In Saturday’s Australian Open final, Elena Rybakina won 92 points. Aryna Sabalenka won 92 points. Rybakina won 76% of her first serve points; Sabalenka won 75%. Both players held on to 48% of their seconds. Even their average first serve speeds were nearly identical, Rybakina’s 178 km/h nipping Sabalenka’s 177 km/h.
Only a few moments really mattered. Sabalenka converted two of eight break points. Rybakina converted three of six.
With such narrow margins, we should be cautious to draw conclusions about tactics and player skills. Flip one or two of those break opportunities, and it would have been a very different trophy ceremony. Anybody who tries to tell you “why” Rybakina won should keep that in mind. Still, Sabalenka would surely like to know how to secure another half-dozen points and put the result out of the range of luck. Rybakina will hope to do the same.
Pick target, hit target
Rybakina is the best server in the women’s game. Her ace rate over the last year is better than 10%–a percentage point ahead of second place (Osaka), and miles ahead of Sabalenka’s 6%. Rybakina has won nearly 75% of her first-serve points, while no one else cracks 73% and only a few players are on the north side of 70%.
At key moments on Saturday, Rybakina dazzled with her ad-court serves out wide. She saved the only two break points she faced in the first set with back-to-back unreturned serves, both wide. She finished the match with another signature delivery, acing Sabalenka out wide on match point.
If you’re looking for a “why,” it’s tempting to focus on those wide ad-court serves. Rybakina made 18 first serves when she aimed for that corner, and she won 14 of those points.
But! It’s not the ad-wide corner, specifically. Rybakina was even deadlier when she targeted Sabalenka’s backhand corner in the deuce court. She landed 14 of those first serves, winning 13.
Here’s the Rybakina method for defeating the world number one:
- Have a world-class serve
- Aim first serves at the backhand corner
- Make half of them
Easy, right?
Apparently not easy
Fair enough, most players don’t have anything like Rybakina’s serve. A few–Osaka, Noskova, Qinwen–can do a decent impression on a good day. Still, it’s an uphill battle to knock off Sabalenka with aggression from the line.
What’s striking, though, is that most opponents don’t really try.
Across 120+ charted matches since the beginning of 2024, Sabalenka’s opponents aimed their first serve at her backhand corner about 40% of the time. (That doesn’t mean they aimed 60% at the forehand corner: A fair number of first serves don’t land close to either corner.) In the vast majority of matches, her opponent aimed half or fewer of their first serves at her backhand corner.
On Saturday, Rybakina targeted the backhand corner 63% of the time.
The first serves that landed in were so devastating in part because she took a low-margin approach. Rybakina already misses more first serves than almost anyone on tour: Her 57.4% first-serve-in rate is worse than 45 of the top 50 women. Against Sabalenka, she succeeded exactly half the time when she fired in that direction. Corner-aimed serves are (unsurprisingly) lower-percentage for everybody, but her 50% was even worse than tour average.
It’s a smart tradeoff. Combine the two numbers, and we see that on 32% of her service points, Rybakina put a first serve in play to Sabalenka’s backhand corner. Those, as we’ve seen, are as close to guaranteed points won as you can find. Sure, that leaves 68% of service points to worry about. Yet as much as Rybakina’s premier weapon glitters, she’s a solid average at everything else. She’ll pick up a lot of those other points with quality second serves or rocket-powered firsts to the forehand corner, or by winning baseline rallies.
In the past two years, only a handful of players have managed to put first serves to Sabalenka’s backhand corner on as many as 32% of points. Even then, it doesn’t always work: Marketa Vondrousova, for instance, is unparalleled at hitting her targets, but her deliveries are softballs in comparison. For the players who can serve big, though, Rybakina may have pointed the way to tougher challenges against the world number one.
Ka-zam
This might be a recent refinement to Rybakina’s match tactics. We have over 80 charted matches for her since the beginning of 2024, and she has rarely aimed so many of her serves at the backhand corner. To be clear, she doesn’t need to. She straight-setted Sabalenka for the year-end title in November with only 44% of first serves pointed at that target.
But suggestively, Rybakina hit nearly as many first serves to the backhand corner in her Australian Open quarter-final match against Iga Swiatek. While she wasn’t quite as successful, landing just 40% of those attempts, the end result was encouraging. Even with all the misses, backhand-corner firsts accounted for a quarter of her service points. And she was as eye-poppingly successful on those points against Iga as she was in the final. Swiatek salvaged just 1 of 12.
It remains to be seen whether this is repeatable. When Rybakina is serving at her best, peppering the backhand corner is probably a good way to take advantage. (Unsurprisingly, since this is something tennis coaches tell twelve-year-olds.) If she’s misfiring, low-percentage first serves are probably not the way to fight her way through.
And surely, the world number one will start taking a few more backhand-return reps. She doesn’t have to turn into Andre Agassi to negate Rybakina’s new-found advantage. She just needs to defend that corner a little better. 94 or 95 points would have gotten the job done on Saturday. Even against a world-class serve and superb tactical execution, Sabalenka won 92. The two women will continue jostling for an edge, and it looks like the battle will increasingly take place with Sabalenka leaning to her left.
Good work indeed
Always good work, Jeff. Pete Sampras would have approved of that AO women’s final outcome.
Hi Jeff, I really appreciate your work, and as a student using your atp datasets for my final year stats project, I was wondering if your GitHub repository is still active and if you were going to upload a 2025 version?