Win the first set, win the match? 26,900 settled Polymarket tennis matches

The player who won the first set won 82.9% of 26,900 settled Polymarket singles matches from November 9, 2025 to October 10, 2026, give or take 0.45 points. Better than four in five. But on the 17,826 best-of-three matches where both of the first two sets had a market, the second-set winner won 85.8% and the first-set winner 83.4%, because the player who levels the match takes the decider 53.9% of the time.

So the first set is not the most telling set in a best-of-three match; the second is. Read a set-all match as close to a coin flip with a lean to the player who just won the second set, and let the two players matter more than the order of the sets. Best of five is the forgiving format: there the first-set winner won 74.6%.

73.4% of Polymarket's 37,231 settled singles match markets had a settled set market. The sample is every singles match with a settled match market and a settled first-set market, from the first on November 9, 2025 to October 10, 2026: 26,900 matches. 13,293 of them, 49.4%, are on the ITF tour, whose matches enter the sample in June 2026; before then it was ATP and WTA only.

Set markets name players by surname and match markets by full name, so a set result maps to a match result when the full name ends with the surname. Doubles and matches between two players with the same last name are left out, and so is any market that settled 50-50. Every rate carries a plain binomial half-width.

Tour is the series Polymarket files the match under, and each is broader than its name. ATP here means tour events and their qualifying rounds, junior boys' matches at the Grand Slams, and smaller events such as Seville and Genoa. WTA includes qualifying and junior girls, and ITF is the ITF World Tennis Tour for men and women.

Format comes from the markets. A match counts as best of five when Polymarket lists a market only a best-of-five match can have: a Set 4 or Set 5 winner market, a -2.5 set handicap, a Total Sets line of 3.5 or 4.5, or a match games total of 30 or more. That flags 532 matches. A tournament title would not do the job: Polymarket files Australian Open qualifying, which is best of three, under the same title as the men's main draw.

The pooled rate climbed as the ITF's share of the sample grew: 82.2% in the first quarter of 2026, when it held no ITF matches, 82.1% in the second, 33.5% ITF, and 83.3% in the third, 66.5% ITF. The first ten days of October came in at 84.5%, 71.6% ITF. The first set decides most on the ITF tour, so a sample that tilts toward it raises the total even if no tour changes.

Hold the mix still and most of the rise goes away. Best-of-three matches from September 14 to October 10, 2026 came in at 85.1%, against 82.5% before; weighted by the earlier mix of tours, they come to 82.9%. 2.2 of the 2.6 points are the mix.

The tours tell the two apart. The ATP and WTA together ran between 78.3% and 83.3% a month from January, with no rise to match the total. The ITF ran between 83.8% and 86.6% a month, and its own rate did rise: 86.7% since September 14 against 84.6% before, a gap wider than both half-widths, so the part of the rise that is not the mix is an ITF rise. Read the rate for the tour you are trading, not the total.

Any of these rates is an average across every kind of first set, and the average hides the matchup. A favorite who wins the first set has done what the price expected. An underdog who wins it still has to beat the favorite again. This study has no prices to split those cases, so read the tour's rate as the base rate a live price starts from, not the price itself.

The first-set winner won 85.3% of 13,293 ITF matches against 80.4% of 8,131 best-of-three ATP matches, a gap wider than both half-widths together. The WTA, at 81.6% over 4,944 matches, overlaps the ATP, so this data shows no difference between women's and men's best-of-three tennis.

The ITF gap comes with lopsided first sets. 40.2% of ITF first sets ended 6-2 or wider, against 26.3% in the ATP, and a first set won that easily is the hardest to come back from, as the first-set score section below shows.

Best of five is the exception. In 532 best-of-five matches the first-set winner won 74.6% of the time, with a half-width of 3.70 points. A player who drops the first set has four sets left and needs three, instead of two left and needing both, and that room shows up as about 6 more points of comebacks.

Polymarket's Total Sets Over/Under 2.5 settles Over when a best-of-three match needs a third set. Of the 25,737 settled on best-of-three singles matches, 31.8% went Over, with a half-width of 0.57 points: 36.8% in the ATP, 34.9% in the WTA, and 27.7% on the ITF tour. The ITF's lower third-set rate goes with its higher first-set rate, and the mix moves this total too: 33.1% of matches up to September 13, 2026 went to a third set, and 27.4% of those since, 75.4% of them ITF.

In best of five, 55.0% of 509 settled Over 3.5 markets went to a fourth set and 23.9% of 331 Over 4.5 markets to a fifth.

The set markets and the Total Sets market agree. On the 17,486 best-of-three matches that had both, a split of the first two sets matched an Over every time, and a player winning both matched an Under every time. Two kinds of market, settled separately, tell the same story about the same matches.

17,826 best-of-three matches had both a Set 1 and a Set 2 market settled. In 69.2% of them one player won both sets and, as the format requires, the match. The other 5,490 went to a deciding set.

In those deciders the player who had won the first set won the match 46.1% of the time, with a half-width of 1.32 points, so the player who won the second set won 53.9%. The whole interval sits below 50%, and the lean holds on both sides of the tour split: the first-set winner took 46.6% of ITF deciders and 45.2% of ATP and WTA ones.

Winning the first set counts for a lot before the second set is played. Once the second set has gone the other way, the player who leveled the match is the slight favorite in the decider. This data does not say whether that is momentum or the better player taking over.

