. These are the candidates. Slow, rotational moves rarely leave tradeable gaps.
Check the three-candle geometry
Candle 1 (pre-move), Candle 2 (the thrust), Candle 3 (post-move). For a bullish FVG, low(C3) > high(C1). For a bearish FVG, high(C3) < low(C1). The middle candle's body is almost always the largest of the three.
Mark the gap range
Draw the zone, from Candle 1's high to Candle 3's low (bullish), or Candle 1's low to Candle 3's high (bearish). That untested range is your FVG. The free Fair Value Gap indicator does this automatically; doing it by hand a few dozen times first builds pattern recognition.
Validate with context
A gap that forms at a structural level, yesterday's high, the session VWAP, a pivot, is worth more than a gap in thin-air midday. Context is the difference between a high-probability setup and a 50/50 pattern.
Five-minute ES or NQ is the training wheels. Enough gaps print per session to get reps; the timeframe is high enough to avoid pure noise. Spend a week circling gaps by hand on replay before deciding the indicator does it correctly.
Step 1. The 10:00 five-minute bar on NQ drops thirty-eight points on a data headline, three times the size of the bars around it. That relative range flags it as a candidate; you don't even need the geometry yet to know it's worth checking.
Step 2. Candle 1 (09:55) bottomed at 18,410. Candle 2 is the thrust down. Candle 3 (10:05) tops out at 18,402. Because high(C3) 18,402 < low(C1) 18,410, the geometry confirms a bearish FVG.
Step 3. Mark the range: 18,402 to 18,410, an eight-point pocket, with the midline at 18,406. That's the zone price has to climb back into to offer a retest.
Step 4. Context check: 18,410 was also the overnight low. The gap forms right on a level the whole session was watching, so it grades up from a fair setup to a good one. A gap in blank space at, say, 18,340 would have failed this step and gone on the skip list.
Grading a gap once you've spotted it
Identification tells you a gap exists. Grading tells you whether it's worth an order. Two gaps with identical geometry can be miles apart in quality, and three variables do most of the sorting.
The thrust candle should have a large body and a small wick, decisive, not indecisive. A wide bar that closes near its extreme reflects real one-sided pressure. A wide bar with long wicks on both ends is volatility, not displacement, and its gap is weaker.
Measure the gap against the average bar range for that timeframe and session. A gap a fraction of one bar's ATR is noise; a gap that's a meaningful multiple of it came from genuine imbalance. Bigger isn't automatically better, but too small is automatically ignorable.
An untouched gap carries the most weight. Each time price pokes in and reacts, the imbalance gets partly consumed. By the second or third tag there's little left to defend, so prioritise the first clean retest and downgrade a gap that's already been visited.
A quick shorthand: a top-grade gap has a decisive thrust behind it, a size that stands out from the session's noise, and no prior touches, sitting on a level the market already respects. Miss on one of those and it's still tradeable with smaller size; miss on two and it usually belongs on the skip list from the previous section.
Nesting: a gap inside a bigger gap
The single fastest way to upgrade a gap is to check the timeframe above the one you trade. Gaps nest. A 5-minute FVG that happens to sit inside an unfilled 1-hour FVG is a far better bet than the same 5-minute gap floating in open space, because the higher-timeframe imbalance is a level the whole market is drawn toward, and your entry timeframe just gives you a tighter place to act on it.
The workflow is two passes. First, mark the unfilled gaps on the timeframe one or two steps up from your execution chart, on a 5-minute trader that's the 15-minute and the 1-hour. Those are your zones of interest. Then drop to the execution timeframe and only take the gaps that print inside or right at the edge of one of those higher zones. You're using the big gap to say where, and the small gap to say when.
The 1-hour chart left a bullish FVG from 5,150 to 5,162 on yesterday's afternoon rally, still unfilled. That's your zone of interest, twelve points wide, way too coarse to enter against directly.
Today price rotates down into it. On the 5-minute, a fresh bullish gap prints at 5,156 to 5,159, right in the middle of the hourly zone. Now you have a precise three-point pocket to work, backed by a twelve-point level the whole market is watching. Same 5-minute pattern you'd normally grade as ordinary, promoted to a high-conviction setup purely by where it sits.
The three false positives
Most spurious FVG calls fall into one of three buckets:
Chop mistaken for imbalance
A tiny gap in a sideways session isn't a fair-value gap, it's rotation. If the move behind the gap wasn't a real impulse (large relative range, volume behind it), the gap is structurally weak. Skip it.
Timeframe mismatch
A 30-second gap on a 1-minute chart is noise by the time the 5-minute bar closes. If you're trading 5-minute setups, work from 5-minute FVGs, or higher. The gap has to be meaningful on the timeframe you're actually trading.
No structural context
A clean FVG in the middle of nowhere, away from any level anyone watches, is often a wait-out. The retest happens, the gap fills halfway, and price wanders somewhere else. Gaps that confluence with other references do the work.
Conceptual takeaways
- Four steps: scan for impulse, check the geometry, mark the range, validate with context.
- A gap without structural context is a 50/50 pattern. The level matters as much as the shape.
- Practice on replay. Hand-circle fifty gaps before trusting the detection to an indicator.
- Grade after you spot: displacement quality, size versus ATR, and freshness sort tradeable gaps from ignorable ones.
- A fresh, untouched gap carries the most weight. Each prior tag consumes the imbalance, so favour the first clean retest.
Next chapter: how to actually trade a gap once it's identified, entries, stops, targets, and the single most common mistake that eats the edge.