Current ELO Ratings

[Wuornos]

1. Roger Federer 2796 Peak Rating and still climbing
2. Rafael Nadal 2690 Peak Rating and still climbing
3. Andy Roddick 2667 Down from a peak rating of 2676 at the start of the year.
4. Novak Đoković 2654 Peak Rating and climbing fast
5. Nikolay Davydenko 2641 Peak Rating and still climbing
6. Lleyton Hewitt 2627 Down from a peak rating of 2678 in 2005
7. Fernando González 2613 Down from a peak rating of 2616 at the start of the year.
8. David Nalbandian 2609 Down from a peak rating of 2637 achieved at the French Open in 2006.
9. Juan Carlos Ferrero 2601 Down from a peak rating of 2652 at the start of 2004.
10. Marcos Baghdatis 2600 Previous peak at 2006 Wimbledon 2618.


1. Justine Henin 2716 Peak Rating and still climbing
2. Maria Sharapova 2664 Has dropped back from 2684 since the 2007 French Open.
3. Serena Williams 2662 No where near her 2003 peak of 2733
4. Venus Williams 2647 Also well off her 2003 peak of 2720
5. Amélie Mauresmo 2639 Finished last year on 2673.
6. Svetlana Kuznetsova 2617 Is currently inmproving. This is her best rating.
7. Lindsay Davenport 2603 Jut returned. Doubt she can again approach her best of 2701 in 2001
8. Ana Ivanović 2587 Best rating so far and climbing fast.
9. Marion Bartoli 2564 Great Wimbledon. Playing the best Tennis of her life.
10. Nicole Vaidišová 2563 Has dopped a few points since start of year when she was 2569.


The Mens and Womens ratings are not interchangable.

Just ran the system I use for Elo ratings and thought I would share them with you.

Take care all

Tim

 

[brownyazn]

what are elo ratings?

 

[pabloJD]

Jelena Jankovic wouldn't like ELO ratings

 

[Dedans Penthouse]

what are elo ratings?

I'm lost too; I was thinking "Electric Light Orchestra"

 

[jaggy]

I'm lost too; I was thinking "Electric Light Orchestra"

Now I cant get Mr Blue Sky out my head

 

[Rhino]

I'm lost too; I was thinking "Electric Light Orchestra"

Me too, perhaps this is how the band ELO rate the current players.

 

[David L]

what are elo ratings?

It's a rating system used in the world of chess. I believe Wuornos believes they give more accurate rankings than the current ATP system. Good work, it's interesting to see what the rankings would look like under a different system.

 

[Rhino]

It's a rating system used in the world of chess. I believe Wuornos believes they give more accurate rankings than the current ATP system. Good work, it's interesting to see what the rankings would look like under a different system.

Well putting Roddick above Novak says to me that the system is highly flawed.

 

[Paul Goldstein]

1. Roger Federer 2796 Peak Rating and still climbing
2. Rafael Nadal 2690 Peak Rating and still climbing
3. Andy Roddick 2667 Down from a peak rating of 2676 at the start of the year.
4. Novak Đoković 2654 Peak Rating and climbing fast
5. Nikolay Davydenko 2641 Peak Rating and still climbing
6. Lleyton Hewitt 2627 Down from a peak rating of 2678 in 2005
7. Fernando González 2613 Down from a peak rating of 2616 at the start of the year.
8. David Nalbandian 2609 Down from a peak rating of 2637 achieved at the French Open in 2006.
9. Juan Carlos Ferrero 2601 Down from a peak rating of 2652 at the start of 2004.
10. Marcos Baghdatis 2600 Previous peak at 2006 Wimbledon 2618.


1. Justine Henin 2716 Peak Rating and still climbing
2. Maria Sharapova 2664 Has dropped back from 2684 since the 2007 French Open.
3. Serena Williams 2662 No where near her 2003 peak of 2733
4. Venus Williams 2647 Also well off her 2003 peak of 2720
5. Amélie Mauresmo 2639 Finished last year on 2673.
6. Svetlana Kuznetsova 2617 Is currently inmproving. This is her best rating.
7. Lindsay Davenport 2603 Jut returned. Doubt she can again approach her best of 2701 in 2001
8. Ana Ivanović 2587 Best rating so far and climbing fast.
9. Marion Bartoli 2564 Great Wimbledon. Playing the best Tennis of her life.
10. Nicole Vaidišová 2563 Has dopped a few points since start of year when she was 2569.