Best of five shows the same pattern on fewer matches. When the first two sets were split, the first-set winner won 40.7% of 108 matches. When one player won both, they went on to win 92.0% of 137 matches, so a comeback from two sets down happened 8.0% of the time.

497 of the 532 best-of-five matches came from the men's main draws at Wimbledon, Roland Garros, the Australian Open and the US Open. Most of the rest are the Next Gen ATP Finals, which plays best of five with short sets, and a batch of Wimbledon qualifying matches from June 25, 2026 that Polymarket listed with best-of-five markets. So the best-of-five rate is a Grand Slam rate, and it grows mostly during four fortnights a year.

Two are ATP Finals matches, a best-of-three event, that Polymarket listed with games totals of 36.5 and up, one of them also with a -2.5 set handicap. They stay in the best-of-five group because the rule reads the listing, and two matches out of 532 barely move its rate.

How the first set was won matters nearly as much as who won it. A first set won 7-5 or 7-6 was worth 11.1 points less to its winner than one won 6-2 or better, 77.4% against 88.5%, a gap far outside both intervals. When the match price moves after the first set, check the score: a 7-6 set should move it less than a 6-1 set.

A third set came in a little more than one ATP match in three, so a Total Sets Over 2.5 price there starts from about 37 cents before the matchup moves it, and nearer 28 on the ITF tour. Two close players push it up, a mismatch pushes it down, and a close first set is the first sign of which you are watching.

It has no prices. It does not say what a player was priced at after winning the first set, or whether that price was right, because the set markets on record settle at the same time as the match and carry no time for the end of the set. This study counts results, not prices.

It does not split by player strength. The 82.9% mixes favorites who won the first set with underdogs who did, and the right number for any one match sits above or below it.

The samples are the matches Polymarket listed set markets for, and that list has changed. ITF matches were 17.7% of the 12,640 matches this page first reported and are 49.4% of the 26,900 now, so a pooled rate tracks what Polymarket lists as much as how tennis is played; the tour rates are the steadier figures. The 17,826 best-of-three matches with a Set 2 market went to a third set 30.8% of the time and followed the first set 83.4% of the time, close to the 31.8% and 83.1% of the larger samples, so the decider figures describe ordinary matches. Tour labels come from Polymarket's series, which mix tiers, and doubles are left out.

Among singles matches with a Completed Match market, about one in 17 did not finish, and each market on it settles that differently. The match market pays $1 on the player who advances, including by retirement, default or disqualification, and a walkover settles it 50-50. A set winner market settles only on a completed set, and settles 50-50 if that set is not finished. Total Sets settles on any result already decided when play stopped.

So if a player retires during the second set, the match market pays their opponent, the Set 2 market settles 50-50, and a Set 1 bet already won stays won.

Polymarket lists each match as an event: a match market titled with the tournament and both players, "US Open ATP: Frances Tiafoe vs Ben Shelton", and around it the set markets (Set 1 Winner, Set 2 Winner, a set handicap, Total Sets, Total Games, and games totals for single sets). Because the match and each set settle to a result, what the first set says about the match is a count, and that count is this study.

Tennis is the live sport on the feed. From September 11 to October 10, 2026, large tennis buys landed mid-match about twice as often as large buys across sports, and nearly every large tennis sell did. If you trade tennis, you trade it during play.

The tennis board lists live and upcoming matches that have a Polymarket market, with both players' prices, and live matches show the games of each set, the point and who is serving. With Pro, a match's event page lists its markets by family (first-set winner, set handicap, Total Sets, Total Games), each with its price change over the last hour and 24 hours, and a Volume by family bar; without Pro, that section shows labeled sample data.

The large-trade feed lists recorded taker fills of $10,000 or more that meet its market-volume-share floor with the grade of the wallet behind it, live on Pro and 24 hours behind without it. Cards on the tennis board list the top five holders on each side, with grades, for every visitor.

The timing study and the cash-out study measure the same tilt toward live tennis over the spring and summer, and the tennis guide covers each tennis market. Since most of the large money arrives after the first serve, the set score is the context every large tennis trade needs.

Every figure in the text comes from one read-only run on a copy of the production database at 15:29 UTC on October 11, 2026, except how matches ended, which comes from an earlier run at 13:34 UTC on October 11, 2026. The query text and the raw output of both are public in the 0xinsider research repository, under polymarket-sports-markets, in the files whose names carry the words tennis-sets; this reading is the tennis-sets-rerun pair.

The universe is every settled Polymarket tennis singles match with a match market and a first-set market, both settled to a player, where the two players' last names differ. The best-of-three and best-of-five splits use the format flag described above; query 7 lists the tournaments behind it.

This page first ran at 19:46 UTC on September 13, 2026, on 12,640 matches, with the first-set winner at 80.7%. Re-run on October 11 with the same rules, the same window holds 20,777 matches, because 0xinsider has since recorded settled set markets for 8,137 more, 6,437 of them on the ITF tour, and the weeks since add 6,123. Every figure on this page is the October 11 reading. The first set still wins about four matches in five, it still decides most on the ITF tour and least in best of five, and the second-set winner still takes the decider; the pooled rates moved with the mix of tours, as the sections above show.

The best-of-five rate will move most, since it grows only during the four Grand Slams; none of the 6,123 matches from September 14 to October 10, 2026 was best of five.