Men's ELO Ratings

Wuornos

Professional
ELO Ratings of Major Winners for last 10 years

Deleted as found unuseful by board members. :cry:Deleted as found unuseful by board members. :cry:
 
Last edited:

Wuornos

Professional
What the numbers mean

Deleted as found unuseful by board members. :cry:
 
Last edited:

PCXL-Fan

Hall of Fame
It would be interesting so see mathematic equations to see how accurate ELO ratings have been at predicting actual tennis matches.

You wouldn't by any chance be able to post those would you Wuornos?
 

msunderland71

New User
ELO rating are great, thanks. Has it ever been calculated for specific surfaces? For example who was the player with the highest rating for clay/grass etc. Also for the 1st thread, what date did they each reach those ratings? Sorry if you've already posted this info.
Thanks
Matt
 

akv89

Hall of Fame
I have no managed to buy a much larger database of results from the open era and incorporated these results into my ELO Ratings system.

The following data show peak career ELO Ratings for all players who have achieved a 2600 ELO Rating based on results within the Open Era up to and including the Australian Open of 2008.


1 Roger Federer 2776
2 Ivan Lendl 2769
3 Rod Laver 2763
4 Pete Sampras 2749
5 Björn Borg 2748
6 Mats Wilander 2739
7 John McEnroe 2736
8 Stefan Edberg 2731
9 Boris Becker 2725
10 Jim Courier 2719
11 Ken Rosewall 2706
12 Jimmy Connors 2703
13 Arthur Ashe 2691
14 Andre Agassi 2687
15 Tony Roche 2681
16 Rafael Nadal 2678
17 John Newcombe 2673
18 Novak Đoković 2670
19 Guillermo Vilas 2665
20 Stan Smith 2662
21 Jan Kodeš 2660
22 Lleyton Hewitt 2658
23 Vitas Gerulaitis 2658
24 Andy Roddick 2656
25 Pat Cash 2654
26 Michael Chang 2652
27 Patrick Rafter 2651
28 Ilie Năstase 2650
29 Miloslav Mečíř 2649
30 Andrés Gimeno 2637
31 Todd Martin 2634
32 Yevgeny Kafelnikov 2633
33 Marat Safin 2633
34 Juan Carlos Ferrero 2632
35 Henri Leconte 2630
36 Tom Okker 2630
37 Roscoe Tanner 2625
38 Michael Stich 2625
39 Nikolay Davydenko 2621
40 Goran Ivanišević 2619
41 David Nalbandian 2617
42 Johan Kriek 2615
43 Manuel Orantes 2614
44 Carlos Moyà 2613
45 Richard Krajicek 2613
46 Kevin Curren 2611
47 Sergi Bruguera 2607
48 Alex Metreveli 2606
49 Cédric Pioline 2604
50 Clark Graebner 2603
51 Yannick Noah 2603
52 Gustavo Kuerten 2601

I will endeavout to keep these data updated quarterly as and when I receive database updates.

Good to see you back. I have familiarized myself with the ELO rating system a bit recently and I must say that it is probably the best rating system that I have seen yet because it factors the strength competition that a person is facing. Thanks very much for the effort.
I can think of only one problem with the ELO system; surface definitely makes a difference in a player's success based on their type of play. Hence, when a player like Roddick loses to a player like Volandri on clay he would see a sharp drop in his ELO rating despite the fact that he is playing a better player on clay.
 

Wuornos

Professional
Elo Ratings as a predictor.

Deleted as found unuseful by board members. :cry:
 
Last edited:

llgc8080

Rookie
Deleted as found unuseful by board members. :cry:


-Sorry Wuornos, i record your material in my disc.
You can't erase it:


Men's ELO Ratings

--------------------------------------------------------------------------------

I have now managed to buy a much larger database of results from the open era and incorporated these results into my ELO Ratings system.

The following data show peak career ELO Ratings for all players who have achieved a 2600 ELO Rating based on results within the Open Era up to and including the Australian Open of 2008.