The Mens and Womens ratings are not interchangable.

Just ran the system I use for Elo ratings and thought I would share them with you.

Take care all

Tim

are you the guy who owns setratings.com? if so please get it back up. if not, how did you run the ratings?

 

[David L]

Well putting Roddick above Novak says to me that the system is highly flawed.

Well yes, it's only interesting from an academic point of view.

 

[djsiva]

Nice!

Who has the highest ELO rating ever? For tennis, of course.

Maybe this will end all arguements of who's best.

If I remember correctly Kasparov rating was better than Bobby Fischer, right?

 

[vive le beau jeu !]

I'm lost too; I was thinking "Electric Light Orchestra"

eheh... me too !

are you the guy who owns setratings.com? if so please get it back up. if not, how did you run the ratings?

this site was relatively interesting... would be cool to see it back.

 

[Wuornos]

are you the guy who owns setratings.com? if so please get it back up. if not, how did you run the ratings?

Thanks all

Yes ELO ratings are the rating system that was adopted by the American Dr Arpad Elo as system that measures quality of results in chess and are calcualted with overlapping normal distributions of performance for the two players involved in the individual match concerend.

Yes SET RATINGS was a good site and did use the ELO system in the calculation of performance but unfortunately I am not the person who ran it. These ratings may be slightly different to those that were shown on SET Ratings asthe ELO system is only really measuring the difference between player performances and whether the base starts at 2200 or 1000 is really irrelevant. I suspect my base will be different to set ratings.

Yes Jelena Janković would be disappointed with ELO ratings. The official ratings are a volume based system meaning if you play lots and lots of tennis you achieve a higher rating than if you say play a small number of tournaments each year. ELO ratings just measure the quality of performance and only the speed of rating change is effected by the volume of matches played. Jelena who would be number 11 in the Elo ratings is a high volume player and therefore benefits from this in the official ratings.

Who has been the highest rated player? Dunno, my data only covers the open era, but in the open era Roger Federer's current rating is the highest that has been recorded in my database.

ELO is just a different way of looking at things based purely on quality of results rather than volume. I wouldn't say its better or worse than the official rankings. Just different. The official rankings being volume based encourage the player to compete more which is in the fans interest.

Just thought people would be interested in seeing the output after I ran the system up to the end of the US Open.

Take care all

Tim

 

[Wuornos]

Nice!

Who has the highest ELO rating ever? For tennis, of course.

Maybe this will end all arguements of who's best.

If I remember correctly Kasparov rating was better than Bobby Fischer, right?

Yes that's spot on. Were you like me chess fan ?

 

[Wuornos]

Well yes, it's only interesting from an academic point of view.

Agreed and that's why I posted the results.

Take care

Tim

 

[Wuornos]

Well putting Roddick above Novak says to me that the system is highly flawed.

Well yes, and you're entitled to your opinion. The Elo ratings are showing Novak climbing very fast though while Andy just maintaining his standard. Novak will probably overtake Andy in the very near future.

Take care

Tim

 

[ecy0004]

1. Roger Federer 2796 Peak Rating and still climbing
2. Rafael Nadal 2690 Peak Rating and still climbing
3. Andy Roddick 2667 Down from a peak rating of 2676 at the start of the year.
4. Novak Đoković 2654 Peak Rating and climbing fast
5. Nikolay Davydenko 2641 Peak Rating and still climbing
6. Lleyton Hewitt 2627 Down from a peak rating of 2678 in 2005
7. Fernando González 2613 Down from a peak rating of 2616 at the start of the year.
8. David Nalbandian 2609 Down from a peak rating of 2637 achieved at the French Open in 2006.
9. Juan Carlos Ferrero 2601 Down from a peak rating of 2652 at the start of 2004.
10. Marcos Baghdatis 2600 Previous peak at 2006 Wimbledon 2618.


