Peter Bodo's TennisWorld - Headbanging with Rosangel
Home       About Peter Bodo       Contact        RSS       Follow on Twitter Categories       Archive
Headbanging with Rosangel 04/16/2007 - 5:10 PM

Rosangel, one of our most devoted and astute comment posters, had an intriguing idea after she watched Novak Djokovic's impressive performance at the two recent Masters Series events (Indian Wells and Miami). Knowing how tough it is to play your way to the top and stay there, she decided to do some serious number crunching (that skill appears to be emerging as a real TennisWorld avocation!) on the head-to-head front. In fact, she charted the H2Hs of the Top 12 against the Top 50 players, and further broke things down (not published) by gathering the H2H of the Top 50 vs. the current Top 5, Top 10 and Top 20. Below is her chart of the Top 12 and their H2H stats in a variety of key areas. It is accurate as of April 9.  In addition, here are some of Rosangel's comments on what she was attempting to do, and what she discovered:

- I have also looked at certain players we at TennisWorld have focused on recently - young guys, late-career winners and the former world number ones. I may post some of those stats in the Comments if you all are interested.

Rosia1-  Among the Top 50 players, Roger Federer, Rafael Nadal, Andy Murray, David Nalbandian, Guillermo Canas, Tim Henman, Gaston Gaudio and LLeyton Hewitt all have a 50%  or better winning percentage against Top 5 opposition.

- Andy Roddick has only a 40% winning percentage against the Top 5, although he performs at a better than 50% rate against the Top 10. The other Andy (Murray) has a 53% WP against the Top 5 but against the Top 10 his WP drops to 41%

-Two off-the-radar players with winning records against the Top 10: Guillermo Coria and recently retired Andre Agassi.

-  Outside those elites who have a winning record against Top 5 players, only three players - Djokovic, Marat Safin and Ivan Ljubicic  - have a 50% or better WP against the Top 20.

Rosia2- The table includes players' ages, and the year they turned pro. Note that Nadal turned pro in 2001, but with one exception, was not playing on the main tour until 2003.

-  Andy Murray, on a shorter career (time-wise), has built a stronger record against the top dogs than his friend and rival Djokovic. But Djokovic's numbers vs. the Top 10 will benefit on April 16th, when Murray enters the Top 10.

Here are some other oddities for you to savor:

1) Wawrinka is the highest-ranked (No. 36) never to have had a win against a Top 5 player

2) Benjamin Becker gets a similar "honor" if you expand the pool of opponents to the Top 10. In fact, Becker has the peculiar distinction of having only one win against a Top 20 player - No. 20 Jarkko Nieminen.

3) Dominik Hrbaty has the unique honour of being the only Top 50 player with a winning H2H against both Federer (2-0) and Nadal (3-1). But he has losing a losing H2H against the rest of the top 5, and the other three members of the Top 10 whom he has played.

I also started thinking about those top players, personified by Human Litmus Test  Kolya The Obscure (Nikolay Davydenko), who have losing H2Hs against the Top 5, Top 10 or Top 20. My investigations were sparked by the work of Sam and Andrew, and the debate they raised over Roddick's place in the scheme of things.

So I looked at top players' records against the current Top 50, to see if they beat the players who they ought to be beating. In this table the Complete Dominance column tracks the number of players who have not been able to post a single win over the player in question.

H2hs2aFederer's numbers versus the Top 50 are astonishing; Nadal's nearly as impressive - especially when you consider his age. But Roddick is also at least 10% ahead of the other contenders. He is well ahead of Kolya, who has a losing H2H (75-76) against the current Top 50 - thus proving, as if we didn't already know it, that his ranking is perhaps artificially inflated by the "Best-18" system used for ranking (that is,  the computer uses a player's 18 best results, no matter how many tournaments he plays). The only other players who have losing records against the Top 50 are Tommy Robredo, Murray and Djokovic (for the latter two, age undoubtedly has something to do with this).

Federer is Tier 1, Nadal is Tier 2 and Roddick is Tier 3? I looked at their career winning percentages to see if they shed light on this issue. Of course, this is not authoritative for assessing a career-in-progress, but it's interesting nonetheless. Federer has a career winning percentage  of 79.6%, and Nadal sits at  78.1% - closer than I expected. Roddick has won 76.1%,  while Hewitt has slipped to 76.0%. The remainder of the Top 12 have winning percentages between 60% and 68% - with the outstanding exception of Davydenko, who's at 56.8%.

Looking at the game's recent greats, Pete Sampras had a 77.4% winning percentage, and Agassi 76.0% (that's weaker than Roddick's). The highest winning percentage of all was Bjorn Borg's 82.3%. He is closely followed by Jimmy Connors at  82.0%  - truly astonishing, given that Connors had a long career and ended up playing twice as many matches as Borg. McEnroe came in third at 81.7%. From there, the next players are Federer and Nadal. With the exception of Hewitt, our still-active former Number Ones - Moya, Ferrero and Safin - have career winning percentages in the mid-60s.

Finally, I couldn't notice how thin the actual H2H records of today are. The most substantial current "rivalry"  is Federer-Hewitt at 11-7 (favoring Federer),  followed by Federer-Nalbandian at 8-6 (again, for Federer). I don't know that 13-1 Federer-Roddick counts as a rivalry, or even 11-3 (Federer owns Ljubicic). Tommy Haas and Hrbaty, though, have something interesting going, with Haas leading 8-6.

If memory serves, Rod Laver and Ken Rosewall played so often that professed to not know their H2H, though they seemed to agree that Laver had won more. More recently, Sampras led Agassi by a healthy margin, 20-14, while Jim Courier-Michael Change finished 12-12. Stefan Edberg nosed out Chang at 12-9 and Borg bested Connors, 13-8.  Mats Wilander prevailed over Edberg, 11-9. Given their wildly different styles, that rivalry had it all. 