1 Roger Federer 2776
2 Ivan Lendl 2769
3 Rod Laver 2763
4 Pete Sampras 2749
5 Björn Borg 2748
6 Mats Wilander 2739
7 John McEnroe 2736
8 Stefan Edberg 2731
9 Boris Becker 2725
10 Jim Courier 2719
11 Ken Rosewall 2706
12 Jimmy Connors 2703
13 Arthur Ashe 2691
14 Andre Agassi 2687
15 Tony Roche 2681
16 Rafael Nadal 2678
17 John Newcombe 2673
18 Novak Đoković 2670
19 Guillermo Vilas 2665
20 Stan Smith 2662
21 Jan Kodeš 2660
22 Lleyton Hewitt 2658
23 Vitas Gerulaitis 2658
24 Andy Roddick 2656
25 Pat Cash 2654
26 Michael Chang 2652
27 Patrick Rafter 2651
28 Ilie Năstase 2650
29 Miloslav Mečíř 2649
30 Andrés Gimeno 2637
31 Todd Martin 2634
32 Yevgeny Kafelnikov 2633
33 Marat Safin 2633
34 Juan Carlos Ferrero 2632
35 Henri Leconte 2630
36 Tom Okker 2630
37 Roscoe Tanner 2625
38 Michael Stich 2625
39 Nikolay Davydenko 2621
40 Goran Ivanišević 2619
41 David Nalbandian 2617
42 Johan Kriek 2615
43 Manuel Orantes 2614
44 Carlos Moyà 2613
45 Richard Krajicek 2613
46 Kevin Curren 2611
47 Sergi Bruguera 2607
48 Alex Metreveli 2606
49 Cédric Pioline 2604
50 Clark Graebner 2603
51 Yannick Noah 2603
52 Gustavo Kuerten 2601

I will endeavour to keep these data updated quarterly as and when I receive database updates.

--------------------------------------------------------------------------------
Last edited by Wuornos : 04-05-2008 at 05:35 AM.


Wuornos
View Public Profile
Find More Posts by Wuornos
Add Wuornos to Your Buddy List

02-02-2008, 12:28 AM #2
Wuornos
Semi-Pro


Join Date: Aug 2007
Location: England
Posts: 654 Methodology

--------------------------------------------------------------------------------

ELO Ratings were adopted and used by Professor Arpad Elo of America to calculate performance standards of chess players for the US Chess Federation. They were subsequently adopted by the International Chess Federation to rate all international chess players. They have since been adopted by other sports where two opponents compete against each other in head to head matches.

See links:
http://www.fide.com/ratings/top.phtml?list=men
http://www.eloratings.net/


ELO Ratings cannot be used to compare players from different eras unless the assumption is made that the standard of the general population is a constant. If this assumption is made it allows for the quality of opposition faced and the results achieved. As our underlying assumption has always been in all ratings produced that the standard of the general population has remained constant I think these ratings are statistically sounder than anything that has gone before regarding cross era comparison of players.

Please note that by assuming the general population remains constant we are not assuming the standard at the top of the game remains constant and therefore credit will be given to players who play frequently against other highly rated players.


For a fuller explanation of ELO Ratings and how they work please see :
http://en.wikipedia.org/wiki/Elo_rating_system


I will attempt to keep these data updated each time I receive the updated tournament data which should be quarterly, amending both rating list and commentary.



Regards

Tim


Wuornos
View Public Profile
Find More Posts by Wuornos
Add Wuornos to Your Buddy List

02-02-2008, 12:29 AM #3
Wuornos
Semi-Pro


Join Date: Aug 2007
Location: England
Posts: 654 ELO Ratings of Major Winners for last 10 years

--------------------------------------------------------------------------------

ELO Rating for the winner of each major for the past 10 years at the point in time when they won the major.

French Open 1997 Gustavo Kuerten 2575
Wimbledon 1997 Pete Sampras 2749
US Open 1997 Patrick Rafter 2604
Australian Open 1998 Petr Korda 2598
French Open 1998 Carlos Moyà 2600
Wimbledon 1998 Pete Sampras 2725
US Open 1998 Patrick Rafter 2626
Australian Open 1999 Yevgeny Kafelnikov 2613
French Open 1999 Andre Agassi 2627
Wimbledon 1999 Pete Sampras 2712
US Open 1999 Andre Agassi 2667
Australian Open 2000 Andre Agassi 2683
French Open 2000 Gustavo Kuerten 2590
Wimbledon 2000 Pete Sampras 2704
US Open 2000 Marat Safin 2593
Australian Open 2001 Andre Agassi 2678
French Open 2001 Gustavo Kuerten 2601
Wimbledon 2001 Goran Ivanišević 2613
US Open 2001 Lleyton Hewitt 2599
Australian Open 2002 Thomas Johansson 2584
French Open 2002 Albert Costa 2571
Wimbledon 2002 Lleyton Hewitt 2632
US Open 2002 Pete Sampras 2669
Australian Open 2003 Andre Agassi 2671
French Open 2003 Juan Carlos Ferrero 2592
Wimbledon 2003 Roger Federer 2581
US Open 2003 Andy Roddick 2625
Australian Open 2004 Roger Federer 2620
French Open 2004 Gastón Gaudio 2567
Wimbledon 2004 Roger Federer 2637
US Open 2004 Roger Federer 2664
Australian Open 2005 Marat Safin 2633
French Open 2005 Rafael Nadal 2569
Wimbledon 2005 Roger Federer 2685
US Open 2005 Roger Federer 2701
Australian Open 2006 Roger Federer 2714
French Open 2006 Rafael Nadal 2600
Wimbledon 2006 Roger Federer 2730
US Open 2006 Roger Federer 2741
Australian Open 2007 Roger Federer 2752
French Open 2007 Rafael Nadal 2652
Wimbledon 2007 Roger Federer 2767
US Open 2007 Roger Federer 2776
Australian Open 2008 Novak Đoković 2670


