The 38 Greatest Tennis Players of All Time: A New Comprehensive and Methodical Survey

chaognosis

Semi-Pro
The Greatest Tennis Players of All Time?

Rated by Major Singles Titles and Years No. 1 - including Amateur, Pro and Open data*

by CHAOGNOSIS

Male
01. Pancho Gonzales (United States)
--. Bill Tilden (United States)
03. Rod Laver (Australia)
04. Pete Sampras (United States)
05. Ken Rosewall (Australia)
06. Don Budge (United States)
--. Roger Federer (Switzerland)
08. Jimmy Connors (United States)
--. Ellsworth Vines (United States)
10. Jack Kramer (United States)
11. Henri Cochet (France)
--. Roy Emerson (Australia)
--. Ivan Lendl (Czechoslovakia)
--. Fred Perry (Great Britain)
15. Bjorn Borg (Sweden)
--. John McEnroe (United States)
17. John Newcombe (Australia)
18. Frank Sedgman (Australia)
19. Rene Lacoste (France)
--. Tony Trabert (United States)
21. Andre Agassi (United States)
--. Stefan Edberg (Sweden)
--. Bobby Riggs (United States)

Female
01. Margaret Smith Court (Australia)
--. Steffi Graf (Germany)
03. Helen Wills Moody (United States)
04. Martina Navratilova (Czechoslovakia)
05. Chris Evert (United States)
06. Billie Jean King (United States)
07. Suzanne Lenglen (France)
08. Maureen Connolly (United States)
--. Monica Seles (Yugoslavia)
10. Margaret Osborne DuPont (United States)
11. Justine Henin (Belgium)
12. Maria Bueno (Brazil)
--. Lindsay Davenport (United States)
--. Martina Hingis (Switzerland)
15. Serena Williams (United States)

*In compiling these rankings, I have limited myself to the most widely available, and accepted, sources. (All data, except for the most recent few years, can be found in the 2003 edition of Bud Collins's Total Tennis: The Ultimate Tennis Encyclopedia.) Prior to the start of the Open Era, my formula gives equal weight to the four major amateur championships (Australian, French, Wimbledon, United States) and the three major professional tournaments (French Pro, Wembley, U.S. Pro); also, it considers both the most authoritative amateur rankings—published by the London Daily Telegraph since 1913—and the results of the best-documented professional head-to-head tours. Rankings are provided for all players, male and female, who scored above a certain threshold according to my calculations. Although more male players made the cut than females, the top four female players all received higher scores than the top-rated male player.

As with any methodology, the quality of the results is contingent upon the accuracy of the data. Since I have limited myself to the most trusted indicators, for the most part this is not a problem; however, in a few instances the records of the early professional years are vague or conflicting. In one notable example, whether Vines is tied with Connors for eighth place, or whether he is rather tied with Kramer for ninth, depends upon the contested status of the 1936 Wembley event. In such cases I follow Collins, who himself follows the research of the late Joe McCauley (The History of Professional Tennis, 2000), though I am aware that more compelling evidence has surfaced since the last edition of Collins's encyclopedia was published. Hopefully, more authoritative sources will emerge in future years that can help set the record straight regarding many of these inconsistencies and puzzles. The reader should therefore treat these rankings not as absolutely precise, finite, or stable, but rather as an approximation based on the present state of our reconstruction of the complex history of the sport. As more research is done (by Robert Geist, Ray Bowers, and others), minor changes in the "pecking order" may occur.

I am slightly discontent that in both the male and female lists, the formula fails to produce a definitive No. 1. However, it seems reasonable to assert, at the very least, that Gonzales and Tilden on the men's side, and Court and Graf among women, represent the most impressive combinations of dominance and longevity yet seen in the game of tennis. Certain other results may surprise you, as they surprised me: the high rankings of Vines and Cochet indicate that these two are among the most underrated players of all time, whereas the relatively (even shockingly) low positions of Borg, McEnroe, and Agassi probably mean that these three players have been overrated by the current generation of fans and analysts--including myself, I might add! Likewise, DuPont and Bueno are names that may not be familiar to many, but these two players are probably more accomplished than the comparatively over-hyped Serena Williams. To add a comparative perspective, according to my calculations, Serena has achieved barely a quarter of what Court, Graf, or Wills achieved in their careers. Agassi, too, despite his long career, was only about a third as accomplished as Gonzales, Laver, or Tilden.
 