1. Justine Henin 2716 Peak Rating and still climbing
2. Maria Sharapova 2664 Has dropped back from 2684 since the 2007 French Open.
3. Serena Williams 2662 No where near her 2003 peak of 2733
4. Venus Williams 2647 Also well off her 2003 peak of 2720
5. Amélie Mauresmo 2639 Finished last year on 2673.
6. Svetlana Kuznetsova 2617 Is currently inmproving. This is her best rating.
7. Lindsay Davenport 2603 Jut returned. Doubt she can again approach her best of 2701 in 2001
8. Ana Ivanović 2587 Best rating so far and climbing fast.
9. Marion Bartoli 2564 Great Wimbledon. Playing the best Tennis of her life.
10. Nicole Vaidišová 2563 Has dopped a few points since start of year when she was 2569.


The Mens and Womens ratings are not interchangable.

Just ran the system I use for Elo ratings and thought I would share them with you.

Take care all

Tim

Is it possible to get graphs or history of the players' rating from somewhere?
bg

 

[Rhino]

Why does it take so long for players to rise and fall in this system, for example JCF still in the top 10? And why would Hewitt peak in 2005 and not in 2002?

 

[Thud and blunder]

Wuornos, very interesting; thanks for posting.

I don't want to hijack the thread, but does anyone have a feel for how NTRP ratings would map to ELO ratings, eg 3 = 1600 = average club player (?), 5 = 2100, 6 = 2300, 7 = 2500 = professional level in both chess and tennis...well thats my rudimentary guesstimate for how things would map out.

These ELO ratings are FIDE ratings - international rating list. just to confuse matters, USCF ratings are higher than FIDE ratings (by about 100 points(?))

 

[Rhino]

Wuornos, very interesting; thanks for posting.

I don't want to hijack the thread, but does anyone have a feel for how NTRP ratings would map to ELO ratings, eg 3 = 1600 = average club player (?), 5 = 2100, 6 = 2300, 7 = 2500 = professional level in both chess and tennis...well thats my rudimentary guesstimate for how things would map out.

These ELO ratings are FIDE ratings - international rating list. just to confuse matters, USCF ratings are higher than FIDE ratings (by about 100 points(?))

wtf? I've never heard of NTRP, ELO, FIDE, or USCF ratings! Maybe I should have a rating system too: The Rhino Ratings...
#1. Federer 10000000pts
#2. Safin 2000pts
#3. Agassi 1800pts
#4. Ivanisevic 1500pts
#5. Nadal 1000pts

......#9999. Roddick 0 points

It's very scientific you know.

 

[Thud and blunder]

Why does it take so long for players to rise and fall in this system, for example JCF still in the top 10? And why would Hewitt peak in 2005 and not in 2002?

This system is more favourable for less active players. Your rating doesn't drop through inactivity, only results. You have an expected score of 50% against players of the same rating, 45% against a player 50 points higher (or something like that...), change in rating = K * (expected - actual), where K is a multiplication factor, which is higher for juniors and new players to allow them to find their true level quicker, and is lower for established players.

The big advantage of ELO vs rankings is you get ranked according to who you play, not which round of which tourney you get to. If you win three rounds of Wimby against nonentities, you gain fewer points than beating seeds in rounds one and two and losing to a seed in round three. To my mind, this is a huige advantage.

 

[Thud and blunder]

You've never heard of NTRP??

 

[ecy0004]

This system is more favourable for less active players. Your rating doesn't drop through inactivity, only results. You have an expected score of 50% against players of the same rating, 45% against a player 50 points higher (or something like that...), change in rating = K * (expected - actual), where K is a multiplication factor, which is higher for juniors and new players to allow them to find their true level quicker, and is lower for established players.

The big advantage of ELO vs rankings is you get ranked according to who you play, not which round of which tourney you get to. If you win three rounds of Wimby against nonentities, you gain fewer points than beating seeds in rounds one and two and losing to a seed in round three. To my mind, this is a huige advantage.

True, but very difficult for a 'normal' tennis fan to understand. who would bother about ratios and multiplication factors when a simple 100, 70, 45, 25 etc points system is much easier to understand.
Guess ATP is more worried about tennis popularity than effectiveness
bg

 

[Wuornos]

Wuornos, very interesting; thanks for posting.

I don't want to hijack the thread, but does anyone have a feel for how NTRP ratings would map to ELO ratings, eg 3 = 1600 = average club player (?), 5 = 2100, 6 = 2300, 7 = 2500 = professional level in both chess and tennis...well thats my rudimentary guesstimate for how things would map out.

