Measuring the non-overlap on gold and EURUSD
Bullish FVG height = Candle1.high - Candle3.low, valid only when Candle3.low > Candle1.high (no overlap). Example: C1 high 2,341.20, C2 displacement, C3 low 2,344.80. Gap = 3.60 points from 2,341.20 to 2,344.80. Equilibrium is 50% at 2,343.00. Bearish: valid when C3.high < C1.low. EURUSD H1: C1 low 1.08460, C3 high 1.08390, gap 7.0 pips. If C3 high is 1.08465, overlap is 0.5 pip and the FVG is rejected.
Minimum gap filter: 0.15x ATR(14) of that timeframe, so a 1.2-pip H1 EURUSD "gap" inside a 12-pip ATR does not clutter the book. Gold H1 ATR 18: minimum about 2.7 points.
Fill rules: 50% equilibrium versus full fill
Many books fade the first touch of 50%. Venasri scores inversion after a full fill through the opposite gap edge plus a CHoCH. Partial fills that only nick 20% of the gap remain valid. If NFP trades through the entire 3.60-point gold gap and closes beyond, the FVG is filled and deleted. Holding a limit in a filled gap is hoping, not executing.
ICT-style "inversion FVG" after a failed bullish gap becomes a short zone. The engine can invert rather than delete when the displacement BOS that created it is later violated on the same TF.
Stacking FVG with Order Block and Volume Profile
Related reading: Order Block guide and LVN guide.
Stops outside the gap, not inside the paint
SL belongs beyond the origin OB distal or beyond the FVG invalidation edge, not in the middle of the rectangle. A 3.60-point gold gap with a 16-point structural SL still sizes as $500 / (16 x 100) = 0.31 lots. Venasri Risk Management & Equity Guard plus 24/7 MT5 Bot Control attach that SL if you market into the gap from iOS. start free on Venasri before the next CPI so a 30-point news hole cannot land you naked in the imbalance.
start free on Venasri so fair value gap fvg imbalance is not a browser-only exercise: Venasri Risk Management & Equity Guard owns the flatten, Venasri Performance Analytics stores MAE and R, and 24/7 MT5 Bot Control is the 24 hour execution path on ZeroMQ or MetaAPI.