Wuornos
View Public Profile
Find More Posts by Wuornos
Add Wuornos to Your Buddy List

02-02-2008, 12:30 AM #4
Wuornos
Semi-Pro


Join Date: Aug 2007
Location: England
Posts: 654 What the numbers mean

--------------------------------------------------------------------------------

The following table shows the expected outome of a match between two players depending on ELO Rating differential on a random surface.

The first figure is the difference between the two players ratings, the second figure is te probability the stronger player will win in straight sets, then the probability the stronger player will win in three sets, then the probability the weaker player will win in three sets and finally the probability that the weaker player will win in straight sets.

ELO Dif 2-0 2-1 1-2 0-2
10 26% 26% 24% 24%
20 28% 26% 24% 23%
30 29% 27% 23% 21%
40 30% 27% 22% 20%
50 32% 28% 22% 19%
60 33% 28% 21% 18%
70 35% 28% 20% 17%
80 36% 29% 19% 16%
90 38% 29% 18% 15%
100 39% 29% 18% 14%
110 41% 29% 17% 13%
120 42% 30% 16% 12%
130 44% 30% 15% 11%
140 46% 30% 14% 11%
150 47% 30% 13% 10%
160 49% 29% 13% 9%
170 51% 29% 12% 8%
180 53% 29% 11% 8%
190 54% 29% 10% 7%
200 56% 28% 9% 6%
210 58% 28% 9% 6%
220 60% 27% 8% 5%
230 62% 26% 7% 5%
240 64% 26% 6% 4%
250 66% 25% 6% 4%
260 68% 24% 5% 3%
270 70% 23% 4% 3%
280 72% 22% 4% 2%
290 74% 20% 3% 2%
300 77% 19% 3% 2%
310 79% 18% 2% 1%
320 81% 16% 2% 1%
330 83% 15% 1% 1%
340 86% 13% 1% 1%
350 88% 11% 1% 0%
360 90% 9% 0% 0%
370 93% 7% 0% 0%
380 95% 5% 0% 0%
390 98% 2% 0% 0%
400 100% 0% 0% 0%

We can therefore see that when a players rating is 140 points higher than their opponent, the odds of the weaker playing winning the match should be 3-1. While the probability of the weaker player winning the match when they go into the match with a 360 point deficit is negligible (i.e less than 1%).

Finally by comparing this table with the ELO Ratings for the open era we can see a 51 point difference between Roger Federer and Boris Becker. By consulting this table we can conclude that ELO estimates 60% chance of a peak Roger Federer beating a peak Boris Becker and a 32% chance that this will be in straight sets, assuming a stable playing population over time.

Regards

Tim
 

Wuornos

Professional
-Sorry Wuornos, i record your material in my disc.
You can't erase it:

:) Laugh. Feeling much better today and it's really the best piece of feedback I have ever received that someone thought my data was worthy of copying for their own personal records.

I shall be reposting the data today and chould be getting my next update of results next month so I will revise them then.

Take care

Regards

Tim
 

matchmaker

Hall of Fame
I guess it all boils down to what event you pick to calculate your ELO rankings.

Gustavo Kuerten won the masters defeating all the best players on the way. I guess if you retain that tournament he would be in the top ten instead of being at the bottom of the list.

Wuornos, would you have an ELO appreciation for that masters cup in 2000?
 

matchmaker

Hall of Fame
I guess it depends on what data you use to calculate the ELO points.