So Federer-Hewitt is a pretty hefty rivalry spoiled by one inconvenient fact: Federer won their last 9 matches. No former number ones, it seems, are able to provide a big or close challenge to the current number one.

-  Rosangel

(Ed. note: marvelous work, Rosia. Assist to Steggy for helping get these tables into viewable form - PB).

Feed You can follow this conversation by subscribing to the comment feed for this post.
1 2      >>

Posted by L. Rubin 04/16/2007 at 05:27 PM

Impressive stuff, rosangerl.

Posted by Zola 04/16/2007 at 05:30 PM

truely impressive. I love these number crunching articles. gives a very scientific view of tennis. Of course was not surprised by Fed's numbers, but I am glad Nadal is doing so well at such a young age. Hope he continues that way.

This is good especially after the previous , equally interesting , discussion of tier 1, tier 2 , etc. looking at these numbers I guess Roddick deserves to be in the same league as Nadal and Federer.

Canas's percentage wins against top ten is the same as Nadal's! see if he can keep it that way. he should win Federer more!

Posted by Sam 04/16/2007 at 05:31 PM

Ros: Great work - I'll comment after taking a closer look tonight.

Posted by Florida_Paul 04/16/2007 at 05:31 PM

Interesting stats. I'd like to add that the Agassi-Sampras head-to-head has some subtleties that may be obscured by the final 20-14 margin. For example: after their first 16 matches, the score was 8 wins each. Then, from 1995 US Open through the end of 1999 Sampras won most of their matches (9 of 12). Their last 6 matches - all in the current century - were split evenly.
The Grand Slam head-to-head is 6-3 in favor of Sampras; the 3 Agassi wins were at Paris and Melbourne.

Posted by Zola 04/16/2007 at 05:33 PM

well, some might not put Roddick with Nadal and Fed, but his numbers are ahead of most in top 10. interesting to see how these numbers will be at the end of the year.

Posted by Arthur 04/16/2007 at 05:37 PM

Missed Ivan Lendl and Boris Becker (as well as Guillermo Vilas) in the lifetime winning percentage paragraph. Lendl's winning percentage was 81.8. Becker's was 76.9 and Guillermo Vilas's was 76.6. Thanks for the great research!! Love those statistics nad truly, Connors's lifetime winning percentage is absolutely awesome, especially considering he played into his 40s. Had he retired even three or four years earlier that percentage would be considerably higher.

Posted by Danielle 04/16/2007 at 05:41 PM

Damn Ros - that's some seriously impressive number crunching there. Where's Lucy with your Warrier Badge?

Posted by codepoke 04/16/2007 at 05:45 PM

This is cool! Thanks.

Posted by evie 04/16/2007 at 05:58 PM

Hey, Rosangel,

I wonder if you do queries via sql or some such for relational database. That is the best way you can do all this numbers crunching you're doing. Great work and Nice analysis!

Posted by Ray Stonada 04/16/2007 at 05:58 PM

Great work, Rosia. Millions of things to comment on, but here are a couple:

I think you've unearthed something really important: the top players used to play each other much often. That is something tennis needs to concentrate on. We've got to see the premier matchup in tennis more than once in eight months (Federer and Nadal, whose Shanghai match was their only one since Wimbledon).

One other point: how do you make of the disparity in career winning percentages between today's players and Borg/Connors/McEnroe? What if we use the winning percentage of the top players to gauge the depth of the field? Does it tell us the tour is deeper now, that the number fifty player is better than the number fifty of 1980? And why is Roddick's number almost as high as Sampras', and higher than Agassi's? Was the tour at its peak depth in 1995?

Posted by RedClaw 04/16/2007 at 06:00 PM

As per the discussion from before, Roddick is arguably not in the same category as Nadal and Federer - but at the same time, he's clearly ahead of everyone else on tour except for those two!

Posted by GSte 04/16/2007 at 06:03 PM

Thanks for all the stats Ros!

Your work really hurts Davydenko and Robredo in terms of how "top playerish" they are!

I think what accounts for why so many of the current top 10 players have low % in the areas Ros highlighted is because outside of Federer and Nadal, most of them are on a very level playing field. Additionally, there's been so much shuffling among the top player rankings that not many long-standing rivalries have gotten off the ground. Yes, Roddick has a 70 winning % against top 50, but his numbers don't stand out more than others against the top 5 and top 10. Really, Roddick has shown that he's vulnerable to any top 20 player.

I think that after this year, we'll get a better feel for how the young guns can be rated.

Posted by GSte 04/16/2007 at 06:05 PM

Also interesting to note, Hewitt's numbers are better than Roddick's, which shows that it's really injury that's been keeping Hewitt out of the top 5.

Posted by Ray Stonada 04/16/2007 at 06:07 PM


Agassi and Sampras: Sampras won all their meetings at the U.S. Open and Wimbledon, Agassi all their meetings at Roland Garros and the Australian. In finals, the number was 9-7 for Sampras.

Did Borg's percentage edge out the others because of his retiring so young, before the gradual slide that every player goes through? This could be tested by checking McEnroe's in 1985, I suppose.

And Dominik Hrbaty should retire right now.

Posted by media 04/16/2007 at 06:10 PM

Federer and Nadal are both Tier 1 players

Roddick is Tier 2 alonside Nalbandian,Ljubicic and Davydenko(except at Wimbledon) because their results at the slams have been great and on all the surfaces
Murray at this stage looks better than Blake I'm sorry to point to you Bodo;1 quarters on 1 surface and once is nothing.
Murray plays well on all surfaces
Haas,Djokovic,Gonzalez and Robredo can play well on Hard;Clay and grass and RA
Blake plays well on Hard court alone and is a Tier 3 player no matter how much you like him(the man gets beaten by clay courters on Hard Courts at home man) and by the time Hamburg arrives he will be out of the top 10 and will make way for Ancic,Berdych,Gasquet,Ferrer and Youzhny to do better than him.

