Is Sebastian Korda Making Progress?

Also today: Talking Tennis interview

Sebastian Korda in 2023. Credit: Hameltion

It wasn’t long ago that Sebastian Korda was considered one of the best prospects in the men’s game. He won a tour level title before his 21st birthday, then fell one match short at the 2021 NextGen Finals. He reached two more finals in 2022, then began 2023 with a near-miss, a momentous three-hour clash with Novak Djokovic in Adelaide that ultimately went to the veteran.

That result, plus a quarter-final run in Melbourne and another runner-up finish last October in Astana, nudged Korda up to a career-high ranking of 23. While he has since dropped the points from Down Under and fallen out of the top 30, my Elo ratings keep him in the top 25, just ahead of the man who defeated him in Kazakhstan, Adrian Mannarino.

This all represents a step forward for the American, especially since he struggled throughout last year with a wrist injury. Compared to expectations, though, it’s a bit underwhelming. Korda’s father, Petr, is a grand slam champion; Sebastian has said he’d like to surpass him and win two. At age 23, he has plenty of time to develop, but eight of the men ahead of him in the rankings–including three of the ATP’s top seven–are younger still. For all the veteran exploits we’ve seen in the last decade of the ATP tour, superstars tend to make themselves known at an early age.

Last night, Korda recorded his 100th tour-level victory, a milestone that reminds us how much he has accomplished in his budding career. The match itself, however, pointed at some of his limitations. The American edged out big-serving French qualifier Hugo Grenier in the Marseille first round, 6-3, 2-6, 7-6(3). The player Korda aims to be would have progressed with ease. As it happened, he won 89 points to his opponent’s 92, marred by an error-spattered string in which he lost seven straight games. Grenier played well, but he is ranked outside the top 150. Korda didn’t look much better.

What’s missing? The 23-year-old has all the tools to climb higher: a six-foot, five-inch frame; an overpowering serve including a hard slice delivery that looks as if it were inherited directly from his left-handed father; a flexible, assured backhand; and a willingness to step into the court to take control of points. To watch him play, there’s very little separating Korda from, say, Taylor Fritz, yet Fritz is a top-tenner. Is it just a matter of time until Korda closes the gap, or does his game need to change?

Progress report

Let’s start with the positive: Korda’s serve is getting the job done. Yesterday, more than one-third of his serves didn’t come back. That’s in line with the average of the several other charted matches from the last 52 weeks. Only a handful of men end the point so often with their first shot; Fritz and Ben Shelton top 30% but still trail Korda. In a losing effort against Hubert Hurkacz in Shanghai last fall, more than 45% of the American’s serves were unreturned.

Those numbers represent a major step forward. Facing Hurkacz at the Australian Open last year, fewer than one-quarter of his serves ended the point. Korda finished below the 25% mark in matches against Daniil Medvedev and Karen Khachanov at the same event, too. It’s ironic that he won that one against Hurkacz and lost in Shanghai, but there’s no counter-intuitive moral to glean: Unreturned serves are an incontrovertible good.

The 23-year-old’s results are less reliable when the ball comes back. Even when presented with an attackable return, Korda sometimes hesitates. In two matches against Hurkacz last fall, Korda didn’t hit a single plus-one winner or forced error behind the second serve. (The high rate of unreturned serves means that his best deliveries aren’t coming back as sitters, but that hardly means that the remaining returns are all so daunting.) Grenier put 21 second serves back in play yesterday, only one of which the American ended with his second shot. Despite the qualifier’s overt aggression–he occasionally swung wildly for winners against Korda’s seconds–the average point on Korda’s deal ran to 4.3 strokes, an unusually high figure for such a strong server.

Taking first and second serves together, how much does Korda sacrifice with his conservative-seeming mindset on plus-ones? I calculated the percent of 3rd shots (plus-ones) and 5th shots that went for winners or forced errors across all charted matches since 2020. Here are results for Sebi, plus those of a few comparable players and the tour average:

Player                 3rd W%  5th W%  
Hubert Hurkacz          19.0%   20.0%  
Stefanos Tsitsipas      18.5%   20.1%  
Taylor Fritz            18.2%   17.1%  
Sebastian Korda         17.1%   16.9%  
-- Average --           17.1%   17.3%  
Felix Auger-Aliassime   16.8%   17.1%

Korda is just not as aggressive as the more successful of his tall, big-serving peers. He out-winners Felix Auger-Aliassime, but I would argue (and will do so at length, one of these days) that the Canadian’s approach is holding him back, as well. It isn’t that Korda is entirely passive on the plus-one, but given the relatively weak return quality he faces, he should be putting away more than a tour-average rate of second shots.

