The headline of this piece should end with a question mark. But that seemed churlish, so I saved my skepticism for the next couple thousand words. (Is it a deterrent to warn you of the word count in the first paragraph? Your call!)
Last week I wrote about the giant-killing game of 19-year-old Jakub Mensik. He kept winning, defeating Jack Draper, Arthur Fils, Taylor Fritz, and Novak Djokovic (among others) to claim the Miami title. His serve numbers boggle the mind. He cracked aces on at least 24% of points in every match before the final. Djokovic held him to a mere 18%.
While I generally steer clear of the word “clutch,” it was undeniably a clutch performance. The young Czech won all seven tiebreaks he played, including two each against Draper, Fritz, and Djokovic. He allowed Arthur Fils five points in their breaker, but none of the top-tenners managed more than four in a single tiebreak. Awesome as his serving was in general, it was even better at the tail end of those sets.
Mensik’s tiebreak record is almost as stunning as his ace numbers. Coming into Miami, his career tally at tour level was 23-13, a 64% winning percentage in a category where non-elite players tend to stick around 50%. Djokovic is the all-time leader at 66% (minimum 400 breakers), and only a handful of stars have posted career marks above 60%.
(I’m excluding tiebreaks from the NextGen Finals event last December. Mensik won just two of eight in Jeddah. The under-21 tourney is played with different rules and doesn’t award ranking points, so it seems logical to leave it out.)
Having run the table for the last two weeks, the Czech is up to 30-13, a 70% win rate. Mensik sure seems like a young master of the tiebreak. He’s cool under pressure, and he has the monster serve. Is he going to spend the next decade 7-6-ing his peers into oblivion?
The next 43
Let’s get some context. Mensik has played 43 tiebreaks, so here are the players (born 1975 or later) with the best records in their own first 43:
Player W-L Win% Pablo Cuevas 33-10 76.7% Novak Djokovic 32-11 74.4% Marcelo Rios 31-12 72.1% Lucas Pouille 30-13 69.8% Jakub Mensik 30-13 69.8% Sergiy Stakhovsky 29-14 67.4% Tommy Haas 29-14 67.4% Sebastien Grosjean 28-15 65.1% Marcos Baghdatis 28-15 65.1% Bernard Tomic 28-15 65.1% Milos Raonic 28-15 65.1% Botic van de Zandschulp 28-15 65.1% Alexei Popyrin 27-16 62.8% Kei Nishikori 27-16 62.8% Roberto Bautista Agut 27-16 62.8% Felix Auger-Aliassime 27-16 62.8% Lukas Lacko 27-16 62.8% Philipp Petzschner 27-16 62.8% Kristof Vliegen 27-16 62.8% Dominik Koepfer 27-16 62.8%
Not bad company, but something of a mixed bag. Mensik is tied for fourth behind an all-time great, a near-Hall of Famer, and a clay court stalwart. The rest of the list includes both top-tenners and journeymen.
How did these guys fare in their next 43 tiebreaks?
Player First 43 Next 43 Win% Pablo Cuevas 33-10 22-21 51.2% Novak Djokovic 32-11 33-10 76.7% Marcelo Rios 31-12 23-20 53.5% Lucas Pouille 30-13 22-21 51.2% Jakub Mensik 30-13 - - Sergiy Stakhovsky 29-14 19-24 44.2% Tommy Haas 29-14 20-23 46.5% Sebastien Grosjean 28-15 26-17 60.5% Marcos Baghdatis 28-15 25-18 58.1% Bernard Tomic 28-15 22-21 51.2% Milos Raonic 28-15 25-18 58.1% Botic van de Zandschulp 28-15 19-23 45.2% Alexei Popyrin 27-16 22-21 51.2% Kei Nishikori 27-16 25-18 58.1% Roberto Bautista Agut 27-16 25-18 58.1% Felix Auger-Aliassime 27-16 26-17 60.5% Lukas Lacko 27-16 19-24 44.2% Philipp Petzschner 27-16 21-22 48.8% Kristof Vliegen 27-16 19-21 47.5% Dominik Koepfer 27-16 9-14 39.1% TOTAL 570-290 422-371 53.2%
(Some of them didn’t make it, or haven’t yet made it, through another 43.)
