Polymarket Trading Volume, Explained

• Volume counts trades that actually filled, by contract count (shares).
• The official CLOB executes orders; fee and rebate mechanics are separate from data access.
• A recorded row's cumulative traded volume is the running total for that contract at that timestamp.

Executions and settlement are written to Polygon — someone, from Dune's Spellbook tables to exchange explorers, can rebuild the full execution history. That is why fill-based analytics are the most reproducible form of Polymarket history.

The order book that produced those fills is off-chain and gone unless recorded. Volume tells you flow happened; depth tells you at what price it could have.

A measured illustration: in the August 23 repricing event, five crypto markets hit lower bounds in 26 seconds while BTC took another 37 minutes. The volume story and the depth story diverge exactly where a bare volume chart stops asking questions.

Every row in this archive carries cumulative traded volume alongside bid, ask, and mid, so volume can be charted at the same 250ms cadence as depth. That pairing — what traded, next to what was quoteable — is the combination most single-endpoint feeds lack.

For research-grade work, cross-check the fills on-chain (executions) against the recorded quotes (the book): the two should agree, and the places they diverge are exactly the events worth studying.

• Pull one day of frame volume for a market and compute the total; then split the day into five-minute buckets and find the bucket with the largest single jump — that is your reprice window candidate.
• Compare the frame-level volume column to the sum of the tape fills for the same window; agreement is the reconciliation that keeps volume honest.
• Benchmark the daily total against the median trade size (42,254 shares); volume in shares divided by the median is an approximate trade count for the day.
• Store the per-bucket volumes in a small table; the volume ladder across the day is the market's activity signature.

Volume on Polymarket is cumulative-per-frame and tape-reconcilable, which makes the venue's volume claim independently checkable: sum the fills and the frame-volume column must agree, or the pipeline is wrong.

The measured size texture — median 42,254, p90 65,081 — is the column behind most volume statements: a 200,000-share day is statistically a handful of big trades on a thin book, and the page exists to make that obvious.

The volume-signal habit is bucket-and-scan: chunk the frame volume into five-minute windows, find the largest jump, and pull the ladder at that second — the reprice window announces itself in volume before the mid chart even flinches.

Volume also serves as the venue's clock: the frames with the most contracts traded are overwhelmingly frames where the reference moved, so a volume spike with no reference change is rare enough to be studied as an anomaly. Pairing the tape and the reference on the UTC second keeps that clock disciplined.

Volume is the venue's activity truth, and the archive's frame-level column makes it self-checking.

Read volume through the size distribution: the median trade of 42,254 shares is the unit that actually moves these books.

Bucket the day and find the reprice windows by volume jump; volume is the cheapest reprice detector on the site.

Volume is the activity column of every study: the metrics page carries it per frame and the flow page reads it as a signal.

This page pairs with the tape page for reconciliation and with the VWAP page for what volume-weighted means on the venue.

The median-trade lens (42,254) is the venue constant this page turns into a figure of speech for the rest of the stack.

Volume is the venue's activity truth, self-checking between the tape and the frame column.

Read it through the size ladder (median 42,254) and bucket it to find the reprice windows; volume is the cheapest reprice detector.

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