The Rating of Singles Tennis Players Past and Present

[Wuornos]

Most of you know how I evaluate player performance but by way of a brief summary here are the salient points:

1 Ratings are calculated using a Multi-iterative process with initial iterations used to calculate initial player weightings which then feed into the formulae that derive event weightings. Further iterations are then used to derive player ratings given the event weightings that were derived in the initial iterations.
2 Ratings are not calculated from a single year but instead use a diminishing return on results with results from more distant tournaments carrying a lesser weighting than more recent performances.
3. Players who remove themselves from the main player population cannot have an accurate rating calculated due to a lack of relevant reference points. This most importantly applies to the pre open Pro Circuit.
4 Players who have retired retain their rating which declines overtime given the distance in the past of their most recent results. This is important given the possibility of return to match play and the impact this would have on an initial iteration event weighting.
5 The result of the above provides an estimate of the player’s peak playing standard as measured by their distance from the 50th ranked player at any point in time. While within the top 5 in the game this movement overtime tends to change significantly given ‘strong’ and ‘weak’ periods the level at number 50 while still changing tends to remain more constant.
6 This effect is entirely expected given a normal statistical distribution of player’s standards. E.g. the following would be a typical distribution at the top of such a sample:

1 2700
2 2670
3 2652
4 2640
5 2630
6 2622
7 2615
8 2610
9 2605
10 2600
11 2596
12 2592
13 2589
14 2585
15 2582
16 2580
17 2577
18 2574
19 2572
20 2570
21 2568
22 2566
23 2564
24 2562
25 2560
26 2559
27 2557
28 2555
29 2554
30 2552
31 2551
32 2549
33 2548
34 2547
35 2546
36 2544
37 2543
38 2542
39 2541
40 2540
41 2539
42 2538
43 2537
44 2536
45 2535
46 2534
47 2533
48 2532
49 2531
50 2530

If we were measuring the standard as a distance from the 4th ranked player and this player was removed from the game/distribution the inflation in ratings would show an increase of 4th place player – 5th place player or 10 points. If however we are measuring from the 50th ranked player and one player is removed from the higher positions we would only see an inflation of 1 point.

I hope this explains what the ratings mean and why sometimes players who are currently inactive remain in the ratings albeit with a gradually declining score.

The ratings are an attempt to measure a players playing standard given the conditions and equipment that prevail at the time.

The all time top 10 list simply shows the peak rating achieved for each player concerned.

 

[Wuornos]

Current Men's Singles Ratings

1 Rafael Nadal 2786 (2772) 10 3 3 4
2 Novak Djokovic 2771 (2707) 3 2 7 6
3 Roger Federer 2760 (2781) 16 7 5 5
4 Andy Murray 2716 (2692) 0 3 4 2
5 Robin Söderling 2653 (2670) 0 2 0 4
6 Juan Martín del Potro 2644 (2694) 1 0 1 2
7 Jo-Wilfried Tsonga 2636 (2638 0 1 2 2
8 Tomáš Berdych 2630 (2660) 0 1 1 2
9 David Ferrer 2595 (2517) 0 0 2 3
10 Mikhail Youzhny 2594 (2550) 0 0 2 2

Correct up to and including Wimbledon of 2011.

1 Initial figure shows current rating.
2 Bracketed figure shows rating at this point last year.
3 Following figures show major wins, losing major finals, losing major semi finals and losing major quarter finals respectively.

 

[Wuornos]

Current Women's Singles Rating

1 Kim Clijsters 2687 (2620) 4 4 7 3
2 Serena Williams 2686 (2738 13 3 3 12
3 Li Na 2647 (2580) 1 1 1 3
4 Francesca Schiavone 2634 (2637) 1 1 0 5
5 Vera Zvonareva 2631 (2603) 0 2 2 1
6 Venus Williams 2604 (2637) 7 7 5 14
7 Caroline Wozniacki 2603 (2577) 0 1 2 1
8 Petra Kvitová 2590 (2541) 1 0 1 1
9 Maria Sharapova 2586 (2612) 3 2 7 3
10 Svetlana Kuznetsova 2574 (2616) 2 2 1 7

Sorry Caroline.

Correct up to and including Wimbledonof 2011.