Last edited:

anointedone

Banned
I stopped reading when I saw Borg down at #15, especialy with Connors who Borg proved he was superior to at #8, and even Roy Emerson ahead of Borg.
 

drakulie

Talk Tennis Guru
I stopped reading when I saw Borg down at #15, especialy with Connors who Borg proved he was superior to at #8, and even Roy Emerson ahead of Borg.

I have to agree.


chaognosis, although I don't agree with much of the list, I do appreciate the time and effort you took in making and posting it here.
 

vkartikv

Hall of Fame
Did Justine Henin change nationality?? You have her listed as a US citizen.

The men's list is extreeeeemely biased. I think you know towards who...
 

chaognosis

Semi-Pro
I stopped reading when I saw Borg down at #15, especialy with Connors who Borg proved he was superior to at #8, and even Roy Emerson ahead of Borg.

That was the real shocker for me, too--as I pointed out already in my original post. The reason is obvious: Borg finished on top of the computer rankings only twice in his career, as opposed to five times for Connors. If you disagree with the result, then what is needed is a critique of the ranking system. (I always prefer constructive critiques, rather than just "stopping reading" when I am troubled by something.) I often find the ranking system flawed myself, particularly in those early years of its implementation, but as it seems to be the most widely accepted system since 1973, I resigned myself to abide by it. Just as a counter-argument, Borg's dominance over Connors in the late-1970s doesn't technically "prove" anything, except that he beat Connors more often than Connors beat him during those years. It is the superior longevity of Connors, clearly, that puts him ahead of Borg by this methodology.

Good observations, though.
 

chaognosis

Semi-Pro
I have to agree.


chaognosis, although I don't agree with much of the list, I do appreciate the time and effort you took in making and posting it here.

Thanks.

I'm well aware I can't fully please everyone (not even myself, most of the time), but I thought it would be interesting to offer a ranking based on a truly comprehensive survey of the available data in what are widely considered the two most important categories: major singles titles and years at No. 1.

I myself am pretty pleased with the results--particularly the top five. After that things always get fuzzier.
 

chaognosis

Semi-Pro
Did Justine Henin change nationality?? You have her listed as a US citizen.

Grrrr... thanks for pointing this out; I just fixed it. I thought I proofread my list pretty well, but it's an awful lot of text. At least one mistake was bound to happen.

The men's list is extreeeeemely biased. I think you know towards who...

Really? Towards whom?

I literally just plugged the numbers into a very simple formula, which seemed to me to be pretty fair to all players dating back at least to 1913 (when the first world rankings were compiled by A. Wallis Myers... before that it's much harder to devise any sort of objective standard, though I would commend the work of SgtJohn in this respect).
 

Moose Malloy

G.O.A.T.
The reason is obvious: Borg finished on top of the computer rankings only twice in his career, as opposed to five times for Connors. If you disagree with the result, then what is needed is a critique of the ranking system.

Well, considering the fact that the ATP actually gave its 'Player of the Year' award to Borg instead of Connors in '78 (who finished #1 on their computer) does say quite a bit about how respected that ranking was.

I admire your effort chaognosis, but many weigh peak level of play more than longevity(or at least try to weigh both) Your system is all about longevity. And it seems strange to weigh Emerson's majors the same as Borg's.
 

chaognosis

Semi-Pro
Well, considering the fact that the ATP actually gave its 'Player of the Year' award to Borg instead of Connors in '78 (who finished #1 on their computer) does say quite a bit about how respected that ranking was.

I absolutely agree. But faulty or not, you must recognize that the computer ranking is still the most widely accepted system available. Until another ranking system for those years is adopted, I don't see how I can abandon it--at least not in this kind of approach. I could just offer my own personal rankings, but I thought that would be less interesting (and less open to criticism) than this sort of methodical survey. My goal here was to be historical, which means working with the best sources I have, not inventing my own.

I admire your effort chaognosis, but many weigh peak level of play more than longevity(or at least try to weigh both) Your system is all about longevity. And it seems strange to weigh Emerson's majors the same as Borg's.

Thanks, but the system rewards both dominance and longevity equally. Clearly playing for a long time gives one more chances to win major titles or finish seasons at No. 1, but one also needs to be able to deliver. An example of a player with considerable longevity but a low "peak level" would be Agassi, who doesn't fare so well by this method. The players who came out on top are those with both a high peak level and a long career... isn't that better than a deficiency in either area?
 