These numbers are… less impressive. Djokovic was unstoppable in the early going; he has managed to win “only” 61% after that second batch of 43. Grosjean and Auger-Aliassime came close to matching their first 43s. But the rest of the group made a beeline for mediocrity. Even counting the three standouts, the group won only 53% of their follow-up tiebreaks. 53% is fine–it worked for Marin Cilic and Marat Safin–but if Mensik had won 53% of his career tiebreaks so far, we’d be celebrating a different champion in South Florida.
Breakcasting
Bigger picture, there is almost no correlation between a player’s record in their first 43 tiebreaks and their next 43, or their ensuing career.
Djokovic maintained his breathtaking record because he was extremely good at tennis, not because he had secret tiebreak mojo. Players who aren’t on the way to double-digit slams (and even some who are!) can’t count on winning two-thirds of their tiebreaks.
This doesn’t mean, however, that everybody trends toward 50% on the dot. Better-than-average pros win more points, both in and out of busters. I introduced a stat a few years ago called Tiebreaks Over Expectation (TBOE), which hinges on that notion of “expectation.” Take a player’s rate of serve and return points won in a given match, plug it into a tiebreak simulator, and you get the likelihood that he’ll be first to seven. No clutch, no wizardry, just the assumption that playes are about the same in breakers as they are the rest of the time.
Mensik’s idol Tomas Berdych is a good example. He won 225 of his 374 career breakers, a 54% success rate. While it’s a pedestrian number compared to Mensik’s, it’s solid! And it’s precisely what the formula expects. Run the exercise for each of the 374 breakers, and it predicts 225 wins. (224.86, in fact.) In other words, Berdych was exactly as good in the jeu decisif as he was in those matches as a whole.
TBOE makes for a better–if considerably less exciting–forecast of future tiebreak results. Here’s the top 20 again, with “expected” records for their first 43:
Player Actual Expected Pablo Cuevas 33-10 24-19 Novak Djokovic 32-11 23-20 Marcelo Rios 31-12 24-19 Jakub Mensik 30-13 22-21 Lucas Pouille 30-13 21-22 Tommy Haas 29-14 22-21 Sergiy Stakhovsky 29-14 21-22 Milos Raonic 28-15 24-19 Marcos Baghdatis 28-15 23-20 Bernard Tomic 28-15 22-21 Sebastien Grosjean 28-15 22-21 Botic Van De Zandschulp 28-15 22-21 Kei Nishikori 27-16 24-19 Felix Auger Aliassime 27-16 23-20 Kristof Vliegen 27-16 22-21 Philipp Petzschner 27-16 22-21 Alexei Popyrin 27-16 21-22 Roberto Bautista Agut 27-16 21-22 Dominik Koepfer 27-16 20-23 Lukas Lacko 27-16 20-23
An enormous amount of those all-time-best career starts come down to luck. Cuevas really did win those 33 tiebreaks. But his performance in those matches didn’t merit so many 7-6’s in his favor. Fortune caught up with him, and he won less than half of the 170 breakers he played over the remainder of his career.
Mensik is no exception. Here are his serve and return win rates against his opponents in Miami, along with the resulting probability that he would win a tiebreak in each match:
Opponent SPW RPW p(TB Win) Draper 75.0% 33.7% 65.1% Fils 73.6% 42.0% 74.7% Fritz 74.3% 30.7% 58.9% Djokovic 69.2% 29.7% 48.2%
Replay the tournament, and some of those seven breakers will almost certainly not go the same way, no matter how calm the Czech is under pressure. The odds of converting all seven tiebreaks against this competition is about 2.5%.