1 Initial figure shows current rating.
2 Bracketed figure shows rating at this point last year.
3 Following figures show major wins, losing major finals, losing major semi finals and losing major quarter finals respectively.

 

[Wuornos]

Top 10 Men's Singles Ratings Post WWII

1 Roger Federer 2797 (2010) 16 7 5 5
2 Rafael Nadal 2789 (2011) 10 3 3 4
3 Rod Laver 2776 (1962) 11 6 1 2
4 Ivan Lendl 2774 (1988 8 11 9 6
5 Novak Djokovic 2771 (2011) 3 2 7 6
6 Roy Emerson 2768 (1965) 12 3 4 18
7 Mats Wilander 2763 (1988 7 4 3 6
8 Björn Borg 2762 (1980) 11 5 1 4
9 Frank Sedgman 2762 (1952) 5 3 2 4
10 Tony Trabert 2760 (1955) 5 0 2 2
11 Pete Sampras 2757 (1995) 14 4 5 6
12 Ashley Cooper 2753 (1958 4 2 3 4
13 John McEnroe 2751 (1984) 7 4 8 7
14 Jim Courier 2750 (1993) 4 3 4 4
15 Ken Rosewall 2742 (1956) 8 8 9 5
16 Lew Hoad 2741 (1957) 4 2 3 6
17 Stefan Edberg 2741 (1993) 6 5 8 7
18 Boris Becker 2737 (1991) 6 4 8 5
19 Fred Stolle 2733 (1965) 2 6 3 6
20 Neale Fraser 2731 (1960) 3 4 9 5
21 Jaroslav Drobný 2725 (1952) 3 5 6 3
22 Jimmy Connors 2720 (1975) 8 7 16 10
23 Vic Seixas 2717 (1955) 2 3 5 10
24 Andy Murray 2716 (2011) 0 3 4 2
25 Manuel Santana 2714 (1966) 4 0 4 2
26 Pat Cash 2713 (1988 1 2 2 4
27 Andre Agassi 2708 (1995) 8 7 11 10
28 Ken McGregor 2703 (1952) 1 3 2 1
29 Dick Savitt 2703 (1951) 2 0 3 6
30 Juan Martín del Potro 2703 (2009) 1 0 1 2
31 John Newcombe 2703 (1971) 7 3 6 12
32 Tony Roche 2701 (1967) 1 5 7 8



Correct up to and including Wimbledon of 2011.

Now updated to show all players achieving a peak playing performance in excess of 2700 points since WWII.

1 Initial figure shows peak rating.
2 Bracketed figure shows year peak rating was achieved.
3 Following figures show major wins, losing major finals, losing major semi finals and losing major quarter finals respectively.

 

[Wuornos]

Top 10 Women's Peak Single's Rating Post WWII

1 Martina Navratilova 2809 (1986) 18 14 12 9
2 Steffi Graf 2806 (1989) 22 9 6 5
3 Monica Seles 2805 (1993) 9 4 5 13
4 Margaret Smith Court 2797 (1965) 24 5 7 7
5 Chris Evert 2790 (1983) 18 16 18 2
6 Serena Williams 2785 (2003) 13 3 3 12
7 Maureen Connolly 2779 (1954) 9 0 0 0
8 Venus Williams 2770 (2002) 7 7 5 14
9 Billie Jean King 2768 (1969) 12 6 8 14
10 Martina Hingis 2763 (1999) 5 7 7 5
11 Arantxa Sánchez Vicario 2757 (1995) 4 8 10 13
12 Louise Brough Clapp 2756 (1950) 6 6 9 4
13 Maria Bueno 2755 (1965) 7 5 8 9
14 Doris Hart 2754 (1951) 6 12 6 4
15 Ann Haydon Jones 2752 (1969) 3 6 12 9
16 Justine Henin 2747 (2007) 7 5 5 2
17 Hana Mandlíková 2747 (1986) 4 4 6 9
18 Evonne Goolagong 2745 (1975) 7 11 4 4
19 Lindsay Davenport 2745 (2000) 3 4 11 13
20 Margaret Osborne Dupont 2743 (1951) 6 3 4 7
21 Jennifer Capriati 2739 (2002) 3 0 10 10
22 Gabriela Sabatini 2735 (1991) 1 2 15 10
23 Amélie Mauresmo 2727 (2006) 2 1 5 9
24 Maria Sharapova 2725 (2008 3 2 7 3
25 Althea Gibson 2724 (1958 5 2 0 2
26 Lesley Turner Bowrey 2723 (1965) 2 4 6 10
27 Nancy Richey Gunter 2720 (1967) 2 4 7 9
28 Shirley Fry 2717 (1957) 4 4 6 6
29 Tracy Austin 2715 (1982) 2 0 3 8
30 Darlene Hard 2712 (1962) 3 4 5 5
31 Jana Novotná 2711 (1998 1 3 5 13
32 Helena Suková 2707 (1986) 0 4 3 11
33 Kim Clijsters 2705 (2006) 4 4 7 3
34 Andrea Jaeger 2700 (1983) 0 2 5 3
35 Mary Joe Fernández 2700 (1993) 0 3 6 8

