Our point-by-point tape used to begin in January 2023. It now has a second half behind it: 97,901 matches played between 2013 and 2022, 14,340,663 point rows — the score sequence behind results that were never streamed.
It is a reconstruction from the public record, not a recording, and everything worth
knowing about it follows from that one fact. Nobody watched these matches. Every row is
null on timestamp, win_probability_p1 and danger, and always will be: the table
has no timestamp column to fill, so no later backfill can put a wall clock on a 2016
point. Our rule for the 2023+ tape — a real timestamp means we watched that point —
still holds across both halves of the product.
We said in August that we had no way to rebuild per-point data before 2023. That was true of what we held then; this is what changed.
What is in it. 99.43% of those matches advance point by point; the remaining 556 resolve to one row per game, 555 of them in 2013, and every season from 2015 on is wholly point-granular. A tape averages 146 rows. 31.0% of the corpus — 30,306 matches — is qualifying-draw play: point-by-point qualifying at that scale is genuinely rare, and it is the part of the corpus you are least likely to hold already. ATP contributes 64,785 matches and 9,676,061 rows, WTA 33,116 and 4,664,602 — together, all of it.
What it covers, and what it does not. Of archive matches actually played in those ten years, the tape covers 19.3% overall and 44.9% of tour-level play. Main draws run from 91.6% (WTA International) to 98.7% (ATP Masters); ATP Challenger main draw is 55.3%. Then the thin end: Challenger qualifying 33.6%, and Grand Slam qualifying only 16.0% on the ATP side and 18.1% on the WTA side — so qualifying is deep in volume and shallow at the slams, and those are different facts. Coverage roughly doubles from 2017 onward (12.5% in 2016, 21.3% in 2017, 24.2% in 2022).
ITF and futures are not covered: 25 of 116,575 ATP futures, 68 of 19,162 ITF M15 and 48 of 9,380 ITF M25 matches carry a tape. If your work is at that level, this archive is not for you, and we would rather you knew it here.
How a tape proves it belongs to its match. A tape attached to the wrong match is worse than no tape. Each one proves five things before it is published — the tournament edition, the player names, the round, exact reproduction of the published final score, and a date window — and then passes a 23-invariant interior audit that rejects physically impossible sequences. A tape that cannot prove its binding is refused rather than published against a guess. Across three independent audits, 70 matches were checked against outside sources with zero mismaps; over the full table, 23 of 23 invariants sit at zero violations, with no score mismatches.
What we know is imperfect. 818 within-game point gaps across 633 matches, and 37 tiebreak rows with the same shape — a point occasionally under-reported by the source. Every endpoint is still a legal score, so a tape is conservative, never inflated. The era's records also render a few players inconsistently; Ekaterina Makarova appears as "Kate Makarova" on 270 matches.
How to reach it. One match:
GET /history/archive/matches/{id}/tape — ULTRA, or any Historical Data plan. A whole
year in bulk: GET /history/packages/{YYYY}?kind=archive_tape — core ULTRA, or
Historical Data Pro, Business or an active one-off package. All ten years 2013-2022 are
published. The archive result is unchanged and stays where it was. Ids here are
archive_matches.id, a different space from live match ids.
There is nothing before 2013. Our results archive reaches back to 1968, but no point-by-point record of those years exists to rebuild from, and we do not manufacture one.