Posted by media 04/16/2007 at 06:11 PM

and yes I agree Hewitt should be ahead of Roddick;too bad injuries have held him back.
better player.

Posted by Jane 04/16/2007 at 06:14 PM

I love things like this. The only caveat I'd make is that the either the table for H2H against the top 10 is slightly wrong, or the text is slightly wrong. You say that when you consider winning percentage against the top 10, the list of players with a percentage at or greater than 50 expands to include Andy Murray, Nalbandian, and others like Hewitt and Gaudio. But if I'm reading the table correctly, that should be Andy Roddick (not Murray), and Nalbs shouldn't be on the list (his percentage is 46, 23 of 50). But I'm not trying to be a pain -- I think this stuff is super cool, so thank you for it. Much food for thought!

Posted by jb 04/16/2007 at 06:17 PM

Ros - this is really interesting to look at the stats on the players. Interesting to see roddick's stats up there also... This was a boatload of work to pull together; but I'd love to see it updated quarterly - or at least a couple times a year to see if the trends will change on some of the youngsters.

And props to Roddick - as both Sam and now Ros's work shows he really belongs at the top.

Posted by marieJ 04/16/2007 at 06:17 PM

rosangel, impressive work !!
roger's numbers are close to perfect...
for me roger and rafa broke the tier I category were all the top players used to be and started the super tier I only for the 2 of them ;)
roddick is very close i must admit, the only thing that keeps him from being there is the H2H vs roger, if he beats rafa or roger this year, he will earn the super tier status for me.
since rafa made his brilliant season in 2005 i really though we were having 2 guys for only one spot at numnber 1, they both deserve it.

Posted by Ray Stonada 04/16/2007 at 06:22 PM

MarieJ, I echo you 100% on the "Super Tier." Great name!

Posted by Rosangel 04/16/2007 at 06:25 PM

Jane: The table is correct. Checked and doubled-checked. It's a minor editorial misunderstanding, which I hope can be corrected - have sent an e-mail to the Powers That Be! Obviously the players with a 50% or better winning record against the top 5 include more than Federer and Nadal.

Posted by Rosangel 04/16/2007 at 06:34 PM

By the way, the winning career percentages for the rest of our our current top 12 are:

Gonzalez 64.7%
Robredo 61.7%
Djokovic 67.8%
Ljubicic 60.7%
Haas 65.3%
Blake 60.7%
Murray 66.1%
Nalbandian 66.2%.

Posted by highpockets 04/16/2007 at 06:38 PM


If numbers and charts can be beautiful ... all I can say is WOW! It will take me weeks to digest it, but what an effort!

Steggy, kudos to you for making it all fit perfectly!

Posted by Rosangel 04/16/2007 at 06:43 PM

Career winning percentages for some others beyond the top 12:
Berdych 58.4%
Ancic 59.8%
Gasquet 58.9%
Baghdatis 61.7%
Youzhny 55.2%
Ferrer 55.9%
Henman 65.0%
Coria 67.1%
Safin 63.8%
Monfils 50.5%.

Some other former greats, near-greats or just very good players:

Wilander 72.0%
Rafter 65.2%
Chang 68.0%
Moya 65.5%
Kuerten 65.2%
Gerulaitis 70.2%
Ferrero 66.3%

Posted by Andrew 04/16/2007 at 06:46 PM

Rosangel. Wow, this is yummy stuff.

The player that jumps right out of the pack is Chilly Willy Canas. Not having the decimal places, he shows up as no 2 vs top 5, 2= vs top 10, and 2= vs top 20.

Dude was WAY off my radar screen. These kinds of numbers, if we had them at our fingertips, show he was a major, major threat. All of a sudden, his run through the pack at Miami/Key Biscayne looks much less startling.

This also confirms (as written above) Roddick's ambiguous position, between the top two and the following pack. To an extent, Roddick is victimized by having come up against Federer so often. He's 1-13 (14 encounters).

Suppose, for argument's sake, Federer had taken up soccer - and Roddick had played these 14 matches against other top 5 players, splitting the matches 7-7.

Now his "H2H" vs top 5 is 18-12 (60%), vs top 10 is 42-23 (65%), and vs top 50 is 64-40 (62%).

In my "model of the universe," Roddick is the player whose place in tennis history has been most affected by Federer.

Rich stew.

Posted by sophie 04/16/2007 at 06:50 PM

Well done for the hard work, Rosangel.

Berdych, too, like Nadal, did not come on the tour straightaway. He turned pro in 2002, but played Futures and Challengers throughout 02 and 03 except for 2 tourneys, so came on to the tour full time in 04. He's had to meet top 5 and top 10 almost as much as Nadal in less time - says a lot about draws and whether you are seeded or not :(

And just as you gave credit to Connors' winning career % having played twice as many matches as Borg, Federer's winning % of 79.6 is from 624 matches compared with Nadal's 256 played. Not to say he won't keep that up, but it will be interesting to see Nadal's stats after 600+ matches.

Gaudio's stats I found interesting with over 50% wins in all the tables. He's right up there with the top few guys :) Probably mainly on clay, but still valid.

Posted by Rosangel 04/16/2007 at 06:58 PM

Canas career winning percentage is 58.95%, by the way.

One of the most astonishing finds for me was Tim Henman's record against top players and the rest. It's not just that he's 6-7 against Federer - also that over his career he's generally been consistent, annd seems to have beaten the players he should have beaten, most of the time.

Posted by Snoo Foo 04/16/2007 at 07:00 PM

Thanks Rosangel, this is freaking awesome! I hope El Gato reads this, maybe it will inspire him.

