TA’s Match Charting Stats Leaderboard Discussion

TheFifthSet

TheFifthSet

G.O.A.T.
One of the hidden gems in tennis stat circles IMO:

















Fascinating to see which of our preexisting beliefs about player strengths/weaknesses/tendencies are validated and which ones are challenged. Not the most revealing sample of matches as the more high-profile ones will feature more heavily (and thus usually blunt a players stats) but still quite fascinating.
 
Last edited:
They only used a small sample for quite a few of those players which either makes their stats skewed either too high or too low.
 
Yeah the sample bias is important. I've had an absolute mofo (on a different site) tell me Roddick was inferior to Berrettini using TA collected stats as a justification, notwithstanding that the samples were very far from equally weighted (Roddick = mostly late round matches, hence a much higher percentage of losses; Berrettini = more early rounders, doesn't have many late rounds in big tournaments yet anyway). Still, the stats can be very informative if you approach them correctly with an inquisitive mind. I shuffled through the stats and the found the funniest thing: Federer actually slices his *edit: second serve* returns less than most, in particular clearly less than Djokovic and Murray (6% vs 10% vs 11%). So much for the slicebot repute. Nadal of course barely slices returns (2%), that was expected.
 
Last edited:
AnOctorokForDinner said:
Yeah the sample bias is important. I've had an absolute mofo (on a different site) tell me Roddick was inferior to Berrettini using TA collected stats as a justification, notwithstanding that the samples were very far from equally weighted (Roddick = mostly late round matches, hence a much higher percentage of losses; Berrettini = more early rounders, doesn't have many late rounds in big tournaments yet anyway). Still, the stats can be very informative if you approach them correctly with an inquisitive mind. I shuffled through the stats and the found the funniest thing: Federer actually slices his returns less than most, in particular clearly less than Djokovic and Murray (6% vs 10% vs 11%). So much for the slicebot repute. Nadal of course barely slices returns (2%), that was expected.
Click to expand...

That surprised me too. Naturally Fed is well ahead in rally slice %, but I also expected higher marks off the return.

Another cool one was Federer outdoing Nadal in % of points won once the return was put in play (Fed being a more aggressive returner penalizes Nadal here as the returns he puts in play will have more oomph on them and Nadal puts marginally more in play overall but that’s also largely offset by/a function of the clay skew).


Edit: would be remiss if I didn’t point out Fed’s extraordinarily high serve-in-play win %’s too…in a Top 30 dominated by clay courters and other slow court players, he’s sandwiched between Djoko and Borg for 6th place. Very few bots near the top. Definitively extinguishes any unironic case a casual can make about Fed being a servebot.
 
Last edited:
mike danny said:
Don't know how to feel about this one. Djokovic is way too low in some of these.
Click to expand...

A big part of that is the fact that many of the players atop the ‘board have a very low number of matches charted, so you’ll get some wonky results here and there. If you limit it to players w/100 matches, the Big 3 do about as well as you’d expect in most categories.
 
TheFifthSet said:
Definitely, and I think documenting w-l %’s in the samples and comparing them to career w-l %’s would go a long way in determining just how revelatory they are.
Click to expand...

Yea I would be curious to see how they compare and to see which ones are about the same in both.
 
Lots of really interesting stuff, thanks for sharing. A lot of things jive with what we'd expect, but some things that stood out to me:

-Nadal's rally aggression: find it hard to believe Nadal is less aggressive than the likes of Ferrer, Hewitt, Chang, Murray (guys usually seen as defensive retrievers/counterpunchers). Could Nadal's clay skew (around a half of his charted matches are on clay) be affecting this?

-Federer's second serve aggression: Fed is near the bottom of this with the likes of the clay courters who just spin it in. Doesn't seem to pass the smell test. The legendary Sampras second serve isn't as high as you'd expect either. How is this calculated/determined?

-Karlovic wins 64% of his 10+ shot rallies, which is not only top of the list but an outlier beating the likes of the next best by nearly 7%. Sample size is almost certainly affecting this since he most likely plays the least 10+ rallies and has the lowest average rally length, but still I found it interesting considering he's the most notorious servebot.

-Borg's BH potency per match is another complete outlier. According to the glossary, potency per match is affected by match and rally length, so I guess this is a testament to his often lauded precision and supreme fitness. Just didn't expect it to be by such a large margin. His BH potency per 100 is also near the top. Interesting that his BH metrics are so good but he isn't often brought up in the 2HBH GOAT convo with the usual suspects. Could sample size (only 30 something matches) be affecting this as well? Or maybe it's just a flawed stat?
 