Yeah, but!
None of this is meant to take away from what Mensik has accomplished. He is 8-5 against top-tenners. He just won a Masters 1000. He powered through the last fortnight with one unhittable serve after another, especially when it mattered.
What the numbers do say is that he’s unlikely to keep it up.
Some of you, surely, want to argue that the 19-year-old will defy the odds. He has a huge serve, and big servers do better in tiebreaks, right? He’s clutch, and that doesn’t just go away–it isn’t like we’re rolling dice to get the outcomes of these tiebreaks.
These arguments are appealing, in part because we’ve heard them from players and commentators ever since Jimmy Van Alen figured out a slick new way to end sets. But they are not true.
There’s almost no relationship between a player’s serving ability and his performance in tiebreaks. Of course, the better the serve, the stronger the player, and the better the results, whether we’re talking games or sets or breakers. But when it comes to winning more tiebreaks than expected–getting from Mensik’s expected 22 of 43 to his actual 30 of 43–serving big doesn’t help. (There’s one twist to that rule, and I’ll get there in a minute.)
Next: If clutch ability exists, it is remarkably fickle. Players do clutch things, like reel off seven tiebreaks against higher-ranked players on a big stage. But that doesn’t mean they’ll be clutch the following week, month, or season. We saw that in the “first 43” versus “next 43” comparisons. Every one of those players with a hot start looked clutch. With the exception of Djokovic and maybe a couple of others, their ability to raise their game in tiebreaks disappeared. Even Novak steadily drifted back to earth.
Big Jake and Big John
There are a few big servers who have outperformed tiebreak expectations. The poster boy is John Isner. By my TBOE metric, he won 16% more tiebreaks than he “should” have, flipping the result of about 70 breakers over the course of his career. No one with at least 150 career tiebreaks has consistently stepped up their game so much.
We can learn a lot from that one example. The main thing is that +16% is the best anyone can reasonably hope for. What if Mensik really is the chosen one of the tiebreak? We’ve already seen a chosen one of the tiebreak, and we know how much his magical skills impacted his results.
In his first 43 tiebreaks, Mensik has beaten expectations by 36%, more than double Isner’s career mark. If we assume all of the Czech’s superb tiebreak performance is a mirage, we’d expect him to go 22-21. If we grant him an Isnerian level of clutch performance, he moves up to 25-18, or maybe 26-17. At that rate, he’d be climbing the ranking table, but this week’s storyline would be Djokovic and his 100th career title instead of the teen sensation.
Even +16% is a big ask. Next on the career overperformance list is Nick Kyrgios, at a slightly lower +16%. (I promise it’s not all big servers, and plenty of big servers underperform as well. Ivo Karlovic, Hubert Hurkacz, and Sam Querrey are a few counterexamples.) One thing we know about Kyrgios is that he–sometimes quite blatantly–saves his energy for key points. When he does that, of course he’s going to get better results on the key points, including tiebreaks.
Did Isner do the same? Maybe not as demonstratively, but I can’t imagine he tried very hard when returning at 30-0 or 40-0. Some of the 16% overperformance can likely to be attributed to that. He played better in breakers because he cared more about every point. It’s a sound tactic for a certain type of player; it’s just not as common among those with more modest serves.
Mensik, to his credit, is not so one-dimensional. I don’t get the sense he’s conserving energy for tiebreaks, tanking return games, or anything like that. If that’s true, his overall serve and return stats are a good indicator of his true ability level, unlike Kyrgios’s, which are deflated by his occasional apathy. Accurate serve and return stats make it even less likely that Mensik, or anyone else, can improve on them so much at the end of sets.
The sustain
There’s another reason why Isner–and Mensik–might outperform expectations. It’s not that they get better in tiebreaks, it’s that their serves are so good that they don’t get worse.