There is, however, a reason for his unwillingness to swing bigger, and that’s where we’ll turn next.

Something wild

Here’s the same table with two more columns: one for each player’s unforced error rate on the 3rd shot of the point, and another for the unforced error rate on the 5th shot:

Player                 3rd W%  3rd UFE%  5th W%  5th UFE%  
Hubert Hurkacz          19.0%     12.8%   20.0%     11.8%  
Stefanos Tsitsipas      18.5%     11.4%   20.1%     10.2%  
Taylor Fritz            18.2%     11.1%   17.1%      8.8%  
Sebastian Korda         17.1%     13.8%   16.9%     12.3%  
-- Average --           17.1%     10.8%   17.3%     10.4%  
Felix Auger-Aliassime   16.8%     10.9%   17.1%     11.6%

Yikes! Korda is wilder on these shots than the other players, so much so that he ends more points with the plus-one shot than everyone on this list except for Hurkacz. Of players with some degree of tour-level success, only Marin Cilic misses more plus-ones. Denis Shapovalov and Alejandro Davidovich Fokina are roughly equivalent to Korda in this department.

We’ve taken a roundabout path to reach a more general fact about the American’s game: He misses a lot of shots. As a fraction of all groundstrokes, Korda ends points in his favor about 10% more often than the average ATPer. But he commits 20% more unforced errors. His plus-ones are of a piece with his entire ground game, even if they’re a bit wilder. Racking up so many unforced errors without a correspondingly large winner count means, by definition, that his baseline game is a liability. Only that big pile of unreturnable serves is keeping him above water.

Fortunately, Korda is still young, and his game is not set in stone. He missed 13% of his plus-ones yesterday, but that number is trending in the right direction. Here are his winner and unforced error rates on the third shot of the rally, as ten-match rolling averages going back to the 2021 NextGen Finals:

You don’t need a tour guide to spot the good news here. Korda’s plus-one error rate used to be outrageously high. It’s still higher than he like it to be, but it’s dramatically better, and getting it under control hasn’t cost him much on the other side of the ledger. As he puts the wrist injury fully behind him, there may be even more room for improvement.

The ceiling

I’ve focused on the serve–and Korda’s approach behind it–because that’s the side of his game that will determine how high he climbs. In his career at tour level, he has won 38% of return points, a figure that means he’ll break often enough to win matches when he serves well. Maintaining a 38% rate will get tougher as the quality of his opposition rises, but that may not be a problem: He has already excelled against top tier competition. As Alex Gruskin points out, he’s 18-21 against top-20 players, a record that indicates he’s already able to compete at that level, even if his results against the rest of the pack (and his health) aren’t consistent enough to support a corresponding ranking.

Korda may improve his return game, but if he is to crack the top ten and have a real shot at those two major titles, his serve will make the difference. In the last 52 weeks, he has won 66% of return points and held 83% of service games, numbers that place him among the top half of the top 50… but not much higher. The serve itself needs no improvement, as we’ve seen. The difference between Korda and someone like Fritz or Tsitsipas is what happens when the serve comes back. The 23-year-old is making progress, but he has more steps to take before he can reach the enormous potential that once seemed so assured.

* * *

Talking Tennis interview

I recently spoke with John Silk of Talking Tennis, and in a one-hour interview ,we covered all things Tennis Abstract: how to get the most out of the site, Elo ratings, common beliefs about tennis stats, and the Tennis 128. Watch it here:

* * *

Subscribe to the blog to receive each new post by email:

 

Jelena Ostapenko In the Hands of Fate

Also today: Deciding tiebreaks, a MCP milestone, and assorted links.

Jelena Ostapenko in 2023. Credit: Hameltion

If you’ve ever spent five minutes watching Jelena Ostapenko play tennis, you know she’s as aggressive as it gets. She swings for the fences and sometimes knocks them over. Get her on a hot streak, and opponents can only hope its ends before the handshake. When she’s off her game, spectators in the first few rows duck for cover.

