livetennisapi/tennis-match-outcome-studies

Tennis Match Outcome Studies

Aggregate outcome statistics counted from 116,382 completed best-of-three singles tennis matches on the ATP, WTA, Challenger and ITF tours, played 2023-01-01 to 2026-09-14.

Counted from the point-by-point record rather than estimated from final scores. Retirements and walkovers are excluded.

CC BY 4.0 — free for any use including commercial, with credit. No account, no key.

The finding most people do not expect

The margin of the first set predicts a comeback about three times more strongly than the surface does.

first set lost	comeback rate	matches
7-6	21.92%	14,265
6-4	19.41%	25,546
7-5	18.89%	10,252
6-3	17.39%	25,970
6-2	13.30%	19,739
6-1	11.06%	14,269
6-0	7.18%	5,833

That is a 14.74-point spread. Across surfaces the spread is 0.73 points (grass 17.20%, clay 16.79%, hard 16.47%); across tours about three (ATP 18.21%, WTA 17.04%, ITF men 15.18%).

So "lost the first set" is a weak feature and "lost the first set 7-6" is a strong one.

What is in the file

Four studies, aggregate figures only — there are no per-match rows.

study	measures	corpus
first-set-comeback	comeback rate by first-set score, surface and tour	116,382 matches
second-set-win-probability	first-set loser's chances by second-set score	116,382 matches
hold-rate	service hold % by surface and tour	2,239,440 service games
tiebreaks	share of sets and matches reaching a tiebreak	116,382 matches

A validity check worth knowing

The same corpus reproduces the widely published ATP (79%) and WTA (64%) service hold rates — 78.62% and 64.36% — on a method that was not tuned to match them. Tiebreak games are excluded.

What is wrong with it

Stated here rather than left to be discovered:

• Two breakdowns do not sum to the corpus. by_tour falls 11 matches short and by_first_set_score 508. A re-measurement found no missing tour and no missing scoreline, so this is counting drift in those published rows, not an exclusion.
• The surface split covers 113,288 matches, not 116,382, because roughly 3,100 matches never had a surface stated by the feed. That one is a genuine exclusion.
• The rates are sound. Every tour matches exactly on re-measurement and no first-set rate moves by more than 0.02 percentage points. The conclusions rest on the rates, not the counts.

Use

import json, urllib.request
d = json.load(urllib.request.urlopen("https://blog.livetennisapi.com/studies.json"))
{s["id"]: s.get("matches") for s in d["studies"]}

If you re-run any of it and get something different, we would like to know.

Citation

Live Tennis API (2026). Tennis Match Outcome Studies. CC BY 4.0. https://blog.livetennisapi.com/open-tennis-data

Documentation: livetennisapi.com/open-tennis-data · Source file: studies.json

These are aggregates. The per-match records behind them are in the Live Tennis API's history endpoints; this dataset needs neither an account nor a key.