Tennis data, short essays, analysis

by Matthew Willis

20 posts

Court diagrams of rally forehands hit from the middle or backhand side against right-handers in 2025–26. The top 30 average splits 49% to the deuce side and 51% inside-out to the ad side; De Minaur sends 30% to the deuce side and 70% inside-out to the ad side.

De Minaur's forehand mostly goes one way

This was arguably the most influential point of the match, a break point De Minaur created and dominated until the last shot in set 1.

October 20262 chartsVideo

Table of the average rally forehand for each of the ATP top 10, from contact to bounce: contact distance behind the baseline, ball speed, spin, net clearance and depth. Sinner hits it fastest (82 mph, 3,177 rpm); Medvedev makes contact furthest back (1.30m), hits slowest but one (75 mph) with less spin (2,453 rpm), yet clears the net by the most (76cm) and still lands it 3.67m from the baseline.

Anatomy of a forehand

Things like spin rates and groundstroke speed often get shown on broadcast but they're pretty incomplete in isolation.

September 20262 charts

Table of points won in short (0–4 shots), medium (5–8) and long (9+) rallies for the ATP top 20 plus Fonseca. Sinner wins the most short rallies (57.7%) and, with Fils, the most long ones (56.4%); Alcaraz is best in 5–8 shot rallies (56.5%); Shelton, Mensik and Lehecka win under 47% of rallies of nine shots or more.

Who wins the rallies?

Top 20 (+ Fonseca) win rate by rally length

September 2026

Court map of where Jódar's first-serve returns landed in 2026. 32% land deep (tour 26%, top 20 29%) and he wins 62% of those points (tour 53%). 63% of his returns go in (tour 65%), at an average of 70 mph against 63 mph for the tour and 65 mph for the top 20.

Rafa Jódar's first-serve return

Jódar’s return is finally getting more attention but it’s worth looking at why.

August 2026

Grid of the share of first serves in play that go unreturned, by how many games each set of balls has been used, on men's tour-level hard courts in 2022–26. Fresh balls average 43% unreturned and worn balls 34%, dropping from 42–44% with new balls to the mid 30s after five or more games.

Fresh out the can

This is so interesting because there's a measurable impact from newer/older balls on rates of aces, holds, unreturned, double faults, 1st serve in %, rally length et al.

July 2026

Court diagrams of every wide and T first serve Sinner and Zverev landed in play in 2026, with dashed lines marking each corner's average distance from its target line. Sinner's corner first serves land 50cm from the line on average (52, 48, 51 and 48cm across ad wide, ad T, deuce T and deuce wide); Zverev's land 57cm away (65, 52, 50 and 62cm), with his T serves as tight as Sinner's and the gap coming from his wide serves.

Spot Sinner

Sinner was serving a bit slower than his 2026 averages (in image) today, but it's funny that this matchup is often still framed as Zverev having a slight serving edge.

July 2026

Distribution of where Auger-Aliassime's and Djokovic's rally backhands land, from the net to beyond the baseline. Both make contact about 0.7m behind the baseline, but Auger-Aliassime's heavier, higher backhand (2,368 rpm, 0.83m over the net) lands at a median 7.9m against Djokovic's 8.4m, and he misses more into the net (6.3% against 4.6%) and long (6.6% against 4.4%).

The backhand consistency gap

A huge consistency gap downstream of backhand technique

July 2026

Chart of how often each of the ATP top 20 plus prospects plays a forehand, backhand or volley as their first shot after a first serve, and how often they win the point with each. On average 68% are forehands; Darderi hits the most (81%), Zverev and Medvedev the fewest (52%). Auger-Aliassime's forehand edge is the largest at 15 points (62% against 47%), Medvedev's the smallest at 2.

Serve +1 wing split

How often do the Top 20 find their forehand on serve +1 (the shot after the serve) and how much of a cost is there to them hitting backhands instead.

June 2026

Scatter plot of first-serve returns in play against first-serve return points won for the ATP top 20 plus Jodar, Fils and Fonseca. Sinner and Alcaraz win 35% while getting 68–69% back in play; Jodar also wins 35% from 64% in; Shelton is lowest at 24% won from 61% in. Tour averages are 65% in and 28% won.

First-serve returns: in vs won

1st serve returns in play vs 1st serve return win rate 2025-2026. Top 20 + Jodar, Fils, Fonseca

June 2026

Chart of the share of shots that are drop shots on hard courts and on clay for the ATP top 20 plus Jodar, Tien and Fonseca. Bublik uses them most (5.8% on hard, 8.5% on clay) and every player hits more on clay except Vacherot. Sinner wins 60% of drop-shot points on hard and 66% on clay; Alcaraz 59% and 60%.