Turing said:
Lots of really interesting stuff, thanks for sharing. A lot of things jive with what we'd expect, but some things that stood out to me:

-Nadal's rally aggression: find it hard to believe Nadal is less aggressive than the likes of Ferrer, Hewitt, Chang, Murray (guys usually seen as defensive retrievers/counterpunchers). Could Nadal's clay skew (around a half of his charted matches are on clay) be affecting this?

-Federer's second serve aggression: Fed is near the bottom of this with the likes of the clay courters who just spin it in. Doesn't seem to pass the smell test. The legendary Sampras second serve isn't as high as you'd expect either. How is this calculated/determined?

-Karlovic wins 64% of his 10+ shot rallies, which is not only top of the list but an outlier beating the likes of the next best by nearly 7%. Sample size is almost certainly affecting this since he most likely plays the least 10+ rallies and has the lowest average rally length, but still I found it interesting considering he's the most notorious servebot.

-Borg's BH potency per match is another complete outlier. According to the glossary, potency per match is affected by match and rally length, so I guess this is a testament to his often lauded precision and supreme fitness. Just didn't expect it to be by such a large margin. His BH potency per 100 is also near the top. Interesting that his BH metrics are so good but he isn't often brought up in the 2HBH GOAT convo with the usual suspects. Could sample size (only 30 something matches) be affecting this as well? Or maybe it's just a flawed stat?
Click to expand...

Karlovic atop the leaderboard was jarring, yeah, though I think the minuscule sample covers why (28/44 in rallies over 10 shots, across 35 matches). Still wacky though.
 
Reinforces my infallible observation that Goran's 1st serve is the single most unstoppable weapon of the OE, quite possibly ever. He clearly faced top players/returners more often than Ivo but their career %s of 1st serves unreturned are almost identical (58.9% vs. 59.0%), which validates @slice serve ace's unusual method of looking at aces/1st serves as the next best barometer to the nonstandard (read: rare) URS%. Also shows how otherworldly the Sampras serve was in his 18 Slam finals where he averaged 44.5% of URS - or 46.7% in his 14 Fs won, just a hair short of Karlovic's 48.0% in his 35 charted matches.

But as the forum's go-to URS% wonk I knew all this. A real revelation is Rafter's 26.8% of 2nd serves unreturned, 2nd only to Ivo's 31.6% but arguably more impressive cuz your opponents tend to hold more sway on your 2nd serves and also cuz Pat was a full-time S&Ver, not a part-timer a la (early) Pistol and Boris who stayed back on 2nds outside grass. That nasty kicker explains how Rafter made the top 10 in SGW% every year from 1996-2001 (plus '94).

For moi, though, it's the return stats that come in especially handy. It's become something of a gospel among you mugs that Agassi is overrated as a returner because he (presumably) got aced a lot, even though a closer look at the (corrected) stats shows how laughably clueless this canard is. But let's set those aside for now and dissect the TA stats, shall we? Here are the %s of returns put in play, with the %s of charted matches on clay and grass in brackets:

Agassi - 62.8% [9.6% (14/146), 6.8% (10/146)]
Becker - 64.4% [12.7% (13/102), 25.5% (26/102)]
Borg - 74.0% [40% (12/30), 30% (9/30)]
Chang - 69.2% [20.4% (11/54), 0%]
Connors - 66.3% [17.6% (6/34), 41.2% (14/34)]
Coria - 77.0% [68% (17/25), 0%]
Courier - 66.0% [26.2% (11/42), 4.8% (2/42)]
Davydenko - 73.5% [32.3% (10/31), 0%]
Djokovic - 70.9% [22.2% (88/396), 9.8% (39/396)]
Edberg - 65.9% [13.9% (22/158), 13.9% (22/158)]
Federer - 69.0% [16.6% (88/529), 17.2% (91/529)]
Ferrer - 70.7% [39.3% (42/107), 0.9% (1/107)]
Henman - 63.7% [7.1% (2/28), 28.6% (8/28)]
Hewitt - 67.5% [7.7% (9/117), 21.4% (25/117)]
Kafelnikov - 65.9% [27.9% (12/43), 0%]
Lendl - 65.3% [19.0% (16/84), 15.5% (13/84)]
McEnroe - 70.1% [17.1% (6/35), 31.4% (11/35)]
Medvedev, D. - 72.6% [8.3% (9/108), 6.5% (7/108)]
Murray - 74.4% [12.8% (29/227), 12.3% (28/227)]
Nadal - 74.1% [46.2% (184/398), 8.5% (34/398)]
Nalbandian - 67.8% [25.6% (10/39), 7.7% (3/39)]
Safin - 66.1% [19.5% (8/41), 4.9% (2/41)]
Sampras - 62.8% [4.6% (6/130), 14.6% (19/130)]
Schwartzman - 73.0% [56.5% (52/92), 2.2% (2/92)]
Wilander - 75.3% [35.7% (10/28), 14.3% (4/28)]

(All the big post-1990 names save Korda are carried over from my earlier analysis, plus Becker, Borg, Connors, Coria, Courier, Kafelnikov, Lendl, McEnroe, Safin, Sampras and Wilander.)