Posted by Ruth 04/16/2007 at 07:03 PM

media: The best thing about tables like Rosangel's is that they separate the facts opr reality from wishful thinking and opinions that are based on flimsy data.

Posted by ptenisnet 04/16/2007 at 07:15 PM

Nice work Ros, although I am still looking for way to disagree with you on Kolya.

Posted by Rosangel 04/16/2007 at 07:17 PM

sophie: I noticed that, neatly, Federer's 266 matches against the current top 50 were exactly twice as many as Nadal's. And agree that we should give as much care to assessing the start of Berdych's career as Nadal's. Those 'turned pro' stas are only a guideline.

Posted by FoT 04/16/2007 at 07:18 PM

Wow Ros! I'm very impressed! Great job!

Posted by Beckham 04/16/2007 at 07:25 PM

Great job Ros

I agree with you if TMF was not there Andy would have a better record...but you play who is in your era right...Same could be said for players who played against Borg and Sampras...No?

As for the younguns (Rafa included)...they have great results but I want to see them play a lot more matches like Fed and Roddick before they are annointed as the next great one IMO.

Posted by Beckham 04/16/2007 at 07:26 PM

oops I meant I agree with Andrew

Posted by Lisa 04/16/2007 at 07:28 PM

wow, you did some serious number crunching, Ros. Thanks!

Posted by Flyer 04/16/2007 at 07:35 PM

Roseangel - just beginning to "digest" all this info after a very filling dinner.

Extremely interesting - thank you for all the hard work - I can foresee it providing the foundation for many hours of contemplation & discussion....

Posted by skip1515 04/16/2007 at 07:46 PM

Thanks, Rosangel, for all this work. Anecdotally, or from the seat of my pants, I'd say that through the years the top players exhibit consistent results against everyone, and the #'s 5 through 15 spots (roughly) are constantly changing as players' levels of confidence ebb and flow. And, too, it changes as their competitors figure out how to play them, or not (e.g., Federer and Nadal). Without delving too, too deeply into your findings it appears this is the case.

For more on the Laver/Rosewall h2h, and a good history lesson about the pro years of the late 50's and early 60's, folks should visit Bud Collins' site and click Bud's Notes. ( There is an article there about someone who researched the Laver/Rosewall rivalry, with the results:

"But somebody has counted....Austrian tennis historian Robert Geist. He will soon publish his research, hailing Laver the victor, narrowly, 75-66."

There's much more to the article, all of it interesting. I recommend it.

Thanks again.

Posted by JAG 04/16/2007 at 08:18 PM

"1) Wawrinka is the highest-ranked (No. 36) never to have had a win against a Top 5 player"

Maybe the reason why Fed won't commit to play DC against top teams :)

Posted by Beth 04/16/2007 at 08:25 PM

wow - Rosangel - that was a great deal of work you did . Will take me some time to sort through it all - lots to think about.

Posted by jb 04/16/2007 at 08:34 PM

skip - interesting read - thanks for the bud collins link.

Posted by steggy 04/16/2007 at 08:57 PM

Ros: Changes have been made; Pete probably got a bit confused when he was editing this up earlier today (it was a great deal to sift through *shrug* these things happen..).

Posted by ptenisnet 04/16/2007 at 09:13 PM

Stan Wawrinka beat nalbandian twice in 2006.
in cincy - nalby's rank was 4
in basel - nalby's rank was 3.

Posted by Slice-n-dice 04/16/2007 at 09:38 PM

Fantastic stuff, Rosangel! Very very informative, but also not too far off from what common sense and general observation has been telling us.

I'll play devil's advocate of a ort here and throw in this thought: it dosn''t entirely surprise me that we have a few "giant killers" out there in the likes of Canas, Murray, even Hrbaty (okay, he's done well againthe top two, at least). As any player knows, it's easy to get "up" for a match against a highly ranked player. What's really tough is to play "up" when you're playing down.

It was always a barometer for the young players when they went up against Connors, Chang or, more recently, Agassi. These tough veterans rarely lost to lesser players, even late into their careers. Once they did, it was clearly an indication that it was time to step down. Inversely, a young gun who was able to beat one of these veterans had achieved a "breakthrough" win.

I'd be curious to see what the numbers bear when we look at how the top 10 fair against players in the 20-50 range.

Rosangel, you've served up a beauty!

Posted by Pete 04/16/2007 at 09:47 PM

Hi everyone! Yes, this was somewhat confusing to boil down to bullet points and I apologize for what confusion one or two of the bullet-points might have created. I am in touch with Rosangel and I will continue to revise or update (as I just did) as needed.

Ptenisnet - I'll defer to Rosia on the Wawrinka issue, so let's see what she has to say . . .

Posted by MWC 04/16/2007 at 09:50 PM

Obviously a lot of work was put in to this post, so kudos for that. I think it would be more valid as it pertains to RIGHT NOW, to see the same table over the last 16 months. That would give us a true indicator for RIGHT NOW. Kudos though to Roseangel.

Posted by ptenisnet 04/16/2007 at 09:56 PM

Hey Pete, welcome back and all.

I think I might have made a mistake on Stan. Rosia's stats are for results agains the current top 5.
Not, as I thought, against a top 5 when the opponent was in the top 5 at the time of the win.

Posted by Fefe 04/16/2007 at 10:19 PM

ptenisnet - I thought so too. So this is all based on current rank and not their rank at the time of the match?

Posted by steggy 04/16/2007 at 10:38 PM

Pete: So, basically, you refixed what I fixed that was broken. Brilliant.

Posted by Ramesh 04/16/2007 at 11:20 PM

I have not read through the entire post and the comments section so if it has already been mentioned my apologies.