For instance Guga Kuerten is at the bottom of the list but if you take his master cup win in 2000 with back to back wins against Sampras and Agassi, surely the points must be far higher.

Wuornos, do you have any way to calculate that particular result?

Something else which isn't taken into account is the strength of players on a certain surface. Again, to take the example of Kuerten, he won his first RG with victories over three former champions: Muster, Kafelnikov and Bruguera. How does that translate into ELO?
 

Benhur

Hall of Fame
Deleted as found unuseful by board members. :cry:

So all those threads containing painstaking statistical calculations for ELO ratings are deleted, while hundreds of thousands of posts containing any junk that crosses a poster's mind are preserved for posterity.

Thanks for your great work, Wuornos. Since the airheads who deleted it are likely to keep doing it (stupidity tends to by an incurable ailment) I would encourage you to post updates to it on your own site or blog, and just post the links to it here periodically. There is a chance the censors won't be offended by the links.
 

leapfrog

New User
ELO ratings redone

Hi everyone!

I am new to this forum, please go easy on me if I don't follow the right etiquette.

I have redone the tennis ELO ratings with a database of matches going back to the beginning of the open era 1968. The match data were downloaded from the ATP website.
The database contains 14699 players and 384050 matches as of today.

There are several decisions to make when computing ELO ratings.

1. What distribution to use: I chose the logistical distribution with the standard deviation of 400. (see the wikipedia's article on ELO ratings)

2. What rating to give a new player: I chose 1500

3. What K factor to use: I chose K = 30, independent of the players's experience and surface type

4. What information about the turnament, and the score to use: As a first pass I just use the outcome of the match, without using the set score, the level of turnament or the surface

The ELO ratings have several features that may make them unsuitable for evaluating current player strength:

1. The current rating tends to lag behind the true strength, the size of the lag as to do with the value of K and whether it is adjustable

2. If a player is inactive for a while, his rating does not change. This may be fine for chess but for game in which physical condition is everything, it seems wrong.

That said, ELO does provide a good measure of maximum strength that should be invariant over time. Thus if Roger Federer were transported to 1985 he is likely to beat John McEnroe.

I have all sorts of data computed, please let me know what you would like to see. Just to start out here is the list of the top 100 players of all time, with the date the rating was achieved:

Rank Name Year achieved Rating

1 Roger,Federer,SUI 2007.7 2890.2
2 John,McEnroe,USA 1985.3 2835.6
3 Ivan,Lendl,USA 1990.1 2814.3
4 Rafael,Nadal,ESP 2009.4 2799.6
5 Bjorn,Borg,SWE 1980.5 2771.4
6 Jimmy,Connors,USA 1980.1 2665.1
7 Novak,Djokovic,SRB 2009.9 2619.4
8 Pete,Sampras,USA 1997.2 2609.6
9 Andy,Murray,GBR 2010.1 2584.2
10 Rod,Laver,AUS 1975.2 2567.8
11 Boris,Becker,GER 1991.1 2558.3
12 Andy,Roddick,USA 2009.6 2540.9
13 Stefan,Edberg,SWE 1991.8 2533.0
14 Andre,Agassi,USA 1995.7 2500.5
15 Lleyton,Hewitt,AUS 2005.2 2499.0
16 Guillermo,Vilas,ARG 1979.0 2498.1
17 Nikolay,Davydenko,RUS 2010.1 2485.8
18 Mats,Wilander,SWE 1988.8 2467.7
19 Michael,Chang,USA 1997.3 2416.9
20 Juan_Martin,Del_Potro,ARG 2009.7 2390.9
21 Ilie,Nastase,ROU 1976.9 2382.8
22 David,Nalbandian,ARG 2009.1 2369.4
23 Arthur,Ashe,USA 1976.2 2366.4
24 Fernando,Gonzalez,CHI 2009.7 2365.5
25 Thomas,Muster,AUT 1997.2 2351.0
26 Gustavo,Kuerten,BRA 2001.7 2341.3
27 Jim,Courier,USA 1993.5 2339.1
28 Robin,Soderling,SWE 2010.1 2336.3
29 Vitas,Gerulaitis,USA 1980.5 2334.7
30 Brian,Gottfried,USA 1978.3 2330.1
31 Goran,Ivanisevic,CRO 1997.2 2328.8
32 Jo-Wilfried,Tsonga,FRA 2010.1 2322.9
33 Ivan,Ljubicic,CRO 2007.2 2311.0
34 Carlos,Moya,ESP 2004.5 2309.9
35 Guillermo,Canas,ARG 2007.3 2309.9
36 David,Ferrer,ESP 2008.5 2302.3
37 Fernando,Verdasco,ESP 2010.1 2301.7
38 Juan_Carlos,Ferrero,ESP 2003.8 2301.6
39 Eddie,Dibbs,USA 1978.6 2300.2
40 James,Blake,USA 2008.3 2300.0
41 Michael,Stich,GER 1995.6 2294.7
42 Guillermo,Coria,ARG 2005.6 2290.3
43 Patrick,Rafter,AUS 2001.7 2287.0
44 John,Newcombe,AUS 1975.0 2285.1
45 Tommy,Haas,GER 2007.2 2277.5
46 Jose-Luis,Clerc,ARG 1982.6 2272.5
47 Todd,Martin,USA 1999.3 2272.1
48 Gene,Mayer,USA 1983.3 2270.5
49 Tommy,Robredo,ESP 2009.4 2268.7
50 Radek,Stepanek,CZE 2010.0 2261.4
51 Manuel,Orantes,ESP 1977.8 2261.3
52 Yannick,Noah,FRA 1987.2 2260.7
53 Harold,Solomon,USA 1980.8 2254.4
54 Stan,Smith,USA 1973.4 2254.4
55 Yevgeny,Kafelnikov,RUS 1999.1 2254.3
56 Richard,Krajicek,NED 1999.2 2251.1
57 Marin,Cilic,CRO 2010.1 2250.7
58 Ken,Rosewall,AUS 1975.6 2248.2
59 Tim,Henman,GBR 2002.3 2239.6
60 Sergi,Bruguera,ESP 1994.6 2238.7
61 Tomas,Berdych,CZE 2010.0 2237.0
62 Marcelo,Rios,CHI 1999.4 2236.6
63 Tom,Okker,NED 1974.2 2235.5
64 Marat,Safin,RUS 2005.2 2234.4
65 Brad,Gilbert,USA 1989.8 2230.6
66 Mario,Ancic,CRO 2009.2 2226.6
67 Gael,Monfils,FRA 2010.1 2224.1
68 Alex,Corretja,ESP 2001.6 2221.7
69 Roscoe,Tanner,USA 1979.7 2221.0
70 Andres,Gomez,ECU 1986.6 2215.9
71 Marcos,Baghdatis,CYP 2010.1 2215.4
72 Raul,Ramirez,MEX 1978.5 2209.4
73 Gaston,Gaudio,ARG 2005.6 2208.4
74 Mikhail,Youzhny,RUS 2010.1 2205.4
75 Petr,Korda,CZE 1998.1 2202.0
76 Jose,Higueras,ESP 1983.5 2200.4
77 Richard,Gasquet,FRA 2010.1 2195.2
78 Gilles,Simon,FRA 2009.9 2194.9
79 Thomas,Enqvist,SWE 2000.8 2185.4
80 Wayne,Ferreira,RSA 1996.7 2178.6
81 Miloslav,Mecir_Sr.,SVK 1988.7 2176.5
82 Stanislas,Wawrinka,SUI 2010.1 2175.8
83 Greg,Rusedski,GBR 1999.8 2172.7
84 Joakim,Nystrom,SWE 1986.4 2163.6
85 Nicolas,Kiefer,GER 2008.6 2151.5
86 Eliot,Teltscher,USA 1982.1 2151.4
87 Kent,Carlsson,SWE 1989.3 2150.2
88 Wojtek,Fibak,POL 1980.6 2148.7
89 Juan,Monaco,ARG 2010.2 2142.5
90 Sebastien,Grosjean,FRA 2003.5 2141.0
91 Magnus,Norman,SWE 2000.8 2135.4
92 Philipp,Kohlschreiber,GER 2009.5 2134.9
93 Guy,Forget,FRA 1992.9 2133.6
94 Tomas,Smid,CZE 1984.3 2132.8
95 Jiri,Novak,CZE 2004.8 2132.3
96 Tim,Mayotte,USA 1988.8 2127.6
97 Anders,Jarryd,SWE 1988.2 2126.6
98 Thomas,Johansson,SWE 2006.1 2124.1
99 Dmitry,Tursunov,RUS 2008.8 2123.7
100 Mark,Philippoussis,AUS 2001.2 2123.6
 

leapfrog

New User
ELO ratings for singles men, K makes a difference

I have redone the ELO ratings with a variable K with K = 20 for all turnaments except for grand slams for which K = 40. This is akin to the world soccer teams rating where K depends on the level of the tournament.