So Pistol is the only stud in recent years who failed to get more of his returns in than Dre, and considering how much Pete cruised on return it does seem that Agassi wasn't terribly effective against 1st serves. Surely the numbers don't lie, right?

Except for the tiny fact that Dre played under 10% of his charted matches on clay, vs. Coria's whopping 68%, Diego's 56.5%, Rafa's 46.2%, Borg's 40%, Ferru's 39.3%, Mats' 35.7% and Davy's 32.3% which no doubt skew their RiP%s. Now you know why I in my infinite wisdom included the surface breakdowns.

Of course this is where you jokers point to Muzz's crazy-good 74.4% and say only 12.8% of his charted matches were on dirt. Which is a fair counterpoint... except that Dre's 53.0% in RiP W% (that is, % of points won when the return is put in play) trails only Dustin Brown's 53.6%. So let's assume for argument's sake that these guys play an exact 100 points, successfully return them per their RiP% and win those return points per their RiP W%. This is what you'd get for each:

Agassi - 33.28% (62.8 x 53.0%)
Becker - 31.88% (64.4 x 49.5%)
Borg - 39.22% (74.0 x 53.0%)
Chang - 35.08% (69.2 x 50.7%)
Connors - 34.81% (66.3 x 52.5%)
Coria - 38.19% (77.0 x 49.6%)
Courier - 32.47% (66.0 x 49.2%)
Davydenko - 36.16% (73.5 x 49.2%)
Djokovic - 36.94% (70.9 x 52.1%)
Edberg - 32.62% (65.9 x 49.5%)
Federer - 34.91% (69.0 x 50.6%)
Ferrer - 34.29% (70.7 x 48.5%)
Henman - 29.68% (63.7 x 46.6%)
Hewitt - 34.16% (67.5 x 50.6%)
Kafelnikov - 32.49% (65.9 x 49.3%)
Lendl - 33.30% (65.3 x 51.0%)
McEnroe - 33.23% (70.1 x 47.4%)
Medvedev, D. - 35.57% (72.6 x 49.0%)
Murray - 35.94% (74.4 x 48.3%)
Nadal - 37.49% (74.1 x 50.6%)
Nalbandian - 35.26% (67.8 x 52.0%)
Safin - 32.98% (66.1 x 49.9%)
Sampras - 30.77% (62.8 x 49.0%)
Schwartzman - 37.16% (73.0 x 50.9%)
Wilander - 37.27% (75.3 x 49.5%)

Dre doesn't look so outclassed now, does he? Apart from the dirtmeisters the only guys who you could say surpass him are Chang, Jimbo, Fed, Rusty, Med, Muzz and Nalby, and all of 'em but Med played a bigger % of their charted matches on clay.

And surface indeed explains a lot of this gap. Here are (where available) these guys' %s of games won on hard, the most neutral/equalizing of surfaces, since '91 per the ATP, listed this time in descending order with their career rank:

1. Chang - 32.8%
2. Agassi - 32.3%
3. Djokovic - 32.1%
4. Murray - 32.1%
5. Edberg - 31.6%
6. McEnroe - 30.9%
7. Coria - 30.9%
10. Connors - 30.5%
11. Hewitt - 30.1%
12. Ferrer - 29.7%
13. Lendl - 29.4%
14. Schwartzman - 29.3%
15. Nadal - 29.2%
20. Davydenko - 28.5%
22. Nalbandian - 28.3%
33. Medvedev, D. - 27.4%
35. Kafelnikov - 27.3%
38. Federer - 27.2%
54. Courier - 26.6%
59. Henman - 26.4%
85. Wilander - 25.4%
105. Becker - 24.9%
107. Sampras - 24.8%
239. Safin - 21.6%

Normally for GW% you can expect about a 3-5% drop from PW%, and that's indeed what you see for Novak, Muzz, Rusty, Lendl, Kafelnikov, Henman and even Ferrer with their fairly even surface distribution while Coria, Diego, Rafa and to a lesser extent Davy and Nalby with their clay-heavy focus are subjected to a more dramatic free fall. Jim's and Marat's own nosedive can be explained by the fact that their small sample on TA is mostly culled from their best matches, while Jimbo's and Mats' come with the caveat that they were on their last legs in the '90s. And Boris', Pistol's and Fed's #s are further evidence that they go easy on their botting except in the biggest matches. (Frankly not sure where Med falls among these cohorts.)