I want to point out that Roddick's stats are also skewed for one reason. His dismal record against Roger. If we take that out of this data, things will become really interesting.

Roger has been top 5 consistently since Feb 2003 and from that date Roddick has lost 10 matches to Roger (except for 2003 Montreal. Without that his record would be 12 and 10.

He will be closer to Nadal in this analysis.

Although that doesn't really make him close to Nadal in the actual game.

Thanks for putting this data together.

Posted by Zola 04/16/2007 at 11:27 PM

if you take out the Fed data for Roddick, you should do that for everyone. I think almost everyone not named Nadal has a skewed data when matches with Fed. included. what about Gonzo, Blake?

Posted by MWC 04/16/2007 at 11:28 PM

Suggestion to someone out there, or myself, to do which would augment this discussion. Find out through data analysis of rthe top 10 players now, their average margin of victory per set...obviously the closer it is, the less dominant a player they are. Just a suggestion. Could find all needed information on

Posted by Sam 04/16/2007 at 11:29 PM

Ros: Thanks for putting all this together! Lots of great information here.

Loved the "Human Litmus Test" line about Davydenko. Good to see the data backing up his artificially high ranking. All his wins against the top 5 have come against Gonzalez (he is 0-13 combined against Federer, Nadal, and Roddick).

Posted by Sam 04/16/2007 at 11:36 PM

Records against Fed:

Gonzo: 0-10
Blake: 0-6
Davydenko: 0-8

Posted by ptenisnet 04/16/2007 at 11:47 PM

How exactly would you assess the points on the tour to calculate davydenko's "actual" ranking as opposed to his artificial ranking?

Posted by Ray Stonada 04/17/2007 at 12:32 AM

Word, ptenis. Rankings are rankings. By definition.

Posted by ajv 04/17/2007 at 12:56 AM

Borg's all-time winning percentage is impressive enough at .835 (I'm going by the stats in Collins' tennis encyclopedia), but even that figure is misleading. Borg basically stopped playing full-time in 1981. His record thereafter (from 82-91 he played a total of 7 matches; in 92 he played 8 and in 93 he played 3) was 3-15. Given that he played so few matches in total (455), that 3-15 skews his overall winning percentage big time. If you eliminate those 18 matches, and I would argue that you pretty much have to since he was not a full time player, his overall winning percentage rises to an out of this world 85.68.

The other amazing thing about Borg's record has to do with the role his extreme youth both at the beginning and end of his career plays when you look at his stats. During Borg's full-time playing career he lost a total of 60 matches. Of that total, 34 were suffered when he was between 16 and 20 years of age, as a teenager. His record thereafter is positively Rajah-like in its metronomic minimalism: At 21 and 22, he losses 4 matches each year; at 23, 24 and 25 he loses 6 each year. Done. I would argue that his record overall winning percentage while certainly aided by his early exit, is actually negatively impacted by his precociousness.

Setting Borg aside and moving to Connors, his record of 10 consecutive years (1974-1984, a period of monumental change in the game) with a better than 80% winning percentage in each year (his lowest was .825) is extraordinary. Only two other male players that I could find come close to matching it, and ironically they are two of his bete noires: Lendl from 81-90 did it in every year except 84 when he won .795; and McEnroe from 78-86 had 9 years of over 80%. These three head cases are of course blown out of the water by Evert/Navratilova who exceeded the 80% mark in 18 and 17 consecutive years respectively!! I have to believe that these records of prolongued greatness will be hard to match going forward.

Posted by Suresh 04/17/2007 at 01:07 AM

Rosangel, that was a great piece of research, thanks!

Posted by JTContinental 04/17/2007 at 01:22 AM

Impressive. I feel dizzy just thinking about the math involved.

Posted by abbey 04/17/2007 at 01:41 AM

rosia, aren't you supposed to be a high-powered, busy career woman? i'm very impressed you found the time to do this. what i like about what you, sam and andrew have done is to show how numbers generally don't lie and that the rankings are very good barometers.

really, kudos to you all.

Posted by Rosangel 04/17/2007 at 02:40 AM

ptenisnet: Agred. These numbers are all against the current top 5, top 10 and top 20, though. They are looking at what a player's record is against the top guys of today, and therefore what he has to do going forward. I considered looking at this using rankings of players as and when they met, but preferred the 'current' solution.

Posted by Rosangel 04/17/2007 at 02:56 AM

Sam: The "Human Litmus Test" moniker for Kolya was originally Pete's. He had a post about him last year.

ptenisnet: Kolya plays the rankings system, as he is permitted to do. If you averaged out players' results instead of ignoring the worst, Kolya's ranking would suffer. His match-winning percentages are closer to those of a Youzhny or a Ferrer than his top 10 peers, and objectively he does not have a strong record in winning big matches.

ajv: all my stats on winning percentages came from the numbers on ATP's website. For some reason, it didn't occur to me to include Lendl's winning percentage in the mix, so I'm happy that this was mentioned in the comments here.

Posted by bb 04/17/2007 at 03:17 AM

It is widely believed that statistics don't lie; that's true, but they never convey the complete picture: while the career percentages of Connors and J Mac seem impressive, you would also need to look at just WHO they were playing also.
If memory serves, Connors was 31 for his last slam and he retired at 39 and there were at least 2 years in betwen when he was injured, so who were his victories against in those years and how many matches did he play/ win then relatve to his peak years? and so on.
It also true with Mac who posts high career numbers: he won his last slam at a relatively young age and then seemed to hang on forever not winning anything significant, so what explains his figures?

I guess the point of some of these questions is that the numbers against the top 5, or maybe the top ten are the only ones which have any real significance.
A guy could have a high career percentage from winning against schmucks, but have a low career total of matches played.