Moose Malloy

G.O.A.T.
Until another ranking system for those years is adopted, I don't see how I can abandon it--at least not in this kind of approach.

I wish the Grand Prix Points Race totals were available for those years, they may be a more accurate system. In 1982 they seemed far more fair than the ATP's ranking. Here is a thread about it:

http://tt.tennis-warehouse.com/showthread.php?t=130373

Also the year end prize money leaders are worth examing for those years as well.
 

CyBorg

Legend
That was the real shocker for me, too--as I pointed out already in my original post. The reason is obvious: Borg finished on top of the computer rankings only twice in his career, as opposed to five times for Connors. If you disagree with the result, then what is needed is a critique of the ranking system. (I always prefer constructive critiques, rather than just "stopping reading" when I am troubled by something.) I often find the ranking system flawed myself, particularly in those early years of its implementation, but as it seems to be the most widely accepted system since 1973, I resigned myself to abide by it. Just as a counter-argument, Borg's dominance over Connors in the late-1970s doesn't technically "prove" anything, except that he beat Connors more often than Connors beat him during those years. It is the superior longevity of Connors, clearly, that puts him ahead of Borg by this methodology.

Good observations, though.

Why did you stick by the computer rankings? That's the real question. I don't understand your motivation in this at all. Pretty much everyone knows that these are bunk.

Focusing on longevity is another factor which I strongly disagree with. But again I appreciate your post, although I wouldn't mind more details as to your formulae.
 

CyBorg

Legend
I absolutely agree. But faulty or not, you must recognize that the computer ranking is still the most widely accepted system available. Until another ranking system for those years is adopted, I don't see how I can abandon it--at least not in this kind of approach. I could just offer my own personal rankings, but I thought that would be less interesting (and less open to criticism) than this sort of methodical survey. My goal here was to be historical, which means working with the best sources I have, not inventing my own.

The computer ranking system is incredibly fallacious. I'm also wondering how you weighed tournaments and whether you took winning percentages into consideration (another highly fallacious aspect for consideration).

I think that the best system would be to weigh tournaments based on qualities of draws from year to year and assign an independent year-end ranking for each player based on the added data. What's left are longevity and peak factors, which can be balanced.

Regardless, it is easy to criticize and much harder to actually sit down and formula one's own list. You have done that and that's quite admirable.
 

chaognosis

Semi-Pro
Why did you stick by the computer rankings? That's the real question. I don't understand your motivation in this at all. Pretty much everyone knows that these are bunk.

I would hardly say "bunk," although they do have serious problems. The reason, which I've stated a number of times already, is that I wanted to abide by the most widely accepted indicators. If I eschew the computer rankings, what then do I use? The money rankings are, IMO, less valuable even than the computer system. Should I choose a particular publication's end-of-year rankings? If so, which one is most "definitive"? Do I follow that famous Wikipedia article--what makes that particularly reputable? Bud Collins, in his encyclopedia, printed the computer rankings since 1973 for this same reason, and I think he had good reason to do so. If I were offering rankings based on my own observation and opinion, it would be a different story... I would, for example, concede that Borg was No. 1 in 1978, and Connors in 1982, and Becker in 1989, etc. But as it stands, I am trying to be as historical as possible and appeal to the best available objective standard of reference. Like it or not, for the current era, that is the computer rankings.

Focusing on longevity is another factor which I strongly disagree with. But again I appreciate your post, although I wouldn't mind more details as to your formulae.

Again, I don't see how I'm focusing on longevity here... players are rated more highly for winning more majors (no matter who long/short it takes them to do so), or for finishing more years No. 1. Dominance and longevity are both necessary by these criteria. On the contrary, I don't see how to devise a good system that doesn't take into account a player's whole career, which seems to be what you're calling for. I know you prefer more qualitative judgments (i.e., the praise of Vines, Hoad, etc., offered by peers), and that to me is perfectly valid, but it is not what I am offering in this thread. The stated goal is a comprehensive, methodical survey, which is what this is. I've simply isolated what many consider to be the two most important criteria (major titles and years at No. 1), and included the most authoritative data from the amateur, pro, and open eras.