At this point you should have also noticed that Dre, Edberg, Mac and even Chang suffer a minor drop at most, which may come as a surprise because it's undeniable that service stats have been creeping up - and vice versa for their return counterparts - due to the extra spin made possible by poly. So what explains this anomaly? Take a gander at the Net W%s here:


Out of these dozens of players only one won less than 60% of their net points overall. But how can this be? After all you geniuses remain absolutely adamant that S&V/net rushing is dead!

Of course real connoisseurs like moi always knew better and have been telling you jokers as much. And when you remember that even the very best players struggle to win 55% of their 2nd-serve return points you can see how big that 5+% advantage can be in the right hands.

So what do we do with this info? We know for a fact that attackers extraordinaire like Pete, Boris, Stich, Goran, Krajicek and to a lesser extent Henman S&Ved on many if not most of their 1st serves on hard and carpet, and Mac, Edberg and Rafter also on most of their 2nds. Plus even baseliners (let's not get too anal on the category) like Wilander, Courier, Mecir and Enqvist came in a lot on grass before the early aughts. That's why TA tells us Dre, Stefan, Mac and Chang won only a combined average of 33.28%, 32.62%, 33.23% and 35.08% of their return points respectively even though they stand 2nd, 5th, 6th and 1st among all players since '91 in GW% on hard - let's not forget Mac practically retired at the end of '92 - and also why Dre's 25.8% and perhaps Chang's 25.1% of career GS on grass is at least as good as Muzz's 26.5% and Novak's 26.4%, not to mention Stefan's spectacular 28.0% on the green stuff which may well be the most underappreciated stat of this most underappreciated returner.

And yet the idiots keep dissing the Agassi return as overrated even though just about every pro would laugh at the suggestion. As with the Sampras FH you gotta look beyond "analytics" to understand what made Dre such a nightmare on return.

Speaking of which I really don't give a crap about TA's "advanced" stats like RallyAgg or ReturnAgg which for the umpteenth time are nothing more than faux-objective ways to make you hack(er)s feel more knowledgeable than you really are. Do yourself a favor and stick to the basics instead.
 
As I am wont to do, points with which I agree will likely be glossed over because I'm a d!ck

NonP said:
Reinforces my infallible observation that Goran's 1st serve is the single most unstoppable weapon of the OE, quite possibly ever. He clearly faced top players/returners more often than Ivo but their career %s of 1st serves unreturned are almost identical (58.9% vs. 59.0%), which validates @slice serve ace's unusual method of looking at aces/1st serves as the next best barometer to the nonstandard (read: rare) URS%. Also shows how otherworldly the Sampras serve was in his 18 Slam finals where he averaged 44.5% of URS - or 46.7% in his 14 Fs won, just a hair short of Karlovic's 48.0% in his 35 charted matches.
Click to expand...

1. Surely some truth to the notion that the best ability is availability, no? Their first serve unreturned serve %'s are close, but there's a wide gap in first serve landed in -- 10.4% for their careers. This very obviously has much to do with racquet technology enabling more security off both deliveries...but, even for their days...Goran had an erratic first serve and Karlovic a reliable one...again, not just in absolute terms but also relative to their peers.

2. Number of top players/returners faced is specious, reason for this outlined here:

TheFifthSet said:
That’s sound in theory but Karlovic’s service stats against the top 20, 10 and 5 remain pretty stable.

His holding % against the top 10 in 80 career matches is 90.9%, not appreciably worse than his 92.1% career rate. In 37 matches against the top 5 its 90.5%.

If you’d like a larger sample, it’s 90.6% in 138 matches against the top 20.

Fed’s at 88.8% for his career but it’s 85.8%, 84.6% and 81.9% against the top 20/10/5.

The funny thing is that the better players beat Ivo with the most frequency not necessarily because they break his serve the most but just as much or more so because they have more stable service games of their own and limit the amount of freak breaks he notches. Fed’s an excellent example of this, he has broken Ivo 9% of the time in their 15 matches yet has won 14 of them.

Nobody can really do much to Karlovic’s serve. The opponent scarcely matters with regards to the service stats.
Click to expand...

Same applies to Goran. Neither men's %'s budge a whole lot against top players, both drop about 0.8-2.0% at most.
 
You must log in or register to reply here.

Similar threads

Rovesciarete
Lorenzo Musetti and the One Hand to Rule Them All by Jeff Sackmann
Replies
4
Views
650
Rovesciarete
Rovesciarete
Back
Top