Correct up to and including Wimbledon of 2011.

Now updated to show all players achieving a peak playing performance in excess of 2700 points since WWII.

1 Initial figure shows peak rating.
2 Bracketed figure shows year peak rating was achieved.
3 Following figures show major wins, losing major finals, losing major semi finals and losing major quarter finals respectively.

 

[Bursztyn]

You seem to develop a kind of system to measure players dominance over their peers. In this respect your ranking is similar to ATP ranking based on performance in the previous 52 weeks, hovewer your ranking is much more complicated and allows to look at players achievements from different perspective.

What I find intriguing at first sight is that players rating is influenced by his/her results from the distant past, although the more distant in time are past results the less they influence the overall rating. I would like to ask what kind of function have you used to link the amount of time that passed from the past achievements and the rate of decrease of their value. It seems to me that your final ratings depend on such a function, and the choice of this function is somewhat arbitrary.

So it is not surprising that Nadal is ahead of Sampras in your rating list (despite the fact that Sampras won more GS titles), because Sampras has never been as dominant as Nadal was last year.

 

[robow7]

IMO, you have a system that rates Sampras at no. 10 and exclude all together the individual with the most professional tournament wins in the Open era from your top 10, therefore your system is deeply flawed.

 

[pc1]

Wuornos,

The other flaws is that I believe it rates majors too high since the players pre 1968 could not play the majors. Also with boycotts and the fact that players often skipped majors in the 1970's and 1980's does not make everything a level playing field.

For example Ken Rosewall's actual stats are 136 tournament victories and if you include Pro Majors, 23 total majors.

Rod Laver has 199 tournament victories and 19 total majors. Shouldn't dominance include a calendar year Grand Slam of which Laver had three if you include the 1967 Pro Grand Slam?

I do appreciate the attempt and I respect your work.

Jimmy Connors has I believe actually 148 tournaments won and Lendl 146.

Bill Tilden at one point was virtually unbeatable and during one period he won 138 of 192 tournaments played with a 907-62 match record. He won almost every major he played for a period but often decided not to take the very long oversea trips to go to the Australian, the French and Wimbledon because it would take many many weeks. It's probable he would have won at least one calendar year Grand Slam, perhaps more. In total Tilden won over 160 tournaments in his career. This is only below Laver.

 

[NLBwell]

I really like and appreciate your efforts.
From your list I'm assuming that you are excluding the pro tournaments prior to 1968.
To get a gauge on things, at what time point does a tournament win decrease to half its value with age?

 

[fed_rulz]

IMO, you have a system that rates Sampras at no. 10 and exclude all together the individual with the most professional tournament wins in the Open era from your top 10, therefore your system is deeply flawed.

he has explained his methodology. Aside from your apparent liking for Sampras, do you have any valid objection as to why Sampras should be ranked higher in terms of peak rating achieved? Sampras was never as dominant as Federer or Nadal or even Wilander (he won 3 slams in 87), so his rating makes sense. Same applies for your argument about the individual with most professional tournament wins.

 

[pc1]

he has explained his methodology. Aside from your apparent liking for Sampras, do you have any valid objection as to why Sampras should be ranked higher in terms of peak rating achieved? Sampras was never as dominant as Federer or Nadal or even Wilander (he won 3 slams in 87), so his rating makes sense. Same applies for your argument about the individual with most professional tournament wins.