What you might not realize is just how aggressive she is. A few years ago I tuned Lowell West’s Aggression Score metric so that the numbers fell in a range between 0 and 100. In theory, 0 is maximally passive; 100 is go-for-broke, all the time. Ostapenko’s career Aggression Score in rallies is 175.

This sort of extreme style lends itself to all sorts of narratives. She can beat anybody, any time, as she showed when she won the 2017 French Open as an unseeded player, and again last year when she upset Iga Swiatek at the US Open–her fourth win in as many matches against the Pole. That makes her a perennial dark horse pick at majors. Even though she hasn’t reached a semi-final since 2018, neither Iga nor Coco Gauff–who exited the Australian Open after an Ostapenko barrage last year–would like her find her in their section.

(Sorry Iga: Guess who you might face in the quarters!)

Hyper-aggressive players also appear to be works in progress. Especially early in Ostapenko’s career, commentators would talk about her stratospheric potential if she could only improve her footwork, or play a bit more “within herself.” That is, not quite so many winners, not quite so soon, more point construction, fewer unforced errors. But players rarely change much, and as they age, they are more likely to become more aggressive, not less. The Latvian is now 26 years old, beginning her ninth year on tour. What you see is what you get.

What you get, it turns out, is a lot of close matches. Ostapenko played 30 three-setters last year, including four in a row to reach the Birmingham final and another four straight to start the US Open. Alona’s apotheosis came at Indian Wells, when she faced fellow super-aggressor Petra Kvitova in the third round. Both women tallied exactly 75 points; Kvitova won, 0-6, 6-0, 6-4. Tennis ball fuzz could be seen floating over the desert for days afterward.

That particular scoreline was an oddity, but the margin of victory was not. Ostapenko’s tight matches are not a result of streakiness, flightiness, or anything of the sort. They are an unavoidable function of her game style. It’s almost impossible to hit lots of winners without also committing piles of unforced errors. (We’ll come back to that.) When you do both in such numbers, you personally account for a substantial majority of point outcomes. The winners and errors (very approximately) balance each other out, and unless your opponent does something remarkable–or remarkably bad–with the limited influence you leave her, you end up winning about half the points played.

No one takes the racket out of an opponent’s hand like Ostapenko does. Once the return is in play, the Latvian ends nearly two-thirds of points herself, with a winner or unforced error, or by forcing an error. No one else comes close. Drawing on Match Charting Project data, I’ve listed the active players who end the most rallies:

Player                 RallyEnd%  
Jelena Ostapenko           65.9%  
Petra Kvitova              61.6%  
Madison Keys               60.8%  
Liudmila Samsonova         60.0%  
Camila Giorgi              59.7%  
Aryna Sabalenka            59.7%  
Veronika Kudermetova       57.5%  
Danielle Collins           57.5%  
Ekaterina Alexandrova      57.2%  
Ons Jabeur                 56.8%  
Peyton Stearns             56.5%  
Caroline Garcia            56.2%  
Naomi Osaka                56.2%  
Varvara Gracheva           55.0%  
Iga Swiatek                55.0%

Here’s another way to look at Alona’s extreme position on this list. The only other woman to grade out so far from 50% is Madison Brengle, who ends fewer than 34% of rallies. Ostapenko’s power turns the rest of the tour into Brengle.

Give and take

Ending even 57% of points on your own racket requires a lot of big swings. When you aim for a line, you might feel confidence about your chances, but you are taking a risk. A few players, like Swiatek, can generate winners without paying the unforced-error penalty, but that takes an unusual combination of patience and power that most players do not possess.

The 66% of points that Ostapenko ends on her own racket divides into roughly 37% winners (and forced errors) and 29% unforced errors. That’s worse than Aryna Sabalenka, who hits nearly as many winners with only a 23% error rate, but compared to the tour as a whole, the ratio is a solid one. For every unforced error she commits, she ends 1.25 points in her favor. Average among players represented in the Match Charting Project is 1.16, and the true mean is probably lower than that, since the MCP is more heavily weighted toward the best players.

The ratio varies among players, but there is a fairly strong relationship. Here are the winner/forced error and unforced error rates–each as a percentage of all points where the return came back in play–for 140 current and recent players:

The correlation between the two rates (r2 = 0.3) would be even stronger if it weren’t for net-rushers like Tatjana Maria–and to some extent Leylah Fernandez–who force their passive opponents into more aggression than they would otherwise produce.