Another thing the statistics don't reveal are the Federer lost years, for example: As we all know he was a slow starter, relative to not only Nadal and Djoko, Sampras, Becker, you name them, then suddenly caught fire mid 03.As many commentators have said, the last 3 years of Feds career could be the best in any 3year pro-tennis career.

Therefore perhaps a player's peak year's percentages should only be included. This is related to Connors's twilight years also. Then the stats would have more meaning.

But these suggestions are not in any way meant to denigrate the very fine effort here.

It would also be great to see some more figures about these mysterious figures from the past-like Laver, Emerson, Gonzales etc.
How do they bear comparison with modern players beyond the anecdotal evidence

Posted by Split Infinitive 04/17/2007 at 04:05 AM

Neat work here Rosangel! It'd probably be interesting to see %age wins against top-5 players defined in terms of playing a top-5 player at the time of the match, rather that records against current top-5. If anything, I suspect, we might see some %ages decline marginally. Another way to slice-n-dice data would be to see data for the last, say, two years only - as an indicator of current form.

Not that I'm suggesting that you actually do this, just wishful thinking!

Posted by Annabelle 04/17/2007 at 04:28 AM

Thanks so much for your efforts, Rosangel! Very impressive! And I was worried I'd misjudged you and you'd hit the town and went out headbanging...

Props to you, Rosia! Props to Andy Roddick! I was of course very excited about the concrete evidence available for your statement: "Federer's numbers versus the Top 50 are astonishing; Nadal's nearly as impressive - especially when you consider his age. But Roddick is also at least 10% ahead of the other contenders." I'll keep your "The Top 12 versus the Top 50" table as an essential handy (and frequently required) reference if I may! Thank you!!

Must go back to your post and study the non-Roddick details.

I did enjoy reading the oddities! Any more? :-)

Posted by Rosangel 04/17/2007 at 06:20 AM

Annabelle: a couple of things I left out due to trying to keep my wording concise:
(1) The other member of the top 50 not to have wins over a top 5 player are Gilles Simon and Marc Gicquel. Simon does have one win over a top 10 player (Tommy Robredo); not Gicquel, though.
(2) Pete Sampras has a winning record against the current top 50 (52%) of 13 wins to 12 losses. In spite of a losing record against Federer, Roddick and Hewitt, he managed enough wins over Haas, Youzhny and Nieminen to swing the numbers in his favour:)

I found it interesting that we had to go down as far as the mid-30s before finding a player with no top 5 wins (and even then, as ptenisnet points out, Wawrinka had beaten a former top fiver). It would seem to confirm one of the things that various people keep saying in different ways (and I generally agree with) - that on the ATP tour, the depth is great enough that any player, on his day, can pose a threat to the best players. I wonder if the stats would look like that in the WTA. I suspect not, but when I have time, I might look into that one by the numbers as well.

Posted by Ray Stonada 04/17/2007 at 06:56 AM

Sampras is one player with whom I trust the numbers outside of Slams very little - he sleepwalked through so many smaller tournaments, only to come back and put a beating on the same players when he met them in Slams. In the 2001 US Open, for instance, he just came out and waxed Roddick in the quarters. Sampras' inconsistent valuation of tournments means his overall numbers understate his big-match dominance, I think.

Posted by Andrew 04/17/2007 at 07:32 AM

Split Infinitive: the database that Sam and I are working on should have H2H results against then top-5 players. Stay tuned...

bb: Federer's "lost" years is a concept which has some currency, but when you look at the data for multiple GS winners, you find that Federer is not very atypical. His rapid rise at age 21 was about a year later than Sampras (or Roddick, who was a nearer contemporary), but was earlier than the similar rise for Ivan Lendl.

Work like Rosangel's is really interesting because it shows us lots of stuff we didn't expect to see - for example, the large gap between Hewitt and Safin, two former Tier I players.

Posted by roGER 04/17/2007 at 07:48 AM

Thanks so much for the stats, and the interpretation (including the commenters).

Really interesting and a great way of quantifying extraordinary skill and tenacity.

Posted by roGER 04/17/2007 at 07:48 AM

Thanks so much for the stats, and the interpretation (including the commenters).

Really interesting and a great way of quantifying extraordinary skill and tenacity.

Posted by Sam 04/17/2007 at 08:12 AM

Ray: Good point about Sampras. To me, his 14-4 record in Slam finals speaks to his big-match dominance. Also, the USO match with Roddick was in 2002. Roddick had beaten him in Houston earlier that year.

Posted by Allie 04/17/2007 at 08:13 AM

Rosangel - Wow! Soo much work! And very interesting results.

OT, sorry steggy, feel free to move it: for the Fed lovers- he's back to vogueing. he's on the cover of Men's Vogue this month. This time photos by Annie Liebovitz.

Posted by Rosangel 04/17/2007 at 09:00 AM

Thinking about the relatively thin H2Hs noted, I thought it might be interesting to list some of the head-to-head matchups that haven't yet happened:


Posted by Allie 04/17/2007 at 09:03 AM

fingers crossed for Safin -Nadal this week! Would love to see that one.

Posted by Annabelle 04/17/2007 at 09:05 AM

Thanks Rosangel! So many details! Really interesting. I agree that there is great depth on the ATP tour, wonderful to have it confirmed in such a manner. Thanks.

Posted by RedEric 04/17/2007 at 10:12 AM

The more charts and graphs poeple come up with (excellent work!), the clearer it becomes to me that Davydenko is not a true top tier player. From these ctables, I see he wins less than 50% of the time against top 5, top 10, and top 20 players. When you scan the stats, his stick out like a sore thumb. I'm not saying he doesn't deserve his ranking, though. It just tells me that work ethic can indeed get you pretty far.

Posted by frances 04/17/2007 at 10:14 AM

Please do the women.

