🎾 What happens at different levels of the competitive ladder — and how players' skills shape the structure of a match.

This is the first in a series on differences in how tennis matches play out at different levels of the sport. I compare women’s and men’s recreational matches, Division 1 college matches, International Tennis Federation events, and WTA/ATP tour level events. Data can tell great stories: here’s one of them.

Point Outcomes

What changes in the structure of points as a match level climbs from recreational to professional? We often talk about speed, power, and consistency, but few analyses show how rallies, points and matches evolve across those levels. Using shot-by-shot data from recreational, collegiate, and professional men’s and women’s matches, I look at what happens as the game matures.

We often hear that most points are won through an opponent making an error. This adage comes from the professional level. But recent computer vision technology finally make it possible for non-pros to track shot-level data. Here, I leverage a relatively new data source to test when intuitions about errors and rallies for recreational and college play are correct and where they aren’t. (More on the data sources at the bottom.)

The results reveal distinct patterns that at face value are what we would expect, For coaches, players, and fans, the nuances in this data highlight not just how matches are playing out, but where the training emphasis at each level could be as well as what the process of improvement looks like in the aggregate.

First, we'll look at groundstrokes.

Groundstroke Balance: Power vs. Precision

These datasets reveal a subtle evolution in shot selection and risk between the forehand and backhand sides. As is widely cited by coaches and commentators, tennis is a game of limiting errors. Forehand and backhand errors are indeed the first and second most common ways that points end, and the share of all points that end in an error indeed hover around 50% (not including double-faults) at all levels and for both men and women. But their contributions at different levels is worth noting.

For women, forehand errors make up just about 30% of every 100 rally-ending shots (I’ll just call these: points or point outcomes) at all levels. But forehand winners show a different pattern. The number of both women’s forehand winners are even through the D1 college level. But as a share of points, forehand winners are fewer through the D1 level and then jump from 12% to 16% of point outcomes in the WTA. 

Backhand errors as a percentage of all rally-ending shots (points), on the other hand, climb through DI before settling back down around 23% of points in the WTA. And at the same time, the number and percent of backhand winners increase through the D1 level and settle around 9% of points. There’s a few different things going on here…so what’s the story? 

Women
	Rec 3.5 and 4.0	Rec - all	D1	ITF	WTA
BH winners/ errors	0.37	0.40	0.35	0.39	0.39
FH winners/ errors	0.47	0.43	0.42	0.52	0.56

While the share of backhand winners-to-errors is fairly stable at each level at just over 1 winner for every 2 errors, forehand winners-to-errors for women grow noticeably in the average WTA match. The number of forehand errors increases through the pros, but only at the ITF level do forehand winners catch up. 

All of this (plus some evidence on serving below) suggest that women’s college and professional players must use their backhands more often, creating more errors and winners. As the backhand winners-to-errors ratio doesn’t change much, the backhand appears to just be used more overall–perhaps by necessity as women’s forehands (and serves–see below) are more potent at higher levels of tennis.

For men, the shift is even clearer. From the rec level to the ATP tour, the share of forehand errors steadily decline by 5 percentage points, while forehand winners also rise by 6 percentage points. [Due to the few (6) men’s ITF matches in the MCP database, those data are shown in the figures for interest, but not referenced or used to draw any conclusions.]

Similar to the women, the number and share of backhand errors and winners both increase through men’s D1 matches. Then, the number of backhand winners and errors is slightly more at the ATP level, but the share is back on par with recreational players. And while the winner-to-error ratio for backhands stays fairly constant, the forehand winner-to-error ratio nearly doubles from 1-to-2 to just under 2-to-1! 

Men
	Rec 3.5 and 4.0	Rec - all	D1	ITF	ATP
BH winners/ errors	0.35	0.36	0.37	0.27	0.37
FH winners/ errors	0.34	0.39	0.47	0.40	0.59

For an average ATP match, forehand winners are on par with service winners after occurring about half as often at the lower recreational level. It’s very clear that in addition to increased potency of the forehand at higher levels of the men’s game, elite men increasingly build patterns to hit forehands rather than hit evenly across both wings.

For both men and women, improving the backhand is clearly part of the difference between less and more skilled play, serving as a rally-neutralizer. But the forehand becomes the engine of improvement on the baseline.

Next, we’ll look deeper into the serve.

----

Recreational and collegiate data come from a selection of anonymized matches supplied SwingVision from February-August of 2025. ITF, WTA, and ATP level data come from the Match Charting Project from 2010-2025. There are 105 men’s recreational and 36 women’s recreational matches in the SwingVision database used here. Thirty-six and 20 of those are for men’s and women’s matches where the SwingVision customer self-rated as 3.5 or 4.0 on the USTA NTRP scale. No matches with a rating lower than 3.5 are used here. Collegiate data are entirely from Division I teams–10 men’s teams and 6 women’s teams. As of September 2025, there are 4932 men’s and 3146 women’s matches in the MCP database since January 2010. Six of those are ITF-level men’s matches and 119 are women’s ITF-level matches.