The formula, as I said, is supremely simple. I count the number of major titles a player won (amateur, pro, open), and I count the number of years a player was considered the "world champion" by the most widely accepted standard of the day (amateur rankings, pro world championship tours, and open-era computer rankings). I then rescale the variables to give them each equal weight, and I add them together. This way I don't have to decide which criterion is more important than the other.
 

drakulie

Talk Tennis Guru
I think that the best system would be to weigh tournaments based on qualities of draws from year to year and assign an independent year-end ranking for each player based on the added data. What's left are longevity and peak factors, which can be balanced.

Wow! we actually agree on something!

I have always thought of this, and wondered how it could be done. I'm quite certain people will argue one tourney is being given too much value vs. another, or won't be able to agree on the "quality" of the draw.

If it ever could be done, it would be interesting to see what the results would turn out.
 

BTURNER

Legend
The ladies list is pretty close to my own views! Except I put Graf ahead of Court. So I think it is genius!
 

CyBorg

Legend
I would hardly say "bunk," although they do have serious problems. The reason, which I've stated a number of times already, is that I wanted to abide by the most widely accepted indicators. If I eschew the computer rankings, what then do I use? The money rankings are, IMO, less valuable even than the computer system. Should I choose a particular publication's end-of-year rankings? If so, which one is most "definitive"? Do I follow that famous Wikipedia article--what makes that particularly reputable? Bud Collins, in his encyclopedia, printed the computer rankings since 1973 for this same reason, and I think he had good reason to do so. If I were offering rankings based on my own observation and opinion, it would be a different story... I would, for example, concede that Borg was No. 1 in 1978, and Connors in 1982, and Becker in 1989, etc. But as it stands, I am trying to be as historical as possible and appeal to the best available objective standard of reference. Like it or not, for the current era, that is the computer rankings.

Populist (widely accepted) indicators are often horrendously skewed. The fact that the game has changed enormously over the decades make many of these virtually unusable. The ranking system has been tweaked pretty much every decade. I think that what you're doing is a good start, but you'll need something a bit more rigorous before making assumptions based on these results.

Again, I don't see how I'm focusing on longevity here... players are rated more highly for winning more majors (no matter who long/short it takes them to do so), or for finishing more years No. 1.

Well, you already know how I feel about the latter... as for the former - which majors? Any majors? The Aussie as well? What about SgtJohn's top four tournaments per year? I am wondering if you've delved into that. I'm curious, because that thread was fascinating and could be taken further.

Dominance and longevity are both necessary by these criteria. On the contrary, I don't see how to devise a good system that doesn't take into account a player's whole career, which seems to be what you're calling for.

Well, forget what I'm calling for. I'm not saying 'either/or' here, nor am I trying to say that my system is better or anything. I just think that you are probably failing to account for something, considering some of the results, which is okay. Maybe the year-end number one is exactly it.

For the record, I never advocated discarding portions of a player's career. I suggested that the peak value should be predominant.

I know you prefer more qualitative judgments (i.e., the praise of Vines, Hoad, etc., offered by peers), and that to me is perfectly valid, but it is not what I am offering in this thread. The stated goal is a comprehensive, methodical survey, which is what this is. I've simply isolated what many consider to be the two most important criteria (major titles and years at No. 1), and included the most authoritative data from the amateur, pro, and open eras.

Oh, I realize this. I am not comparing. I have never posted anything rigorous on this board and have never pretended to. I just don't think that year-end rankings should qualify as good data. In fact, they are very questionable data. Another thing to consider about data is how you use it, so just having data isn't in itself sufficient.

The formula, as I said, is supremely simple. I count the number of major titles a player won (amateur, pro, open), and I count the number of years a player was considered the "world champion" by the most widely accepted standard of the day (amateur rankings, pro world championship tours, and open-era computer rankings). I then rescale the variables to give them each equal weight, and I add them together. This way I don't have to decide which criterion is more important than the other.

What are the 'major' titles for some of these players? At least roughly speaking .. we know that in some eras guys had only three majors of value over a single year.

And what is the weight that you grant to Amateur grand slam titles?

Thanks for your clarifications.
 

CyBorg

Legend
Wow! we actually agree on something!

I have always thought of this, and wondered how it could be done. I'm quite certain people will argue one tourney is being given too much value vs. another, or won't be able to agree on the "quality" of the draw.

If it ever could be done, it would be interesting to see what the results would turn out.

The quality of the draw could be formulated by figuring out the average rating of the participants. There may have to be two variables: a) the average rating of all participants, b) a way to account for the number of top-10 participants.

