A gap in one direction, a stretch of trading, then a second gap the same way — the price action in between left stranded like an island, cut off from the trend on both sides.
Early chartists noticed clusters of bars completely cut off by gaps on both sides, left floating apart from the surrounding trend like an island in open water.
The landmark text treats this as a real but relatively rare signal of exhaustion — trend momentum failing so sharply it leaves visible gaps on both sides.
Because it depends on real gaps, this pattern shows up more often in markets and instruments prone to gapping — earnings-driven stocks, futures reopens, and news-sensitive assets.
In markets that trade nearly continuously, like crypto, true price gaps are less common — this pattern is more at home on stocks and futures with real session breaks.
Price gaps in one direction, trades in a tight cluster for a few sessions, then gaps back the other way through roughly the same price zone — stranding the cluster as an island.
The island cluster is typically brief — a handful of sessions, not weeks. A longer stranded stretch starts to resemble a different pattern entirely, like a rounding formation.
Most gaps are common gaps that fill quickly, or breakaway/runaway gaps that don't reverse. An island requires two genuine gaps in the same direction, both crossing roughly the same price zone.
Around the sharpest days of the crash, a brief, volatile cluster of sessions was cut off by gaps on both sides, marking a sharp, if temporary, turn in the tape.
On futures markets that do close briefly, a short top cluster was cut off by gaps around a local high — a reminder this pattern needs a market capable of gapping at all.
A stock gaps up sharply, trades in a tight three-day cluster, then gaps back down through that same price zone. What is this?
Price gaps up, drifts sideways for six weeks, then gaps down. Is this a genuine island reversal?
A stock gaps down, then gaps right back up through the same zone a day later — no real cluster of trading in between. Is this an island reversal?
A gap, a stranded cluster, and a second gap — watched tick by tick on the left, and the mark it leaves in the ledger on the right. A confirmed island top, a mirrored island bottom — and a plain gap-and-fill that never stranded anything.
A gap appears. Judge whether a genuine second, opposing gap strands a real cluster — then call it: a true island, or just a plain gap-and-fill.
The classic error is trying to predict the island before it's confirmed, or mislabeling ordinary gaps as islands. The discipline is mechanical: require two genuine, opposing gaps around a real, short cluster, then react to the second gap as the tradeable signal, not the cluster itself.
Two gaps, one direction, and a short, genuine cluster left stranded between them — rare, sharp, and usually only confirmed once the second gap has already happened.
No man is an island.