These ELO ratings are FIDE ratings - international rating list. just to confuse matters, USCF ratings are higher than FIDE ratings (by about 100 points(?))

Yes that's right the FIDE and USCF ratings are not an exact match. This is the issue I mentioned earlier in the thread when I was concerned that people didn't see these as the SET RATINGS as they may have a different base.

The main thing with ELO ratings is not the overall figure, it is the difference between the various players whether the starting points is 2200 or 1000 does not impair the accuracy of the rating generated.

See the following wikipaedia entry for a full explanation of how Elo ratings are calculated. This isn't something I have made up. I wish I had, it's just a system with a reasonably sound basis in probability theory for evaluating the performance of players in two player sports or two team sports.

http://en.wikipedia.org/wiki/Elo_rating_system

International Soccer has this system operting as did tennis until the demise of SET Ratings. It's original use, as already mentioned, was chess

http://www.eloratings.net/

Thanks for the interest all

Tim

 

[Wuornos]

Is it possible to get graphs or history of the players' rating from somewhere?
bg

Well I can produce them from my database but I'm not sure how to post them within a thread. Alos backdated lists to a particular time within the open era.

 

[djsiva]

Yes that's spot on. Were you like me chess fan ?

Yeah I kind of lost interest when computers started whooping Kasparov.

In Kasparov's defense, it was unfair.

Kasparov wasn't given access to the games Big Blue played, while the computer knew very well all the games Kasparov had ever played.

Sort of like letting Federer play a match, not telling him who he's playing or what surface he was playing on. While the other player studied Federer over and over for two months. Then you put him against a Nadal at his prime on clay, or a young Pete Sampras on grass.

No wonder Kasparov lost.

So have you figured out elo on Sampras, Lendl, Borg, and Laver?

 

[Thud and blunder]

All very interesting...I never realised the soccer rankings were based on the same system.

I wonder how the ratings would look on the different surfaces. That would be an interesting exercise.

 

[Wuornos]

Yeah I kind of lost interest when computers started whooping Kasparov.

In Kasparov's defense, it was unfair.

Kasparov wasn't given access to the games Big Blue played, while the computer knew very well all the games Kasparov had ever played.

Sort of like letting Federer play a match, not telling him who he's playing or what surface he was playing on. While the other player studied Federer over and over for two months. Then you put him against a Nadal at his prime on clay, or a young Pete Sampras on grass.

No wonder Kasparov lost.

So have you figured out elo on Sampras, Lendl, Borg, and Laver?

I stopped playing chess a little earlier than that when I got married. Never seemed to find the time to be honest.

I agree about Kasparov and computer chess though. It wasn't a level playing field. I also kind of thought of it as being like 100m sprinter versus a motorbike. i.e. a bit irrelevant really.

Funny you should pick those players to ask about in relation to ELO ratings as along with Federer, they are the absolute top 5. Good call.

I'd better say that ELO is a measure of dominance adjusted for the relative dominance of opposition faced rather than total career achievement. More for other readers benefits than yours as I know you already know this.

I know people wont' like this but the top 5 peak ratings on my database are as follows:

1. Roger Federer 2796
2. Ivan Lendl 2789
3. Rod Laver 2783
4. Pete Sampras 2769
5. Björn Borg 2768

Sampras appears lower than most people would probably expect as at his peak of 2769 his main opposition was Andre Agassi who was playing at a level of 2672. Agassi later went on to achive a peak of 2707 although he wasn't playing at that standard at this point in time. Hence Sampras dominance was deflated due to a lack of outstanding competition.

If you compare this with the ratings mid year in 1985 when McEnroe was 2756, I always liked that guy, Lendl was 2721, Connors was 2720 and Wilander was 2693. All four playing at a level higher than the best that Sampras had to make his peak rating off.

I think Federer faces the same problem as Sampras. He's top in the ratings but not by as greater margin as many people might expect.

In Lendl we see the opposite effect as he faced players who had high ratings and therefore his Elo is higher than people who only look at dominance and achievement without quality of opposition might expect.

The strangest career profile belongs to Borg who was dominant but achieved his best rating only with the emergence of McEnroe who challenged his dominance. McEnroe was the second highest rated American born player with a peak rating of 2756.