As Sabalenka shows, it’s possible to seize as many points as Ostapenko does without giving quite so many away, but even that may be a mirage: Sabalenka racks up winners behind an overpowering serve that the Latvian can’t match. If the plot above is any indication, it would be difficult to bring her error rate down without also sacrificing some winners, not to mention the élan that she has ridden to seven tour-level titles.

So we’re left with something of a paradox. A hyper-aggressive player has more control over her fate than her peers do, but that control comes at a cost of a towering error rate, which keeps matches close. One result is a week like this one in Adelaide, where Ostapenko has reached the final by slipping through perilously tight battles with Sorana Cirstea (51.7% of points won) and Caroline Garcia (50.2%). Both matches could’ve gone the other way, something that is true so often when the Latvian steps on court. My tactical advice for Daria Kasatkina in tomorrow’s final: Cross your fingers.

* * *

Deciding-set tiebreak records

AbsurDB asks:

[A]m I right that Hurkacz’s 15 deciding sets going into tie-breaks in one calendar year is a historical record in ATP (10 such tie-breaks won is also probably a record?)?

Indeed, both are records. According to my data, the previous records came from Ivo Karlovic’s 2007 season, when he reached 11 deciding-set tiebreaks, winning eight of them. Here are all the player-seasons with nine or more.

Player              Season  Dec TB  Record  
Hubert Hurkacz        2023      15    10-5  
Ivo Karlovic          2007      11     8-3  
John Isner            2011      11     4-7  
John Isner            2018      11     6-5  
Ivo Karlovic          2014      10     7-3  
John Isner            2017      10     5-5  
Kevin Anderson        2018      10     6-4  
Mark Philippoussis    2000       9     5-4  
Marat Safin           2000       9     5-4  
Ivan Ljubicic         2002       9     2-7  
Ivan Ljubicic         2007       9     8-1  
Ivo Karlovic          2008       9     5-4  
Sam Querrey           2018       9     1-8  
Borna Coric           2019       9     6-3  
Hubert Hurkacz        2022       9     3-6

(Yes, I checked before 2000, as well, but no one reached nine until Philippousis did so that year. The first player-season with eight deciding-set tiebreaks was Tom Gullikson’s, in 1984.)

* * *

MCP Milestones

Earlier this week, the Match Charting Project recorded its two-millionth point:

The milestone match was the Auckland second-rounder between Ben Shelton and Fabian Marozsan, which I charted as a warm-up for my article on Wednesday. We’re not resting on our laurels, of course: We’ve added another five matches (and 800 or so points) in the 48 hours since.

Also worth mentioning is another round number we reached in the offseason: 1,000 different ATP players. Apart from the name syou’d expect, it’s a healthy mix of lower-ranked active players and former tour regulars. #1,000 was Martin Jaite, via his 1987 Rome final against Mats Wilander. We’ve also now charted 800 different WTAers.

We stand about 200 charts away from 13,000 matches overall: approximately 7,000 men’s and 6,000 women’s. 2023 was our most productive year yet, and 2024 would be a great time to start contributing.

* * *

Assorted links

  • Earlier this week I appeared on Alex Gruskin’s Mini-Break Podcast, in which he got overexcited about a number of week one trends, and I tried to talk him down from all the ledges.
  • I wrote about how GPT4 helped me make Tennis Abstract’s new navbar, because you had to know I didn’t do it myself.
  • The tours have introduced a new policy on late matches. I’m underwhelmed: There are an awful lot of exceptions, and there’s no acknowledgement of the underlying problem of longer and longer matches.
  • Two student projects worth a look: Pramukh’s Evaluating Tennis Player Styles in Relation to Tour Averages, based on MCP data, and Amrit’s Aces over Expected model.
  • If you can’t wait until Sunday for grand slam tennis, here’s the Clijsters-Henin 2003 US Open final.

* * *

Subscribe to the blog to receive each new post by email:

 

Dangerous Exponents: A Covid-19 Podcast

Pardon the non-tennis interruption!