Of course, this is flawed as well because the ratings were themselves flawed for many of the past decades. In addition it would be virtually impossible to apply these same formulae to the really old results. Still, it's more rigorous than most systems and at least it would account for all tournaments and not just majors.

One can at least try this for the Open Era Results. I think I'll give it a shot one of these days. It would be a big undertaking.

The most attractive notion is that some tournaments in our history that were not majors were almost as valuable as majors. One such event was Forrest Hills in the 1980s which produced some wonderful draws in the mid 1980s. Good enough perhaps to be worth even more than the Aussie. Or take Boca West in 1986 - a 128-man draw with all the big names. This deserves attention.
 

CyBorg

Legend
One more thing to add: I just don't think that year-end number one is a useful fact, even if altogether representative of the ordering of the best-to-worst players in tennis.

Here's why. Suppose that you have player A who has a fantastic season all-around, winning a major, two masters series shields, six total titles. He plays against a relatively competitive field, but there are no real dominant players aside from himself. Let's assume that each of the majors was won by different players. He winds up as the year-end number one by a smidgeon over other guys, amassing not a terribly high amount of ATP points.

Suppose a player B has almost exactly the same basic results in another year: wins a major, two masters series shields, and six titles in total. But he ends the year as number two or three in the world. The field is fairly competitive, but there are also two other guys who are at or above his level who may actually be preventing him from winning two or three majors that year rather than one.

By giving special value to the year-end #1 ranking one would posit that player A is better than player B (because all other factors are equal).

Quite problematic and I think somewhat representative of Sampras in these rankings. Let's say Federer breaks his leg tomorrow and is never well enough to play again. By virtue of his longevity and year-end number ones Sampras is the better player. This is despite the fact that Sampras finished as #1 often by virtue of bare margins.

We all know that Sampras was not better than Federer. And Federer doesn't have to play two or three more years to become better.
 

urban

Legend
It would be fine, if you would add numbers, Chaog. The system seems to prefer longevity as number one, which is of course a bit speculative for the pre open pros. This was the one vital missing link in the system of Raymond Lee (another could be best year results based on titles won and percentages, and maybe years at staying in the top ten for absolute longevity). The more i think about it, the more i like the approach of Lee, who gives importance to absolute numbers and percentages, longevity for career numbers and domination in a 5 year period.
 

Wuornos

Professional
Hi Chaog

Great list.

This is just the sort of thing I admire. You have devised a system that is comprehensive and independent and can produce results independant of your personal opinion.

You will always get criticism no matter what criteria you select or how you weight it because people have different views on what indicators of outstanding perfromance are most reliablein the evaluation of players. Don't worry to much about the people who criticise without making any attempt of their own or without providing good reason as to why your list may misrepresent achievemnet aside from their personal views.

It's great to see someone trying to look at the past players from an independant an unbiased position. You have already gone where I fear to tread which is pre open era, you get my respect just for that.

My personal indicators for greatness are threefold. Dominance, Opposition Quality and Tournament Status. Each are calculated independently without the interference of human opinion and together derive an overall peak rating for each player. My latest attempt which I am calling the DOT ratings should be completed towards the end of next week and I would very much value your feedback on it as I have always found your posts both constructive and informative.

Great work and keep the good posts like this coming.

Take care.

Tim
 
Last edited:

FedForGOAT

Professional
Great list. I just think it concentrates too much on longevity, year-end rankings, and other stats which, while impressive don't accurately reflect a player's true ability in all cases. it works for players like graf, but may be misleading for players like connors.
just my two cents..
 

Wuornos

Professional
The Greatest Tennis Players of All Time?

Rated by Major Singles Titles and Years No. 1 - including Amateur, Pro and Open data*

by CHAOGNOSIS

Female
01. Margaret Smith Court (Australia)
--. Steffi Graf (Germany)
03. Helen Wills Moody (United States)
04. Martina Navratilova (Czechoslovakia)
05. Chris Evert (United States)
06. Billie Jean King (United States)
07. Suzanne Lenglen (France)
08. Maureen Connolly (United States)
--. Monica Seles (Yugoslavia)
10. Margaret Osborne DuPont (United States)
11. Justine Henin (Belgium)
12. Maria Bueno (Brazil)
--. Lindsay Davenport (United States)
--. Martina Hingis (Switzerland)
15. Serena Williams (United States)

Hi Chaog

Thought you might be interested in a quick comparison of how your rankings compare with the initial output of the DOT ratings I mentioned earlier in this thread.