Posted by Diwakar 04/17/2007 at 10:49 AM

vulnerbility in sticking/staying at the top position is apparent from the variation/increasing %wins. Meaning the more variation you have (+ve that is) going from Top 5 to Top 10 to Top 20 the more vulnerable you are at the current ranking...
Whats very interesting and obvious is that Baghdatis goes from 8-38-45 shows how he hasn't and probably will never improve, sadly... :)

Amazing contrast is that Rosengel already pointed out the young champs have the opposite meaning they are here to don't need to watch their matches...purely statistical...
very intersting... thanks Rosangel!

Posted by Rosangel 04/17/2007 at 11:15 AM

As per your wish, here are the winning percentages (of total matches played) between the Top 12 and the players ranked 21-50, as of last week - all I have done is deduct the Top 20 wins and losses from the top 50 wins and losses, to get these H2Hs.

Federer 81%
Nadal 78%
Roddick 77%
Davydenko 56%
Gonzalez 61%
Robredo 57%
Djokovic 58%
Ljubicic 68%
Blake 66%
Murray 58%
Haas 67%
Nalbandian 66%

Again, it separates Roddick from the rest of the Top 12 below him, though we know his percentage against the Top 5 is lower than some of the other top players. It also puts Kolya at the bottom.

Posted by JR 04/17/2007 at 11:24 AM

Great job Ros. I’ll certainly save this. I thought you had been unusually quiescent.

Re: “I considered looking at this using rankings of players as and when they met, but preferred the 'current' solution.” I’d love to see that. I think it would be even more revealing.

Posted by Rosangel 04/17/2007 at 11:32 AM

Of course, one of the interesting things about Federer's stats is that his match-winning percentages are higher against the top 5 and top 10 than against the top 21-50. The percentage for the top 20 and 21-50 are the same. Federer statistically seems to have been more likely to win the bigger matches than the smaller ones. Not so for anyone else here.

Posted by JR 04/17/2007 at 11:58 AM

Interesting to see how well Canas came out on all scores.

When Djoko won Miami I was thinking of doing something to quantify the significance of a Masters win. Just looking at the past winners shows it’s a pretty elite group (I think there was only one winner who’s name I did not recognize). However, I think it would be interesting to see how many winners did/did not have other significant accomplishments preceding or following their win(s)-—say, multiple top ten rankings, slam semis in the 2 or 3 years preceding or following a Masters win.

Posted by Andrew 04/17/2007 at 12:54 PM

Rosangel: it's possible that the Federer stats against 21-50 are more weighted to the first part of his career, while he's played more of the higher ranked players as a no 1 seed going deep into tournaments from late 2003 onwards.

Also, the intuition that many folks have on Davydenko and Robredo seems to be borne out. Robredo, for example, is 12/21/37/49% vs the top 5/10/20/50 - he's not even ahead vs the top 50...

Posted by omar 04/17/2007 at 12:59 PM

The one thing I would like to bring up is that until the mid 80s all the tennis stats are somewhat funny due to the different governing bodies tennis had. For example, when the Grand Prix was running things, they had their tennis numbers, but when the ATP took over in the mid 80s they changed a lot of the stats that were already in the record books. They cooked the books, so to speak.

When Borg quit, his lifetime record against Connors was listed in every publication as 15-8. It stayed that way until 86 or 87, when it magically re-appeared as 10-7, then it was magically changed again a couple years ago to 13-8, which is where it stands today. Borg beat Connors twice at the WCT Invitational in 79 and 80, which counted as an official tournament according to the Grand Prix, but 8 years after the fact the ATP takes over and rules that it should not have counted. It's a joke. It would be like baseball taking away 10 home runs from Babe Ruth 80 years ago because he hit them in a ballpark that wasn't regulation size according to today's standards. When Borg retired he had 62 listed titles, but now the ATP says he has only 55. So keep this in mind anytime you talk stats and tennis.

Posted by Sam 04/17/2007 at 01:15 PM

"Rankings are rankings."

But they're not always an accurate barometer. We've had number one players who got there without winning a Slam, for example. On the women's side, Clijsters and Davenport in recent years.

Posted by Sam 04/17/2007 at 01:17 PM

Davydenko's finals in 2006. The third column is the opponent's ranking at the time of the match:

W Hrbaty, Dominik (SVK) 27 6-1 6-2 6-2
W Safin, Marat (RUS) 65 6-4 5-7 6-4
W Calleri, Agustin (ARG) 32 6-4 6-3
W Mayer, Florian (GER) 60 7-6(6) 5-7 6-4
W Pavel, Andrei (ROU) 111 6-0 6-3
F Robredo, Tommy (ESP) 9 2-6 1-6
F Nalbandian, David (ARG) 4 3-6 4-6

Posted by omar 04/17/2007 at 01:20 PM

How about Martina Hingis in the late 90's over Venus Williams, even though Venus won the biggies and always beat Martina. Martina played more matches though, so she had more points. Or McEnroe being ranked No. 1 in 82 even though he had 0 GS wins and Connors had 2. The ATP didn't even do a year-end ranking, they should updated the ranking every two weeks, not taking into account the time of year. You could have a guy ranked No. 1 for 50 weeks out of the year, but if the two weeks he isn't No. 1 were the last 2 weeks in December, the computer would not recognize him as the No. 1 player for that year. Mark Twain said it best, "Which would you prefer, the statistics or the facts."

Posted by Rosangel 04/17/2007 at 01:47 PM

Andrew: your point about the weighting could be correct, though to be fair he's played more matches than anyone in times, which naturally means he's played a lot of the lower-ranked players as well, in early rounds.