Actually Connors was very dominant in 1974, winning 3 out of 3 majors and 99 of 103 matches if that's the individual you're talking about. I think he won 14 or 15 tournaments. Too tired to look it up.

 

[robow7]

he has explained his methodology. Aside from your apparent liking for Sampras, do you have any valid objection as to why Sampras should be ranked higher in terms of peak rating achieved? Sampras was never as dominant as Federer or Nadal or even Wilander (he won 3 slams in 87), so his rating makes sense. Same applies for your argument about the individual with most professional tournament wins.

I'm no great fan of Sampras any more than I am of the Fed. But when you put Tony Trabert above a Sampras, don't even include Connors after 74" and 13 years in the top 3, and if mostly built upon dominance over the period of a year or two, you have Laver with two true grand slams and Nadal none, somewhere there's a problem.

 

[Bursztyn]

The problem is that when evaluating players performance you have to make a lot of arbitrary decisions, and your decisions influence significantly the outcome of your model. Hovewer it is interesting to see how the results change when you alter your arbitrary decisions. By changing initial settings you can get a new perspective on things. From the final ratings that you presented it is quite obvious that past results achieved more than 1 year ago are still important, otherwise Serena wouldn't be number one player.

 

[Bursztyn]

It seems to me that past results achieved more than 5 years ago are not important in your rating system. Sampras highest rating was in 1995, and he won his first GS title in 1990. However it was the only GS title he won between 1990 - 1992. So it looks as if the results from the last three years were the most important in evaluating players rating.

 

[Moose Malloy]

Wuornos,

The other flaws is that I believe it rates majors too high since the players pre 1968 could not play the majors. Also with boycotts and the fact that players often skipped majors in the 1970's and 1980's does not make everything a level playing field.

For example Ken Rosewall's actual stats are 136 tournament victories and if you include Pro Majors, 23 total majors.

Rod Laver has 199 tournament victories and 19 total majors. Shouldn't dominance include a calendar year Grand Slam of which Laver had three if you include the 1967 Pro Grand Slam?

I do appreciate the attempt and I respect your work.

Jimmy Connors has I believe actually 148 tournaments won and Lendl 146.

Its like groundhog day around here. wuornos has been posting his rankings for years here & has explained how they work, yet the same questions keep arising over & over again. and he has tinkered with it to take a lot of that into consideration(search his past threads)

his top 10 isn't all that surprising, those names seemed to beat the most quality players during their runs(connors, however dominant, really didn't have a lot of them in '74, we all know how he dodged wct tour, etc)

and 'quality' wins in this discussion isn't just based on atp rankings/players who have won majors etc, but beating players who have high elo rankings which takes time to develop.

whether I agree or disagree with his results I appreciate his work which must have been very time consuming. I wish all those that criticize his methods have at least taken the time to understand them(which may be hard to do since I'm guessing most here don't have the same background/training wuornos does when it comes to statistics - I certainly don't)

Bill Tilden at one point was virtually unbeatable and during one period

his rankings say 'post WWII'

 

[pc1]

Moose,
I thnk Wuornos has done a wonderful job of rating players and I want to let people know I appreciate his work. Sorry for the errors. I guess I missed it in Wuornos post. My bad.

 

[Limpinhitter]

1 Roger Federer 2797 (2010) 16 7 5 4
2 Rafael Nadal 2789 (2011) 10 2 3 4
3 Rod Laver 2776 (1962) 11 6 1 2
4 Ivan Lendl 2774 (1988 8 11 9 6
5 Roy Emerson 2768 (1965) 12 3 4 18
6 Mats Wilander 2763 (1988 7 4 3 6
7 Björn Borg 2762 (1980) 11 5 1 4
8 Frank Sedgman 2762 (1952) 5 3 2 4
9 Tony Trabert 2760 (1955) 5 0 2 2
10 Pete Sampras 2757 (1995) 14 4 5 6



Correct up to and including French Open of 2011.

1 Initial figure shows peak rating.
2 Bracketed figure shows year peak rating was achieved.
3 Following figures show major wins, losing major finals, losing major semi finals and losing major quarter finals respectively.

Trabert and Sedgman above Kramer, Pancho and Rosewall? Wilander above Borg? I do love Emmo, but, above Pete and Borg?