On clay, softer hands pay

Every Top 20 player hits more drop shots on clay than hard courts except Vacherot.

May 2026

Court diagram of where top-10 players make return contact on wide first serves in 2025–26, split by service box and by the server's hand. On the ad side they make contact 2.63m outside the singles line against left-handers and 1.60m against right-handers; on the deuce side it is 1.34m against left-handers and 2.30m against right-handers.

Out Wide, Out Wider

Lefty servers really do seem to have a special something on slice wide serves.

April 2026

Two flow charts of first-serve points for top-20 servers in 2025–26: against each other (72.4% won) and against Sinner or Alcaraz (64.8% won). Against the pair, fewer first serves go unreturned (32% against 38%), and the server's +1 shot wins less often on both the forehand (6.5 points lower) and the backhand (8.4 points lower).

The Sinner & Alcaraz Tax

Sinner and Alcaraz tax the top 20 the way they tax everyone else. And nobody can tax them back.

April 2026Essay, 2 min read5 charts

Clay court maps of where Alcaraz made rally contact at Monte Carlo 2026. Before the final, 60% of his contacts were forehands (416 forehands, 268 backhands); in the final against Sinner, 46% (116 forehands, 134 backhands), with backhand contacts clustered on his left behind the baseline.

Alcaraz Rally Contact

It was so much harder for Alcaraz to play forehands out of his backhand corner (usually a major weapon) in the Monte Carlo final vs Sinner.

April 2026

Court maps of where Sinner and Alcaraz make contact when returning first serves, split by hard court and clay and coloured by point won, lost, return winner or error. Sinner averages 1.63m behind the baseline on hard and 2.78m on clay; Alcaraz averages 1.96m on hard and 3.22m on clay, with a second, deeper cluster around 5m on clay.

Return Contact Point

A nice way to look at differences in hard and clay court strategy is return positions.

April 2026

Dot plots of hard-court hold percentages for top-50 ATP players in 1994 and 2025/26. In 1994, 44% held at least 80% of their service games, with Sampras around 87% and Ivanišević about 84%; in 2025/26, 84% do, led by Sinner at about 92.5%, with Ruud near 87.5% and Davidovich Fokina near 84%.

So much for the serve era

Casper Ruud holds serve more often than Sampras did on hard courts in 1994 & Davidovich Fokina holds serve more often than Ivanišević.

March 2026

Two court diagrams of Alcaraz's ad-side first serves on hard courts over the last 52 weeks. On all points (n=667) he goes 41% wide, 15% body and 44% T, winning 72 to 74% out wide and down the T. On break points (n=56) 71% go wide or body (48% wide, 23% body, 29% T) and he wins only 59% of the wide serves.

Alcaraz has a serve problem

Possibly a crazy thing to be writing about the current world No.1 and recent Australian Open champion. But margins are fine in tennis and his biggest rival has some hard court gifts that can make even small problems loom large.

March 2026Essay, 4 min read3 charts

Scatter plot of the ATP top 20 by average first-serve speed (114 to 130 mph) against first-serve accuracy, measured as average landing distance from the side or centre line (lower is more precise), coloured by player height. Zverev and Mensik serve fastest at 128 mph, Djokovic is the most accurate at about 50cm at the tour-average 119 mph, and Fritz and Mensik combine pace with accuracy of around 51 to 52cm.

Sine of a Great Serve

It's fairly obvious that taller players have an easier time hitting accurate serves (closer to the centre or side line), but I wasn't expecting quite such a stark grouping difference.

March 2026

Bar chart of the share of returns hit with topspin rather than slice or block for the ATP top 20, sorted by topspin share; the tour average is 84% topspin. Sinner hits 95% topspin; Bublik slices or blocks the most at 32%, followed by Musetti at 29% and Ruud at 22%.

The block return

The block return is pretty rare at the top of the men's game these days.

February 2026

Bar chart of the forehand share of groundstrokes for the ATP top 20, against a top-20 average of 53.9%. Auger-Aliassime hits the most forehands (59.0%), and Medvedev is the only player below half, at 45.7%.

Medvedev, the backhand outlier

Medvedev is the only player in the Top 20 who hits more backhands than forehands

February 2026

Line chart of average men's rally length at Wimbledon, in shots per point, alongside relative air density from 2004 to 2025. Rally length moves between about 3.4 and 3.95 shots with no long-term rise and tracks air density, with a correlation of 0.57 overall and 0.74 for year-to-year changes.

Is Wimbledon getting slower?

People love to say Wimbledon keeps getting slower despite consistent grass length (8mm) since 1995, seed type (ryegrass) since 2001, ball pressure since 1995

February 2026