JR: FYI here is Canas' win-loss H2H against last week's the Top 20:
Federer 3-1
Nadal 0-2
Roddick 2-2
Davydenko 3-1
Gonzalez 0-0
Robredo 3-1
Djokovic 0-1
Ljubicic 2-0
Haas 4-1
Blake 0-1
Murray 0-0
Nalbandian 2-2
Ancic 1-0
Berdych 1-0
Gasquet 1-0
Ferrer 0-0
Youzhny 0-1
Baghdatis 0-0
Hewitt 2-5
Nieminen 0-0

Anyone with an interest in another specific top 50 player's H2H versus the same players please say so and I can post similar.

Posted by Rosangel 04/17/2007 at 01:49 PM

In the last post, I meant "in recent times" in the second line.

Posted by Sam 04/17/2007 at 01:57 PM

Oops, sorry about the formatting in my 1:17 post. Looked fine when I cut & pasted it ...

Posted by billy 04/17/2007 at 02:00 PM

Marat just lost the match. What is wrong with this guy. He blew a lead of 6-0 and up a break in the second set. Somebody call a shrink

Posted by ptenisnet 04/17/2007 at 02:00 PM

So let's take a look at 4 of Davydenko's 13 losses against the top 5. Or more to the point, Roddick's wins against the top 5.
Please note, my intention here isn't to bash Roddick, but to make a point. Among the top 5, Federer and Nadal's status was never in question. Kolya is the topic under discussion. And Gonzalez just made it into the top 5 club. It hardly seems fair to haze him right off the bat. So that leaves Roddick

The number after the tournament name indicates davydenko's rank at the time of the match and the number in parantheses indicates Roddick's rank at the time of the match?
What if anything does this tell you about Roddick's current state or for that matter Davydenko's current state?

2002: delray beach - 86 (13)
2003: st poelten - 36 (6)
2004: doha - 44-52 (1)
2005: AO - retired - 26 (2)

I have a similar stat for Roddick's victory over Nadal
2004 US Open - 49 (2)

Further, if by some miraculous happenstance, Giles Muller replaces davydenko in the top 5, Roddick's record against the top 5 drops from a respectable 12/18 to a slightly more disturbing 9/19.
Without Roddick having to lift a finger.

Posted by ptenisnet 04/17/2007 at 02:02 PM

Marat just lost the match. What is wrong with this guy. He blew a lead of 6-0 and up a break in the second set. Somebody call a shrink

Someone somewhere owes my 5 bucks.

Posted by Moderator 04/17/2007 at 02:05 PM

Please adhere to the topic of the current post. Thank you.

Posted by Jamaican Girl 04/17/2007 at 02:18 PM

great stats, but I think that Nadal should be a Tier 1

Posted by JR 04/17/2007 at 03:24 PM

What I meant to highlight (if I’m interpreting and counting correctly) is that Canas is #2 (top 5, h2h %), tied for #2 (top 10) and tied for # 3 (top 20). Tier 1, anybody?

Posted by patrick 04/17/2007 at 03:32 PM

Rosangel, please do Federer,Nadal & Roddick's H2H record against Top 25 just like you did at 1:47PM today if not done already.

Posted by skip1515 04/17/2007 at 03:47 PM

Sam, regarding being #1 and not winning one of the Big Four tournaments in a calendar year: suppose, just for argument's sake, that a player went 90/4 for the year, losing only in all 4 finals of the Grand Slam tournaments, while no one else won more than two.

Number One for the year, or no? If rankings were based on single year performances they'd quite likely be at the top of the list.

Not exactly my scenario, but Henin in '06?

Posted by Zola 04/17/2007 at 03:59 PM

Jamaican Girl -"great stats, but I think that Nadal should be a Tier 1

I thought he was! isn't he?

Posted by Waid 04/17/2007 at 04:01 PM

On what is based the argument that Davydenko's ranking is inflated because he plays a lot? I mean, 18 tournaments are counted in the rankings, only 5 of which are small tournaments. So no matter how many tournaments outside Grand Slams and Masters' he plays, only 5 are counted in his ranking. Every one of the current top 10 have at least 18 tournaments in their ranking. Okay, if a player skips a GS or masters event, it gets counted as 0 points, but that means there are only 13 tournaments that are counted in whether you play them or not. Without any research, I would think that Davydenko's ranking is so high because he gets far in Grand Slams (except Wimbledon) very consistently. Right now, his GS results in his ranking are one semifinal, 2 quarters and one first round (Wimbledon). On top of that, there is a masters win. Those results only are about half of his points.

Posted by Paranoid Android 04/17/2007 at 04:07 PM

Canas had a grand total of 13 wins against top 10 players in 10 years pre-suspension. That averages out to a little more than one a year.

Since coming back from suspension, he has 5 wins over top 10 players.

Before the suspension, Canas had 6 career wins over top 5 players in 10 years; and 3 of those 6 wins came at one tournament('01 Toronto Masters).

Since coming back, he has 2 wins over the number 1 player.

Moreover, despite being one of the oldest players on that list, he had a very small number of matches against the top 10(23) before his suspension. Only Monfils, Gasquet, Baghdatis and the Djoker have fewer matches against the top 10, but they are all very young.

That(and never getting past the quarters of a major and making one Masters final) tells me Canas wasn't consistently going deep into the bigger tournaments(pre-suspension), which limited his number of matches against the upper echelon players.

Since coming back, he is not only playing upper echelon players but beating them. Moreover, he started the season 17-4(5-1 against the top 10), making 2 finals, which is a career 29.

1 2      >>

We are no longer accepting comments for this entry.

<<  Wednesday Racket (OT) Monday Net Post  >>

Wild Women of the U.S. Open
Wild Men of the U.S. Open
Roddick's Imperfect World
"It's Kind of a Dance"
Nadal's Kneeds
The Racquet Scientist: Canadian Tennis
The Long and Short of It
This blog has 3693 entries and 1646148 comments.
More Video
Daily Spin