BACK TO THE DRAWING BOARD!

 

[NLBwell]

Emerson dominated amateur tennis the way Federer dominated the ATP a few years ago.
Pancho and Rosewall were pros most of the time, so much of their career is excluded. (At least if appears he didn't try to put in the pro tours prior to open tennis)
Sampras' greatness was in being year-end number one for 6 (or however many) years in a row. He didn't have a dominant year the way Federer or even McEnroe had. Wilander above Borg is only by one point out of 2763, so virtually even. I do think Wilander is under-appreciated, but, as I said before, I'm wonderiing what the half-life of a tournament win is. Changing it slightly will change the answers.
Nice to see some of the old guys not mentioned much like Sedgeman and Trabert get some notoriety.

 

[Chopin]

Trabert and Sedgman above Kramer, Pancho and Rosewall? Wilander above Borg? I do love Emmo, but, above Pete and Borg?

BACK TO THE DRAWING BOARD!

I said the same thing the other day, Limpin'. Trabert and Sedgman, but no Kramer and Pancho? Back to the drawing board is right.

 

[CyBorg]

Trabert and Sedgman above Kramer, Pancho and Rosewall? Wilander above Borg? I do love Emmo, but, above Pete and Borg?

BACK TO THE DRAWING BOARD!

I'm guessing that Pancho doesn't get much love here because he played in majors only a little bit as an amateur.

 

[Harry_Wild]

I would have to think about Sampass vs. Connors vs. McEnroe vs. Aggasi. Aggasi completed the Grand Slams while either of the others has done that. Lendi did not complete either since he could not win U.S. Open.

Tough to rank the top ten since if you live 200 years you could do it but now it is alway rank the current top players to be on top but you do not know the past so it so tough to see how they compare against one another.

 

[tennis_pro]

Emerson dominated amateur tennis the way Federer dominated the ATP a few years ago.
Pancho and Rosewall were pros most of the time, so much of their career is excluded. (At least if appears he didn't try to put in the pro tours prior to open tennis)
Sampras' greatness was in being year-end number one for 6 (or however many) years in a row. He didn't have a dominant year the way Federer or even McEnroe had. Wilander above Borg is only by one point out of 2763, so virtually even. I do think Wilander is under-appreciated, but, as I said before, I'm wonderiing what the half-life of a tournament win is. Changing it slightly will change the answers.
Nice to see some of the old guys not mentioned much like Sedgeman and Trabert get some notoriety.

Emerson domination shouldn't be compared to Federer's as there were no double circuits in 2003-2007. I bet Emerson wouldn'tve won half the Slams between 1963 and 1967 if he had to face Laver and Rosewall on a consistant basis.

 

[NLBwell]

Emerson domination shouldn't be compared to Federer's as there were no double circuits in 2003-2007. I bet Emerson wouldn'tve won half the Slams between 1963 and 1967 if he had to face Laver and Rosewall on a consistant basis.

I specified amateur.

 

[NLBwell]

Emerson domination shouldn't be compared to Federer's as there were no double circuits in 2003-2007. I bet Emerson wouldn'tve won half the Slams between 1963 and 1967 if he had to face Laver and Rosewall on a consistant basis.

I specified amateur.
I was explaining why he would be ranked so high on the list.

 

[Limpinhitter]

I'm guessing that Pancho doesn't get much love here because he played in majors only a little bit as an amateur.

And as a pro in his 40's when he was still beating some of the best players in the world.

 

[pc1]

I'm guessing that Pancho doesn't get much love here because he played in majors only a little bit as an amateur.

I guess it was hard for Gonzalez to be rated in this system since he was in the Pros since the late 1940's.

 

[Wuornos]

I guess it was hard for Gonzalez to be rated in this system since he was in the Pros since the late 1940's.

Yes absolutely right pc1. As stated in my first post:

3. Players who remove themselves from the main player population cannot have an accurate rating calculated due to a lack of relevant reference points. This most importantly applies to the pre open Pro Circuit.

I have never yet found a work around to tackle this irritating problem and so cannot include the results of the pre open era Pro's. Gonzales had such a short amateur career, i.e. within the main player population, before turning pro that his rating probably does not accurately reflect his peak playing standard.

If anyone has any ideas of how this problem can be tackled, I would love to hear.

