The inverted FVG
The single most valuable pattern extension beyond the standard retest. An FVG inverts when price closes through it decisively in the opposite direction, a bullish gap gets a close-through to the downside, or vice versa. The zone that was acting as support now flips to resistance, or the reverse.
Why it matters: an inversion is the market explicitly telling you that the gap lost its edge as support/resistance. Ignoring the inversion and continuing to trade the original direction is one of the more expensive mistakes FVG traders make. The setup isn't “wait for the retest” anymore, it's “short the retest into the now-resistance zone.”
The Inverse FVG indicator detects the close-through and flips the zone color on the candle it happens, the instant the role reverses, not a confirmation candle later. If you trade a lot of inversion setups, watching for the color flip is orders of magnitude easier than tracking it manually.
ES prints a bullish gap from 5,204 to 5,209 on a morning push. You're watching it as support for a long. Instead of holding, price grinds back down, and a five-minute bar closes decisively at 5,201, a full close below the gap low. The gap is inverted. Your long thesis is dead; that same 5,204 to 5,209 pocket is now overhead supply.
The trade flips with it. Price rallies back to 5,205, tags into the inverted zone, and prints a rejection candle that closes at 5,202. That's the short entry, stop above the gap high at 5,210, targeting the session low. What was a failed long becomes a clean short, read from the other side. The trap is refusing to let go of the original bullish idea and averaging into a level the market just told you had flipped.
Regime-dependent adjustments
The same FVG setup behaves differently in different market regimes. Three that recur:
High volatility (NFP, CPI, FOMC sessions)
Gaps print bigger and retests are deeper, often right through the midline and touching the far edge before the reversal. Widen your stop to beyond the far edge with some margin, and accept that fewer setups will trigger cleanly. In the highest-volatility sessions, sit the first hour out entirely.
Low volatility (summer afternoons, holiday sessions)
Gaps are smaller, cleaner, and retest on the near edge. Tighten targets to the next visible level rather than holding for extended runs. The edge is real but the range is limited; respect it.
Trending sessions
Price doesn't fully retest, the pullback gets to the midline and resumes. Setup 02 (breakout pullback) dominates Setup 01 on trending days. A full-retest entry ends up being an entry after the move has already resumed, which is worse than no entry.
Edge cases worth knowing
Three patterns that show up often enough to earn a name, and are easy to miss without it.
Stacked FVGs
A trending session frequently prints three or four gaps on the way up. The deepest (oldest) gap is usually the one that holds on the retrace; the newer ones fill on the way down. Use stacked gaps as a sequence, not as independent setups.
Gap-inside-gap
Occasionally a small gap forms inside a larger one during the retest itself, a mini-impulse within the primary FVG's range. Rare but high-conviction: both gaps agreeing in direction at the same price almost always holds.
The disappearing gap
Sometimes price enters the gap, passes the midline, passes the far edge, and doesn't reverse. The gap is “filled” but the expected reaction never prints. When this happens, the gap was probably structurally weak, no impulse behind it, or no structural context, and you either avoided it using the validation steps from chapter two, or you paid for the lesson.
Time of day changes the gap
A gap is not the same trade at 09:35 as it is at 13:00. ICT organizes the session into windows, the killzones, and the same geometry behaves differently depending on which one it prints in. You don't need the full ICT time theory to use the core idea: liquidity and participation swing hard across the session, and a gap's reliability swings with them.
The richest window. Real volume, real displacement, and gaps that form here tend to get respected because there are enough participants to defend them. Most of your best setups will print in the first ninety minutes.
The dead zone. Volume thins, gaps get sliced on low participation, and the clean retest you're waiting for often just doesn't come. Gaps that print here are lower-conviction by default, size down or sit out.
Participation returns as the close approaches. Gaps regain reliability, and a level that formed in the morning and never filled often gets its retest here as positioning squares up into the bell.
The practical rule is a discount on the clock. A gap that would be a full-size trade at 10:00 is a half-size trade, or a skip, at noon, even with identical geometry and context. When you review a losing streak, check the timestamps before you blame the pattern, a cluster of lunchtime entries is a scheduling problem, not a strategy problem.
None of these extensions is a standalone signal, they're weights on the same scale. A fresh gap that's nested inside a higher-timeframe zone, prints in the opening hour, and sits at an order block is the trade you size up on. The same gap at noon with a prior touch and no confluence is the one you pass. Advanced FVG trading isn't a new pattern; it's reading how many of these factors line up before you commit.
Conceptual takeaways
- Inversion is the single most valuable FVG extension, watch for the close-through and trade the flipped zone.
- Setups behave differently by regime, high vol needs wider stops, trending sessions favor breakout-pullback.
- Stacked gaps, gap-inside-gap, and the disappearing gap are the named patterns worth recognizing.
- Time of day changes the trade: open and afternoon gaps get respected, lunchtime gaps get sliced. Apply a discount on the clock.
- A failed long is a short once a gap inverts, trade the flipped zone instead of averaging into the old idea.
That's the end of the series. You have the pattern, the identification process, the trading setups, the psychological framework, and the advanced vocabulary. The rest is reps.