Carl Bialik and I–the duo that has brought you the Tennis Abstract podcast–are five episodes into a new show, Dangerous Exponents: A Covid-19 Podcast. We’re attempting to bring our usual analytical approach to issues related to the pandemic, while acknowledging that we’re not doctors, epidemiologists, or anything except for inquiring minds with a penchant for research and some skill at separating valuable research from the rest.

Our most recent episode is on herd immunity: Is there really such a thing, how do we get there, and how will we know we have? Previously, we covered the question of holiday gatherings; the role of exponents R0, Rt, and K; the latest on vaccines; and the trade-offs involved in keeping schools open. Each installment runs approximately 45 minutes.

You can find the episodes and subscribe in all the usual spots, such as iTunes and player.fm.

As we develop this new podcast, we’d love your feedback. If you’ve listened to one of our previous episodes, or after you listen to the new one, please take a moment and answer a few questions to help us hone our efforts.

Thanks for listening!

Economist: Novak Djokovic wins the most thrilling men’s tennis match ever

My latest article at The Economist’s Game Theory blog delves into one way in which the Wimbledon men’s championship match was the most exciting major final of the last four decades:

Sunday’s final registered an [Excitement Index] of 7.5%. Not only was that the highest of the tournament, but it tops every men’s grand-slam final of the last four decades (see chart). (A handful of women’s finals, which are best of three sets, score higher, because the high-leverage deciding set accounts for a larger fraction of the match.) The Wimbledon decider in 1980 between John McEnroe and Bjorn Borg—thrilling enough to spawn films and re-enactments—is next, at 7%. Another clash often dubbed the most thrilling of all time, the Wimbledon final in 2008 between Mr Federer and Mr Nadal, ranks third, at 6.9%.

Read the whole thing.

Economist: Cori Gauff announces herself at Wimbledon

At the Economist’s Game Theory blog, I wrote about Cori Gauff’s historic upset of Venus Williams:

IT IS hard to avoid the impression that the tennis world has witnessed a changing of the guard. On July 1st , the opening day of the 2019 Championships at Wimbledon, Cori Gauff, a 15-year-old American prospect, upset the five-times champion Venus Williams in straight sets. Ms Williams, aged 39, was not the highest-ranked player to fall on the first day of the tournament; that honour belonged to the reigning US Open champion, Naomi Osaka, the second seed. But no first-round winner has garnered more attention than Ms Gauff, whose youth causes her to establish new records every time she steps on court.

Read the whole thing.

Australian Open Coverage at The Economist

I wrote three pieces for the Economist’s Game Theory blog in the last week. The most recent was on Novak Djokovic, who has been dominant on hard courts, but whose few hiccups have come mostly against young players:

Mr Medvedev [followed] a path blazed by Mr Tsitsipas. The Greek prospect allowed Mr Djokovic to hit backhands at a typical 46% clip. But by hitting harder, riskier shots to that side of his opponent, he took Mr Djokovic’s down-the-line weapon out of the game. Mr Djokovic typically sends about one-seventh of his backhands up the line, but against Mr Tsitsipas last summer, that number was cut in half, and Mr Djokovic failed to record a single winner in that direction. In the Melbourne final, Mr Nadal allowed the world’s top-ranked player far more freedom: Mr Djokovic hit one in five of his backhands down the line, and a quarter of those shots ended the point in his favour. Only once has Mr Nadal held his rival’s down-the-line rate below 10%: the 2013 US Open final, the last time the Spaniard got the better of one of their hard-court duels.

After the women’s final, I looked at Naomi Osaka’s accomplishments in comparison to other players in history who were so much younger than tour average. She fares very well by that measure:

Few women have achieved as much as Ms Osaka while being so much younger than tour members as a group. The average age of the top 50 is about 27, nearly six years older than the back-to-back major winner. Only four other players since 1985 have won majors while they were at least 5.5 years younger than the mean of their peers: Ms Williams, Martina Hingis, Maria Sharapova, and Jelena Ostapenko, who won the 2017 French Open but failed to maintain her place in the top ten. None of those players matched Ms Osaka’s feat of following her first grand slam championship by winning another at the first opportunity, and only Ms Hingis claimed her second grand slam within a year of her first. It is too much to predict of any young player that she match the career accomplishments of Ms Williams, whose big-serving style Ms Osaka emulates. But even matching the more modest feats of Ms Hingis and Ms Sharapova, who are tied with five slams apiece, would rank her among the all-time greats.