Laver's peak rating may have been higher than this. My database only covers the open era so he may have a higher peak pre open era. I just don't know. In any event 3rd place for a period covering 40years is pretty hot.

Take care and keep in touch

Tim

 

[Wuornos]

All very interesting...I never realised the soccer rankings were based on the same system.

I wonder how the ratings would look on the different surfaces. That would be an interesting exercise.

Yes that would be interesting, but unfortunately my database does not have the necessary fileds to be able to calculate this, and even if it did it would require a major redesign of the system.

Sorry

Tim

 

[Wuornos]

This system is more favourable for less active players. Your rating doesn't drop through inactivity, only results. You have an expected score of 50% against players of the same rating, 45% against a player 50 points higher (or something like that...), change in rating = K * (expected - actual), where K is a multiplication factor, which is higher for juniors and new players to allow them to find their true level quicker, and is lower for established players.

The big advantage of ELO vs rankings is you get ranked according to who you play, not which round of which tourney you get to. If you win three rounds of Wimby against nonentities, you gain fewer points than beating seeds in rounds one and two and losing to a seed in round three. To my mind, this is a huige advantage.

Absolutely spot on. Wish I could explain things as well as this.

Thanks

Tim

 

[djsiva]

I stopped playing chess a little earlier than that when I got married. Never seemed to find the time to be honest.

I agree about Kasparov and computer chess though. It wasn't a level playing field. I also kind of thought of it as being like 100m sprinter versus a motorbike. i.e. a bit irrelevant really.

Funny you should pick those players to ask about in relation to ELO ratings as along with Federer, they are the absolute top 5. Good call.

I'd better say that ELO is a measure of dominance adjusted for the relative dominance of opposition faced rather than total career achievement. More for other readers benefits than yours as I know you already know this.

I know people wont' like this but the top 5 peak ratings on my database are as follows:

1. Roger Federer 2796
2. Ivan Lendl 2789
3. Rod Laver 2783
4. Pete Sampras 2769
5. Björn Borg 2768

Sampras appears lower than most people would probably expect as at his peak of 2769 his main opposition was Andre Agassi who was playing at a level of 2672. Agassi later went on to achive a peak of 2707 although he wasn't playing at that standard at this point in time. Hence Sampras dominance was deflated due to a lack of outstanding competition.

If you compare this with the ratings mid year in 1985 when McEnroe was 2756, I always liked that guy, Lendl was 2721, Connors was 2720 and Wilander was 2693. All four playing at a level higher than the best that Sampras had to make his peak rating off.

I think Federer faces the same problem as Sampras. He's top in the ratings but not by as greater margin as many people might expect.

In Lendl we see the opposite effect as he faced players who had high ratings and therefore his Elo is higher than people who only look at dominance and achievement without quality of opposition might expect.

The strangest career profile belongs to Borg who was dominant but achieved his best rating only with the emergence of McEnroe who challenged his dominance. McEnroe was the second highest rated American born player with a peak rating of 2756.

Laver's peak rating may have been higher than this. My database only covers the open era so he may have a higher peak pre open era. I just don't know. In any event 3rd place for a period covering 40years is pretty hot.

Take care and keep in touch

Tim

Nice work!!

I'm no Big Blue, but I always had a feeling that these were the players to talk about. I'm sort of like old school Fischer or Capablanca. Not too much into the computers, but have a general big picture feeling. You have both. Sort of what Kasparov was saying. He wanted chessplayers to be able to use calculators and chess programs to play against computers.

Lendl is so underrated by people, even surprising on this board. Maybe not surprisingly because most of the guys here know only Roddick, Nadal, and Federer (and Isner?!?)

I know people won't be interested, but if you have time and its easy could you list the top 100 players with your elo ratings. I'm just curious. If it's too much work don't worry about it.

Thanks

 

[CyBorg]

Yeah I kind of lost interest when computers started whooping Kasparov.

One computer and it barely beat him... with considerable controversy.

 

[David L]

One computer and it barely beat him... with considerable controversy.

Yes, and he beat Deep Blue in the first match, only losing in the re-match in the final game to a primed Deep Blue, as you say, surrounding a lot of controversy.

 

[djsiva]

One computer and it barely beat him... with considerable controversy.