This difference changes the order of the all time greats a bit:

Rank Name Date attained ELO rating
1 Roger,Federer,SUI 2009.7 2764.4
2 Ivan,Lendl,USA 1990.1 2708.4
3 John,McEnroe,USA 1985.3 2704.1
4 Bjorn,Borg,SWE 1980.5 2668.6
5 Rafael,Nadal,ESP 2009.4 2639.1
6 Jimmy,Connors,USA 1984.8 2585.3
7 Pete,Sampras,USA 1997.2 2533.2
8 Rod,Laver,AUS 1975.2 2487.4
9 Stefan,Edberg,SWE 1992.0 2457.0
10 Boris,Becker,GER 1991.1 2457.0
11 Novak,Djokovic,SRB 2009.9 2443.2
12 Andre,Agassi,USA 2003.3 2419.7
13 Mats,Wilander,SWE 1988.8 2416.5
14 Andy,Roddick,USA 2009.6 2411.4
15 Guillermo,Vilas,ARG 1979.0 2408.8
16 Andy,Murray,GBR 2010.1 2392.8
17 Lleyton,Hewitt,AUS 2005.7 2384.2
18 Michael,Chang,USA 1997.3 2320.9
19 Nikolay,Davydenko,RUS 2010.1 2319.3
20 Ilie,Nastase,ROU 1976.9 2301.1
21 Jim,Courier,USA 1993.5 2276.1
22 Arthur,Ashe,USA 1976.2 2273.4
23 Thomas,Muster,AUT 1997.2 2267.0
24 Vitas,Gerulaitis,USA 1980.5 2243.1
25 Fernando,Gonzalez,CHI 2009.7 2237.1
26 Goran,Ivanisevic,CRO 1997.2 2228.9
27 John,Newcombe,AUS 1975.0 2228.4
28 Juan_Martin,Del_Potro,ARG 2010.1 2223.0
29 David,Nalbandian,ARG 2009.1 2215.0
30 Gustavo,Kuerten,BRA 2001.7 2212.4
31 Brian,Gottfried,USA 1978.3 2208.9
32 Patrick,Rafter,AUS 2001.7 2204.0
33 Juan_Carlos,Ferrero,ESP 2003.8 2195.4
34 Manuel,Orantes,ESP 1978.6 2191.2
35 Ken,Rosewall,AUS 1975.6 2190.4
36 Michael,Stich,GER 1996.4 2187.2
37 Eddie,Dibbs,USA 1978.6 2185.8
38 Todd,Martin,USA 1999.7 2184.1
39 Jo-Wilfried,Tsonga,FRA 2010.1 2182.5
40 Tommy,Haas,GER 2009.6 2181.5
 

leapfrog

New User
Another change I would like to make is to have three different rankings for three different surface types: clay, hard and grass. If the ratings are completely decoupled it's like having three different sports with the same players. I am not so interested in this scenario. Tennis on a different surface does, IMHO, qualify as a different sport. So I was thinking of an effective ranking which depends on the surface type.

If the rankings are R_c, R_h and R_g, I would use

0.5*R_c + 0.25*(R_h + R_g) as the rating for clay court tournaments, and so on....

What do you think? Could you imagine other schemes?

Thanks!
 

kOaMaster

Hall of Fame
I love your work! do you have this somewhere online? I'd love to play around with those numbers or see some statistics
 

Aonikenk

New User
Hello leapfrog,

Thanks for your message. I am also planning to study several ranking techniques for tennis : ELO, glicko, Whole-History rating, etc.
And I think indeed that the different surfaces should be included in the rating system.

Also I would like to take into account results available on http://www.itftennis.com/ . There is a lot more of results and the evolution of each player would be more precise (since you have junior results !) But is there a way to download easily all those results ??
 

aldeayeah

G.O.A.T.
Interesting. Nadal having a good H2H against Federer gives a tremendous boost to his ELO.

Another interesting piece of data: Hewitt's ELO peaked in July 2005 (after making the Wimbledon semis). Which proves how much he was stalled by Federer in 2004-2005.

EDIT: On the subject of surfaces, I'd prefer to have totally decoupled rankings, to see the true specialists shine.
 
Last edited:
Top