Polymarket Price Gaps, Explained

• Information gaps — a real repricing event in one tick (the strike ladder reprice moving rungs 0.044 to 0.996 in one second).
• Book gaps — a level vanishes, leaving a 99.6-cent-wide empty grid across the midpoint while the true price is unchanged.
• Resolution gaps — the settle jump to 0.99/0.01, terminal by design, unreadable as a trading signal.
• Opening gaps — thin pre-event books with wide level spacing around the open.

A measured BTC strike ladder repriced from 0.044 to 0.996 in a single second — that is information, and it happened on the whole ladder at once, which only full-book capture could show.

The same measurement recorded a 99.6-cent hole in the middle: the mid moved to 0.4995 and 0.500 not because the market repriced but because the adjacent levels were empty. An analyst charting mids sees a 50-cent move; a book reader sees the ladder intact with one empty rung.

• Always check the book at the gap timestamps: was there size at the levels around it?
• Compare the gap's levels to the pre-gap ladder to classify repriced vs missing.
• Treat near-resolve jumps (0.40 to 0.99) as settlement, not signal — recorded books show depth draining hours earlier.
• Measure gap frequency per market type to build filters; sports event books gap differently than five-minute crypto.

• From the books for a reprice window, find consecutive frames at 250ms with a price step larger than one tick; the gap list is your window's reprice signature.
• Record the biggest single-second gap you find; the maximum documented one-second move (0.044 to 0.996) is the extreme the ladder can produce.
• Check whether the gap coincided with a reference print: join the gap second to the futures level on the same UTC second (the same-second reset pattern).
• Build a small gap table (second, from, to, reference aligned?) for the window; the aligned versus unaligned shares are the gap taxonomy.

A price gap is a jump between captured frames larger than the tick — the compression of the whole ladder in a single second on the extreme end, and the mid floor moving several ticks in a normal repricing.

The measured gaps divide cleanly into reference-aligned (the 14:42:50 second where the futures print and the floor met) and not-aligned (the residual drift), and the archive's exact timestamps are what makes the division possible.

When a gap aligns with a reference print on the same UTC second, it is discovery; when it does not, it is drift or illusion from a stale mid. Period and window boundaries belong in every gap report, because a gap without its boundary is un-reproducible by definition.

Gap studies also need a companion table of the frames before and after the jump: the last executable prices on both sides, the reference, and the second boundaries. With those four columns any reader can confirm the gap was real, aligned, and bounded.

A gap is real only if it is between executable prices; mid gaps on one-sided books are summaries, not trades.

Classify every gap by its reference alignment in the same second; aligned gaps are information, unaligned ones are drift.

The one-second 99.6-cent pass is the documented extreme, and the archive replays it frame by frame.

Gaps are the reprice mechanism measured; this page pairs with the repricing-data page (the recorded cases) and the price-level page.

It gives the arb and order-flow pages their vocabulary: aligned versus non-aligned gaps.

The one-second 99.6-cent pass is its canonical example and it appears in the price-bounds and price-level explainers too.

A gap is real only between executable prices, and it is either reference-aligned or drift.

Classify by the second, and bookmark the 99.6-cent one-second pass as the max the ladder can produce.

• Order book data
• Price history
• Crypto data
• Provider matrix