I only have the initial output of the women at the the moment, and of course it only applies to the open era. I won't publish my list here as I don't want to hijack your thread. DOT ratings are on a scale where 2700 = the mean standard of play of number 1 player during their reign while 2600 = the mean rating of a number 10 player. This isn't accidental, I standardised the ratings with output from 1969 to 2007 inclusive. Only results in majors are taken into account for these ratings and all players are rated on dominance, adjusted for relative dominance of other active players adjusted for tournament strength and recent history. So a very different approach to your own.

Your first three players from the open era are Margaret Court, Steffi Graff and Martina Navratilova. This is quite amazing as my output shows, Margaret Court 2749, Steffi Graff 2749 and Martina Navratilova 2748 at their peaks. Actually Margaret Court was 0.1 point higher than Graff but I round the ratings to the nearest point. So a great start there.

Next you have Chris Evert who has a score at her peak of 2722 in my ratings achieved after the French Open of 1986. We aren't to different here but I have Monica Seles clocking a 2728 rating after the Australian open of 1993. Given the history of Monica and the fact your system looks at total achievement either in number of events or years at number 1, while I look at Domination etc, this difference is not surprising. Maybe Monica would have gone on to be higher in your list had events been different but we will never know.

Next we move down to Billie Jean King on your list. Broadly I can agree that this is about the same as my list, however, creeping in above Billie Jean is Serena Williams with 2709 as opposed to Billie's 2701. The reason for this is simple. It's because Serena scored many of her achievements in a shorter period and therefore was more dominant. This is a fundemental difference in what we value as an indicator of greatness. Nothing wrong with that and again the two systems have shown the differences we would expect allowing for this.

On your list the next open era player is Seles whom we have already discussed above.

Below that you have Justine Henin. Here we have a larger difference. Henin has achieved plenty in her career but loses out three fold in my system. First while being the undoubted top player her achievements have been quite spread out. so her dominance score is low. Second the dominance of other players during her reign have been low. (Assuming the standard of the general population remains constant Justine would have been rated more highly had someone else been a clear number 2, consistently winning those events Justine doesn't, and finishing runner up as often as possible where Justine wins. This player hasn't been present and therefore it is more reasonable to assume that the general competition is slightly weak at the moment rather than the whole population has become stronger andd Justine is in fact outstanding.) Also no Wimbledon title. For this reason Justine appears on my list with a peak rating of 2686 achieved after the Australian of 2004 and ranks at 12th at the moment. I think thats the biggest difference between our two analyses.

Lindsay Davenport equal to Martina Hingis I find surprising based on your criteria. Martina won five majors and appeared in 12 finals, while Lindsay won 3 and appeared in 7. If I understand correctly we are saying that Lindsay's reign at number 1 exceeded Martina's reign by a sufficient margin to compensate for the differences in achievements at majors. My records show Martina Hingis being at number 1 for 209 weeks compared with Lindsay's 98 weeks. Therefore I'm just surprised thath they come out as equal. In the DOT ratings Martina Hingis peaks at 2692 after the French in 1999 while Lindsay is at 2677 after the Australian of 2000. They rank at 9th and 14th respectively.

I hope you find this feedback useful and would appreciate any comments on the DOT ratings when I post them for both Men and Women next week.

Take care and keep in touch

Tim :)
 
Last edited:

CyBorg

Legend
Don't worry to much about the people who criticise without making any attempt of their own or without providing good reason as to why your list may misrepresent achievemnet aside from their personal views.

I don't understand these kinds of comments. The purpose of posting this list on a discussion board is that it would facilitate discussion and criticism.

Criticism is not inherently negative. It is constructive.
 

Wuornos

Professional
I don't understand these kinds of comments. The purpose of posting this list on a discussion board is that it would facilitate discussion and criticism.

Criticism is not inherently negative. It is constructive.

Hi Cyborg.

Yes that's exactly what I meant when I said 'or without providing good reason as to why your list may misrepresent achievemnet aside from their personal views'.

It's the reasons that make the criticism positive and that's great. Your post was good earlier in this thread, you stated the reasons for your thoughts.

Post #2 by the 'annointedone' was I thought negative. It was just personal opinion with no reasoning and no discussion. What was Chaog meant to do with that?

Personally I think we agree with each other on this. I think these posts are just a misunderstanding in what was meant.

Take care.