Finally, I covered Karolina Pliskova’s monumental quarter-final comeback against Serena Williams. There are few, if any, precedents for such a momentum shift in the modern era:

Because collecting point-by-point data for tennis matches is a fairly modern practice, we cannot know for sure where this turnaround ranks in the sport’s long history. But among the 2,300-odd women’s contests that have been manually recorded by volunteers for the Match Charting Project, an online repository of tennis data, there is no example of a greater collapse. Most of the project’s sample is composed of high-profile matches from the 21st century, but there are also a handful of grand-slam duels of yore. Tennis’s most notorious choking incident—when Jana Novotna seemingly lost the ability to hit the ball against Steffi Graf in the 1993 Wimbledon final, after serving for game point at 4-1 in the deciding set—looks unremarkable when compared to Ms Williams’ downfall, with a peak win probability of 95.6%.

Go read them all:

Economist: Announcing his retirement, Andy Murray begins to get his due at last

For The Economist’s Game Theory blog, I wrote about Andy Murray’s legacy, as he approaches retirement:

[B]ecause of the quality of his competition, it is easy to underrate the career on which Mr Murray is calling time. Several of his tallies at the grand slams rank among the top ten in the modern era, including his 11 finals, 21 semi-finals, 30 quarter-finals, and 189 match wins. Only nine other men have spent more time in the top five of the world rankings. His three major titles sit much further down the all-time list, but the rest of the big four blocked him from at least six more. Comparing tennis players across eras is devilishly difficult, thanks to changes in technology, tactics, training regimens, and geographical breadth, but these raw totals probably underrate his standing among the all-time greats. He doesn’t belong in the top three, but few men other than his present-day rivals can unambiguously claim to have been stronger players.

Read the whole thing.

Economist: A fit Novak Djokovic could dominate tennis’s future

At The Economist’s Game Theory blog, I wrote about Novak Djokovic’s return to the top:

But Mr Djokovic is playing better in his 32nd year than Mr Federer did, and his more rounded game means that he can compete on all surfaces—even with Mr Nadal on clay—in a way that Mr Federer could not.  The odds are against the Serb reaching 20 majors, but another two or three seasons at the top could easily give him a final total of 17 or 18—enough to move Mr Djokovic out of his default position in third place. Fans of Roger and Rafa have long dominated the debate about who is the greatest male player of all time. But by the end of the decade, Mr Djokovic’s trophy cabinet could well be as bulging as those of his legendary rivals.

Read the whole thing.

Economist: An overzealous chair umpire overshadows Naomi Osaka’s impressive victory

I wrote something for the Economist Game Theory blog on the controversy in the US Open women’s final. Here’s one part I hope people remember:

The tennis world will probably be debating Mr Ramos’s calls until the next major rolls around in January. But one thing should not be in doubt: Ms Osaka didn’t need his help to earn her first grand slam title. Excluding the five penalty-determined points, she won 60 of the 110 points played, good for 64% on her own service and 45% on return. A ratio of that quality almost guarantees victory. In addition, all five of the points Ms Williams was docked would have been played on Ms Osaka’s serve. Given the level that the 20-year-old sustained, the first point penalty increased her chances of winning by less than half a percentage point, from 97.8% to 98.2%. Even if Ms Williams had been able to raise her level to equal her opponent’s, the impact would have been less than two percentage points. The game penalty was worth barely a full percentage point, boosting Ms Osaka’s probability of victory from 98.1% to 99.2%. By the time the New York crowd started booing, the match was virtually in the bag.

Read the whole thing.

Economist: The new serve clock in tennis appears to be backfiring

At the Economist Game Theory blog, I wrote about the early effects of the new serve clock. The outwardly stricter time policy didn’t speed up Rafael Nadal, nor did it cut down match times in general over its first two weeks:

The Toronto champion wasn’t the only player who slowed down once on the clock. At each of the completed tournaments where the serve clock has been used—Toronto, Montreal, San Jose, and Washington, D.C.—the average point took longer in 2018 than it did in 2017, without the clock. The differences varied from 0.3 seconds per point at the women’s event San Jose (an event that was held in nearby Stanford last year) to 2.0 seconds at the men’s competition in Washington.

Read the whole thing.