Polymarket tennis data: 94k markets across ATP, WTA, and ITF

Polymarket prices tennis tour-wide: match winners, set and map handicaps, and over/under games. The live book for any court is free to read from the official API with no key.

The missing piece is history. Book depth is a current-moment object — the /book route serves now, nothing else — so reconstructing how a set was priced as it played needs an archive that recorded every tick. PolyOrderbooks captures full L2 ladders four times per second for the tennis series it lists.

Tennis is one of the three tours where the 34,000 map-and-set handicaps in the catalog concentrate: a match is priced as a sequence of set markets, and each set prices as its own binary question.

Every one of those questions stores as a complete book — bid and ask ladders with sizes, a captured_at timestamp, outcome metadata on resolution — so a 'tiebreak at 2-1, five games in' snapshot is as recoverable as the opening price.

• Live ladder — official /book, free and keyless, current moment only (what the API serves).
• Historical depth — full L2 books at 250ms over REST, free Starter with 3 days, paid windows to 120 days.
• Discovery — /v1/series for tours; /v1/markets for match questions and slugs.
• Coverage — measured counts per sport on the sports data page.

A Grand Slam match is a set-long repricing machine: the favorite tightens when a break lands, loosens on the return game, and settles at the match point. Every one of those frames leaves the official API the instant the handshake happens, so the archive is the only place the final-set tiebreak exists as numbered rows.

The tour structure explains the market shape: ATP, WTA, ATP Doubles, ITF, and LPGA Tour series each trade match winners, set and map handicaps, and over/under games, and a two-set lead reprises the winner line in a single break game.

The documented behaviors worth verifying against the record include: whether set handicaps trade tighter than match winners, whether favorites drift in their odds between service games, and whether an upset final set produces a clean arbitrage of stale props before the archive reprices.

Pulling a Slam final is two steps: /v1/markets for the match question, then the archive query on the third-set window, which hands back the complete bid/ask ladder with sizes at every recorded second.

The momentum windows are service game breaks, and the tiebreak churn sits inside the last two games of each set — that is where the executable side changes hands most.

With 94k-plus markets, tennis gives studies a deep but bounded sample across tours and surfaces.

The usual tier line applies: 3 days free at 250ms, Pro to 120, and resolved matches stay queryable on both.

Study one: measure whether set-handicap books trade tighter than match-winner books across a Slam fortnight, normalised by the one-sided share that dominates the venue's books.

Study two: replay a final-set tiebreak and count the executable-side flips inside the deciding games, mirroring the 26-second family-collapse timing studied in the crypto corpus.

Study three: test favorite drift between service games — whether an ace hold tightens the line immediately and whether a break against tightens it further — using 250ms frames around the changeovers.

One Slam is the honest minimum sample; its fortnight of matches fits the Pro window, while a single final round is what the Starter free tier can hold.

Start with one deciding set: pull the third-set window of any Grand Slam match from the archive and count how many times the executable price flipped from one side to the other between changeovers. That count is the raw material of every momentum claim on this page.

Then pull the same match's match-winner book and the set-handicap book at one identical second and compare their widths. Two books, same timestamp, different grid discipline — that juxtaposition is the whole tennis market story in one frame.

Turn it into a routine: one Slam fortnight, one window per notable match, and at the end resolve every pulled market so the fortnight itself becomes a queryable mini-corpus.

• Sports data
• Market data
• Order book data