Tim
 

Wuornos

Professional
The computer ranking system is incredibly fallacious. I'm also wondering how you weighed tournaments and whether you took winning percentages into consideration (another highly fallacious aspect for consideration).

I think that the best system would be to weigh tournaments based on qualities of draws from year to year and assign an independent year-end ranking for each player based on the added data. What's left are longevity and peak factors, which can be balanced.

Regardless, it is easy to criticize and much harder to actually sit down and formula one's own list. You have done that and that's quite admirable.

Hi Cyborg

This is along similar lines to what I am currently doing with the DOT Ratings which I will post next week. The only difference is I look for peak Rating withina career allowing for Tournament weighting and population distribution and don't worry to much about year end rankings.

I would value your feedbak when I do post the DOT Ratings. Either Positive or Negative. :)

Tim
 
Last edited:

Wuornos

Professional
Thanks, but the system rewards both dominance and longevity equally.

Hi Chaog

I think your definition of dominanace is perhaps different to what other people mean.

Your system is based upon total achievement, and there's nothing wrong with that, whether it be total majors won or total number of years at number 1.

When I talk about domination I am talking about the concentration of achievement within a limitd time scale.

To make this point clearer. Total Achievement is like total distance travelled in a car. You might measure that in miles or perhaps by another measure like number of state lines crossed. In your analysis majors is miles and years at no 1 is the number of state lines crossed.

Domination, however, is more akin to speed. It's how many majors you clock within limted time constraits. The numebr at number 1 are more impressive if they are consecutive, again representing the rate at which the lines are being crossed or again a rather perverse measure of speed if you will.

Some like myself will then take this one point further and say some majors are worth more than others. Surprisingly this can be measured by looking at the dominanace as whole of the group who populate the latter stages of an event of a hort number of recent years. E.G the Australian Open of the late 1970s and early 1980s has less value than at other times in its history.

Finally I am perhaps alone in adjusting my calculations based on the shape of the population and giving credit to players when resuts suggest that top group of players results are more distant from the population median. I am not going to descibe how this can be calculated independently here as I have already had a number of debates on this topic on this very site, where people insist it can't be done when in fact it is a very basic piece of statistical probability theory and its possibility is not really a debateable point.

Anyway this post started simply on the point of what is meant by domination, so I've probably said enough.

Take care,

Tim
 

chaognosis

Semi-Pro
Hi Chaog
Your first three players from the open era are Margaret Court, Steffi Graff and Martina Navratilova. This is quite amazing as my output shows, Margaret Court 2749, Steffi Graff 2749 and Martina Navratilova 2748 at their peaks. Actually Margaret Court was 0.1 point higher than Graff but I round the ratings to the nearest point. So a great start there.

Very interesting, indeed!

Lindsay Davenport equal to Martina Hingis I find surprising based on your criteria. Martina won five majors and appeared in 12 finals, while Lindsay won 3 and appeared in 7. If I understand correctly we are saying that Lindsay's reign at number 1 exceeded Martina's reign by a sufficient margin to compensate for the differences in achievements at majors. My records show Martina Hingis being at number 1 for 209 weeks compared with Lindsay's 98 weeks. Therefore I'm just surprised thath they come out as equal. In the DOT ratings Martina Hingis peaks at 2692 after the French in 1999 while Lindsay is at 2677 after the Australian of 2000. They rank at 9th and 14th respectively.

The difference is that my formula used year-end No. 1 standings rather than weeks, and Davenport finished one more year on top of the rankings than Hingis did. You make a very good point that, in modern times, absolute weeks at No. 1 may be a better indicator of dominance than year-end No. 1 status. The reason I opted for the year-end measure is that we don't have week-by-week data for most of the history of tennis, and I wanted a basis for comparison that would extend back as far as possible (in this case, to 1913). The system is far from perfect--like you, I am always trying to come up with new and better ways of looking at the data we have--and my ultimate, and probably impossible, goal is to find some acceptable standard by which we can compare players from all eras... in particular, one that can be supported with reference to the best attested sources rather than just my own opinion and observation.

I am looking forward to your next project.
 

chaognosis

Semi-Pro
Hi Chaog

I think your definition of dominanace is perhaps different to what other people mean.

Your system is based upon total achievement, and there's nothing wrong with that, whether it be total majors won or total number of years at number 1.

When I talk about domination I am talking about the concentration of achievement within a limitd time scale.