Tim

 

[Wuornos]

You seem to develop a kind of system to measure players dominance over their peers. In this respect your ranking is similar to ATP ranking based on performance in the previous 52 weeks, hovewer your ranking is much more complicated and allows to look at players achievements from different perspective.

What I find intriguing at first sight is that players rating is influenced by his/her results from the distant past, although the more distant in time are past results the less they influence the overall rating. I would like to ask what kind of function have you used to link the amount of time that passed from the past achievements and the rate of decrease of their value. It seems to me that your final ratings depend on such a function, and the choice of this function is somewhat arbitrary.

So it is not surprising that Nadal is ahead of Sampras in your rating list (despite the fact that Sampras won more GS titles), because Sampras has never been as dominant as Nadal was last year.

Hi Bursztyn

Yes you are right the function of decay in a player performance overtime is a critical aspect of the algorithms used in the calculations.

However I have done my very best to make sure any arbitrary values are kept to an absolute minimum within the calculations.

In the case you refer to it is possible to look at results of players who left the game for a significant period and then returned and to monitor results both pre and post break. Some players decline faster than others. A recent example would be Clijsters and Henin. Clijsters had a peak rating of 2705 prior to break while her current rating stands at 2689. Quite a small fall while Henin had a pre break peak of 2747 and now has a rating of 2580. Admittedly these are two extremes of the decline in practise.

It is by recording this decline across numerous players who have left and returned to the game that we can derive an average decay in playing standard given the duration of absence.

This of course in turn shows the weighting past performances should have for their own sake in that it gives an estimate of what a more distant result is likely to mean in relation to playing standard of the current day irrespective of more recent results.

In this way I ensured that this aspect of the calculation was not arbitrary.

Good question.

Thanks

Tim

 

[Wuornos]

Just to say thanks for all the intersting posts.

I do listen to the various points made and as a result I do plan to revisit the ratings calculated for male players during the pre open era.

I don't think there is anything wrong with the basic methodology but enough has been said to make me want to go back and check the actual formulae being used to make sure that none contain any keying errors which in turn impact on the outputs.

I hope to have completed this by the end of July and will post my findings along with new ratings at that time.

If anyone can suggest any good sources of data for results pre WWII I'd love to hear as it would be great to derive ratings for players like Bill Tilden etc.

Thanks

Tim

 

[Wuornos]

Its like groundhog day around here. wuornos has been posting his rankings for years here & has explained how they work, yet the same questions keep arising over & over again. and he has tinkered with it to take a lot of that into consideration(search his past threads)

his top 10 isn't all that surprising, those names seemed to beat the most quality players during their runs(connors, however dominant, really didn't have a lot of them in '74, we all know how he dodged wct tour, etc)

and 'quality' wins in this discussion isn't just based on atp rankings/players who have won majors etc, but beating players who have high elo rankings which takes time to develop.

whether I agree or disagree with his results I appreciate his work which must have been very time consuming. I wish all those that criticize his methods have at least taken the time to understand them(which may be hard to do since I'm guessing most here don't have the same background/training wuornos does when it comes to statistics - I certainly don't)



his rankings say 'post WWII'

Thanks Moose

Appreciated

Tim

 

[Datacipher]

whether I agree or disagree with his results I appreciate his work which must have been very time consuming. I wish all those that criticize his methods have at least taken the time to understand them(which may be hard to do since I'm guessing most here don't have the same background/training wuornos does when it comes to statistics - I certainly don't)
I'

Well that in itself is a huge problem isn't it? Without understanding the methodology thoroughly, the list is completely meaningless. My suggestion would be for Wournos to write out in explicit detail, the exact calculations which go into his rankings, and then have that posted somewhere. Anyone so inclined can then go read it, and a great deal of hassle can be avoided.

 

[Wuornos]

Well that in itself is a huge problem isn't it? Without understanding the methodology thoroughly, the list is completely meaningless. My suggestion would be for Wournos to write out in explicit detail, the exact calculations which go into his rankings, and then have that posted somewhere. Anyone so inclined can then go read it, and a great deal of hassle can be avoided.

Thanks Datacipher.