I looked at tiebreak tactics a few years ago and discovered that servers, on average, become more conservative. Rallies stretch out. Here were the key findings:
If every player reacted to the pressure in the same way, this would be bad news for someone like Mensik. The longer the rally, the worse he fares. But at least through his first 43 tiebreaks, he has defied the trend. While I don’t have tiebreaks split out for every match, the records I do have suggest he’s winning a whopping 78% of service points in breakers, compared to 65% overall.
His performances in Miami might have been even better. Facing Djokovic in the final, fewer than half of his serves came back. Against Draper, he served ten points in the two tiebreaks. Only two serves came back.
I don’t think it’s realistic to expect Mensik to continue to be the greatest server in the history of the sport every time the score reaches six-all. On the other hand, the best servers have more options than their less-fortunate peers. Roger Federer is another guy who outperformed tiebreak expectations. He, like Isner, could hit a first serve at 90% strength that still left returners flat-footed. Mensik may be in the same category. He can pile up aces without taking on an unacceptable amount of risk.
It’s reigning Mensik
The Miami title moved the Czech up to 24th in the ATP rankings. My Elo ratings–which don’t make any adjustment for whether tiebreak records are sustainable–put him in 13th place.
Assuming he comes back to earth and loses closer to half of his tiebreaks, can he sustain that?
The answer just might be yes–at least for the ATP ranking. Over the last 52 weeks, Mensik has won 50.5% of his points. Here are the ATP top-50 guys who also have total-points won rates in the same range:
Rank Player TPW 14 Ben Shelton 50.9% 49 Jan Lennard Struff 50.9% 25 Sebastian Korda 50.8% 50 Zizou Bergs 50.7% 21 Tomas Machac 50.7% 17 Frances Tiafoe 50.6% 29 Jiri Lehecka 50.6% 44 TM Etcheverry 50.5% 12 Holger Rune 50.5% 24 Jakub Mensik 50.5% 34 Alex Michelsen 50.4% 42 Gael Monfils 50.3% 48 Miomir Kecmanovic 50.3% 43 Nuno Borges 50.3% 40 A Davidovich Fokina 50.2%
The Czech is basically tied with top-20 players Holger Rune and Frances Tiafoe, and he isn’t far behind another in Ben Shelton. Also of note, John Isner finished 2018 in the top ten with a TPW% of just 51.1%.
If anything, that comparison understates Mensik’s level. Thanks to his knack for upsetting seeds and going deep in draws, the 19-year-old has faced tougher competition than almost anyone else. Here are the top-50 players who have played the most difficult schedules, as measured by median opponent rank:
Rk Player MdOppRk 1 Jannik Sinner 27.0 3 Carlos Alcaraz 28.0 7 Jack Draper 30.0 4 Taylor Fritz 30.5 5 Novak Djokovic 35.0 24 Jakub Mensik 35.5 42 Gael Monfils 36.0 11 Daniil Medvedev 36.0 2 Alexander Zverev 37.0 28 Alexei Popyrin 38.0
Mensik isn’t just winning points at a solid rate, he’s doing so against top-tier competition. Alexander Zverev, Casper Ruud, and Stefanos Tsitsipas have all faced weaker average opponents than the Czech has.
One takeaway here is that ATP rankings are awfully noisy. They are so dependent on context that it’s virtually impossible to say what a player’s ranking “should” be. Had Mensik’s tiebreak streak ended in the Draper match, he’d still be about the same player, but his ranking would be 20 places lower. In that plausible counterfactual, he’d be underrated.
As it is, the 19-year-old doesn’t need a 30-13 tiebreak record to be seen as an outstanding player on the rise. He’s a credible top-30 player–maybe more–even without it. Which is good, because he’s not going to keep winning 70% of his breakers. Isner earned his one year-end top-ten finish with a season tiebreak record of just 53%. With a better second serve, Mensik can do the same.
* * *
Subscribe to the blog to receive each new post by email:
Heavy Topspin