Yes, you are right. I was objecting to the claim that my system only rewards longevity, which is not true. As you say, it is based upon total achievement--both dominance (winning much in a short amount of time) and longevity (being around for a long amount of time) contribute to the career totals. But you are correct that my formula didn't give any special weight to "speed"... in other words, a player who wins two majors a year for three years comes out the same as a player who wins one major a year for six years (in such cases it would be up to the year-end No. 1 data to break the tie). I really wish there were some way to do an ELO-like survey for the whole history of tennis, as it's probably the most interesting ranking system I've come across.
 

SgtJohn

Rookie
Chaog, I see what you wanted to achieve, using only widely accepted indicators might be a way to make a 'least common denominator' that most people would agree with...

...that is, it could work in an ideal world, which the world of tennis history is not...I'm afraid I have to agree with CyBorg's criticisms here. Actually I think the way you select your indicators is flawed in that they are a posteriori seen as widely accepted...but were not at the time.
Even if the 'Grand Slam' remained the ultimate achievement from 1938 on, the Australian and the French clearly were not 'indicators' for greatness during most of tennis history. Tons and tons of quotes can confirm that, as late as in the mid 1980s. It's the same reasoning about the US Pro: as it was a great tournament during the last years of the Pro Era ('63-'67), people have come to remember it as a major. But if you could go back in time until, say, 1962, and ask people about it, they would tell you it was a terrible event, very depleted. Then why should it be in the list as a widely accepted indicator?
The discussion above concerning the ATP rankings goes the same way: they might be widely accepted now, because the ATP 'won the war' when it took over the tour in 1990, and thus its ranking was retrospectively considered as the reference, but until the mid 80s, the Grand Prix and WCT were almost as significant...

I can see you wanted to design a list that was free of speculations and subjectivity. I think it is unfortunately impossible for tennis. The problem with your 'accepted indicators' approach is that these indicators have been extremely changing over time. It's hard to compare Myers's rankings, that put the stress on major victories and 'intelligence on court', and the ATP's machine-like, statistical system.

For sure, historical rigor is on your side, but I think the truth is closer to the subjective approach than to this one.

Jonathan
 

joe sch

Legend
I also appreciate the time you have taken to produce another way to look at the all-time rankings. It is also outstanding that you did so for both mens and ladies rankings. I have not yet taken the time to read all the post to figure out how these indicators work, would it be possible to post your algorithms and data ? By doing so, different people could offer improvements or find possible flaws in the algorithms and data and the results would improve over time and become more universally accepted. This is basically why open source software becomes so effective

Thanks again for your work in this area !
 

chaognosis

Semi-Pro
Thanks for your insightful comments, Jonathan. You see exactly what I was trying to do with these rankings--and the difference between "historical rigor" and "the truth" is well noted. My personal rankings, certainly, differ from these results in some significant ways: Laver is ahead of Gonzales and on par with Tilden, Borg is much higher, etc. Nevertheless, I do think this is an interesting and important sort of exercise to do. No, tennis isn't an ideal world... but is there an ideal world in any sport, or in any field?
 

CyBorg

Legend
Hi Cyborg.

Yes that's exactly what I meant when I said 'or without providing good reason as to why your list may misrepresent achievemnet aside from their personal views'.

It's the reasons that make the criticism positive and that's great. Your post was good earlier in this thread, you stated the reasons for your thoughts.

Post #2 by the 'annointedone' was I thought negative. It was just personal opinion with no reasoning and no discussion. What was Chaog meant to do with that?

Personally I think we agree with each other on this. I think these posts are just a misunderstanding in what was meant.

Take care.

Tim

I see. I completely missed antoinedone's comment, which is why I was confused. Thanks for clarifying.
 

CyBorg

Legend
Hi Cyborg

This is along similar lines to what I am currently doing with the DOT Ratings which I will post next week. The only difference is I look for peak Rating withina career allowing for Tournament weighting and population distribution and don't worry to much about year end rankings.

I would value your feedbak when I do post the DOT Ratings. Either Positive or Negative. :)

Tim

This sounds extremely cool.
 

Mingo

New User
Nice list Chaog. I like the attempt at an objective calculation for rankings.

One factor I would like to see added to such a formula accounts for the ability to win on multiple surfaces. I know it is subjective and my opinion, but I always think that the player who wins titles on clay, grass and hardcourt is better than someone who dominates a single surface.
 
Top