Yes I can understand exactly where you are coming from on this. I posted in a lot more depth regarding methodology a couple of years ago and unfortunately it didn’t really help. As the knowledge on this form is more tennis related than maths related it just ended in protracted discussions and criticism about aspects which are the basic building blocks of probability theory and have been known to mathematics for the past two millennia. I really don’t think trying to go on and explain the Bayesian elements to readers of this forum would be of any great interest to them.

On a more positive note I am surprised that no one has mentioned that the proof of the rating lies not in the rating of past players but in its predictive value of current players performances where it differs significantly from the official rankings. This is something that everyone can understand and people within this forum are extremely knowledgeable on.

E.g. at Wimbledon at the moment Serena has returned following injury. The official rankings say she should be ranked at 25 while the expert opinion of the Wimbledon seeding committee seeded her 7th. The rankings I calculate even allowing for a decline due to inactivity say she should still be ranked at #1. With such marked differences her performance of Wimbledon would act as a barometer of the effectiveness of each methodology.

Of course one such example in isolation is not sufficient evidence in itself but a prolonged series of such ‘acid tests’ would provide evidence of the effectiveness or not of the formulae I use.

Thanks for your interest.

Tim

The problem I have now is watching todays match between Serena and Marion Bartoli. I genuine like Marion and love to watch her play and want her to win but I also want Serena to win because her success supports my rating methodology. I suppose mixed emotions for me whichever way the match goes.

 

[pc1]

Yes absolutely right pc1. As stated in my first post:

3. Players who remove themselves from the main player population cannot have an accurate rating calculated due to a lack of relevant reference points. This most importantly applies to the pre open Pro Circuit.

I have never yet found a work around to tackle this irritating problem and so cannot include the results of the pre open era Pro's. Gonzales had such a short amateur career, i.e. within the main player population, before turning pro that his rating probably does not accurately reflect his peak playing standard.

If anyone has any ideas of how this problem can be tackled, I would love to hear.

Tim

Tim,

To be accurate players like Gonzalez did not remove themselves from the main player population, the media removed him and other top pros as far as coverage and recording of records were concerned. The top amateur players wanted the money and the chance to play the top players. There wasn't much newspaper coverage of the pros. The records are there to be found and clearly the pros were far superior to the amateurs. There is no doubt about that.

 

[Wuornos]

Tim,

To be accurate players like Gonzalez did not remove themselves from the main player population, the media removed him and other top pros as far as coverage and recording of records were concerned. The top amateur players wanted the money and the chance to play the top players. There wasn't much newspaper coverage of the pros. The records are there to be found and clearly the pros were far superior to the amateurs. There is no doubt about that.

Hi pc1.

Yes I stand corrected. However the problem remains that with so few cross over points, i.e. when a player moves from amateur to pros, that there is insufficient data to splice the two populations together. Hence the exclusion of the pro players from this era in my ratings.

The main point of my post was to ask if anyone has any idea of how this problem can be surmounted.

Tim

 

[CyBorg]

I think that stats are useful aids. However tennis history is full of too many changes and differences between organizations of tours. So numbers can never give us the answers we seek. They can illustrate and clarify points of view though.

 

[Q&M son]

I think that stats are useful aids. However tennis history is full of too many changes and differences between organizations of tours. So numbers can never give us the answers we seek. They can illustrate and clarify points of view though.

Agree with CyBorg.
And happy that Tim is here, still in this forums.

 

[Wuornos]

Figures updated to include all results up to and including Wimbledon of 2011.

Top 10 players posts now show all players achieving a peak rating in excess of 2700 since WWII.

Tim

 

[Wuornos]

I think that stats are useful aids. However tennis history is full of too many changes and differences between organizations of tours. So numbers can never give us the answers we seek. They can illustrate and clarify points of view though.

Many thanks Cyborg I agree completely. At best they just give a single point of view that is as objective as I can possibly make it. They can never tell the full story.

Thanks

Tim

 

[pc1]

Figures updated to include all results up to and including Wimbledon of 2011.

Top 10 players posts now show all players achieving a peak rating in excess of 2700 since WWII.

Tim

Tim,

I comment you on always trying to improve your system. It's to be admired.

 

[hoodjem]

Goundhog Day--yes!

I have before stated that, while I don't necessarily agree with Tim's rankings 100%, I do find them stimulating, intriguing, provocative, and worthy of much thought and consideration.