Yeah I know. But that was like over ten years ago. Think about what computers could do now.

It like playing times tables with your calculator.

That's why I lost interest.

So did Kasparov. He's into politics and corruption. Stuff you don't need a computer for!! Humans are the full of corruption.

 

[onehandbh]

One game where the top computers are still getting their silicon behinds
kicked is GO. The best computers can even beat the best kids. I've
only played it a few times, but the number of moves and possiblities
are huge. Much larger and much more complicated than chess - yet
the game's rules are quite simple and a 4 year old can even learn to
play.

There's a *Taiwanese foundation that is offering $1 million (US dollars)
to the first person that can make a computer that can beat a selected
twelve-year-old GO player.

For those of you a little more ambitious, Professor Elwyn Berlekamp at
UC Berkeley has upped the ante by offering US$5,000,000 to the first
computer program that can beat a 5-dan player, which is equivalent to
a strong expert player.

In case any of you geniuses out there think you might have an
algorithm/computer that's up to the challenge, here's the info:

The Ing Chang-ki Goe Educational Foundation 4F Kuang Fu Building,
No. 35 Kuang Fu South Road Taipei, Taiwan
Phone: 886-2-761-4117; FAX: 886-2-768-6940

 

[Wuornos]

Nice work!!

I'm no Big Blue, but I always had a feeling that these were the players to talk about. I'm sort of like old school Fischer or Capablanca. Not too much into the computers, but have a general big picture feeling. You have both. Sort of what Kasparov was saying. He wanted chessplayers to be able to use calculators and chess programs to play against computers.

Lendl is so underrated by people, even surprising on this board. Maybe not surprisingly because most of the guys here know only Roddick, Nadal, and Federer (and Isner?!?)

I know people won't be interested, but if you have time and its easy could you list the top 100 players with your elo ratings. I'm just curious. If it's too much work don't worry about it.

Thanks

Thanks for that positive feedback. Yes I could post the top 100. Although it will take time, try to do it by next weekend but will post as a new thread so you can find it.

I have just posted top 20 as an answer to thread

http://tt.tennis-warehouse.com/showthread.php?p=1807071#post1807071

Thanks again for the feedback.

Tim

 

[Wuornos]

One game where the top computers are still getting their silicon behinds
kicked is GO. The best computers can even beat the best kids. I've
only played it a few times, but the number of moves and possiblities
are huge. Much larger and much more complicated than chess - yet
the game's rules are quite simple and a 4 year old can even learn to
play.

There's a *Taiwanese foundation that is offering $1 million (US dollars)
to the first person that can make a computer that can beat a selected
twelve-year-old GO player.

For those of you a little more ambitious, Professor Elwyn Berlekamp at
UC Berkeley has upped the ante by offering US$5,000,000 to the first
computer program that can beat a 5-dan player, which is equivalent to
a strong expert player.

In case any of you geniuses out there think you might have an
algorithm/computer that's up to the challenge, here's the info:

The Ing Chang-ki Goe Educational Foundation 4F Kuang Fu Building,
No. 35 Kuang Fu South Road Taipei, Taiwan
Phone: 886-2-761-4117; FAX: 886-2-768-6940

Interesting.

I suspect the amount will be claimed one day but not by the same methodolgy that is currently used by the top chess computers. I suspect artificial intelligence might crack this egg, with a computer that has been specifically assigned the task of playing Go continuously.

Bit beyond me though! Good luck to whoever.

Tim

 

[Wuornos]

Nice work!!

I'm no Big Blue, but I always had a feeling that these were the players to talk about. I'm sort of like old school Fischer or Capablanca. Not too much into the computers, but have a general big picture feeling. You have both. Sort of what Kasparov was saying. He wanted chessplayers to be able to use calculators and chess programs to play against computers.

Lendl is so underrated by people, even surprising on this board. Maybe not surprisingly because most of the guys here know only Roddick, Nadal, and Federer (and Isner?!?)

I know people won't be interested, but if you have time and its easy could you list the top 100 players with your elo ratings. I'm just curious. If it's too much work don't worry about it.

Thanks

Top 100 peak ratings for open era for both men and women now posted as separate threads in 'Former Pro Player Talk'.

Take care and enjoy.

Tim