Djoko's Delight: ELO Ranks the US Open Fields from 1978

[Meles]

This thread is the successor to the Wimbledon ELO thread:
http://tt.tennis-warehouse.com/inde...-ranks-the-wimbledon-fields-from-1978.569259/

Please see that thread for all references for this OP and much discussion on ELO, etc.

ELO reveals all and tells all; so it has been spoken. These results on first blush shocked and amazed me after seeing the Wimbledon results which seemed to match very well with TTW perceptions. Again the early years must be thrown out. Its little surprise once again that the late 80s and early 90s are strong especially when we consider the early success of Agassi and Courier on tour on top of Sampras. The mid 90s bulge is from Agassi, Chang, and others. One of the issues is that these numbers for the US Open are from the hard court stats which are slam weighted and include Australia. With Agassi's dominate runs in Australia his hard court rating actually goes ahead of Sampras likely due to the slam weighting. In 1995 Agassi crushed everyone in his first Australian Open and then also made the US Open finals where he reached his peak ELO rating before playing Sampras. The later 90s had some weak fields. Agassi's resurgence appears to have bolstered the hard court field nicely in the early 2000s. Federer gets no credit for this as 2004 was his first run into the QFs and the title. The high ratings for the big 4 era should be little suprise as Federer and Djokovic are rated the two best hard court players and Murray, Nadal, and Delpo have ratings on par with Lendl in the slam weighted rankings used for this (it appears those 3 have been quite clutch on hard courts in slams and that has bumped their rating somewhat due to these hard court ratings being slam weighted.) Overall the biggest issue appears to be the mixing of the extensive hard court numbers that cover both the US Open and Australian Open on hard courts. The stats say the big 4 are very solid on hard courts and these field ratings show that. Even Murray is a great hard court player.

Of course we have some weakening of the field in the last few years, but it is still a very strong field. Nadal and Djokovic actually got their best nonslam weighted hard court ratings in 2014. These slam weighted results used for hard courts start in 2012 and so are low for 2013 and beyond. The veteran poly players have been dominating the fast, high bouncing US Open hard courts so no rating adjustments were necessary for the field. (Nadal and Djokovic might have been boosted slightly.) This is the complete opposite of my impression of the tour before seeing these ELO field ratings. I thought the Euro dominated tour was weak on US Open hard courts. I suspect Poly strings are having a huge impact at the US Open.

@Gary Duane this one boggles my mind when I first looked at the graphs. My quick thoughts are that the hard court results are spread across two tournaments both of which are slam weighted. However from 1995 back not all the major players played both tournaments and these ELO numbers are weighted for the five set format. Might that have penalized those who did not play Australia and earlier eras as a whole? I have trepidations about this, but ulitmately I don't think it matters that much. This one took me about 30 minutes to generate the tab and graphs. It actually took a bit longer as I had to manually enter the QF and SF players for 1982 and 1984 due to the data being a bit of a mess in the csv files (not my fault.)

 

[Gary Duane]

In the chart I see an Elo for QF and an Elo for SF.

How do you have an Elo for a match?

I simply can't follow what you are doing.

My quick thoughts are that the hard court results are spread across two tournaments both of which are slam weighted. However from 1995 back not all the major players played both tournaments and these ELO numbers are weighted for the five set format. Might that have penalized those who did not play Australia and earlier eras as a whole?

So are these rankings from ALL HC matches? Are these rankings then used to rate the players at the USO?

If I take these numbers seriously, the 2011 match between Nadal and Djokovic marks the absolute high-water mark of the USO. Well, that and the 2008 match between Murray and Fed.

To be honest, I'm not yet following the logic yet. Pretend I am very VERY stupid. And you may not be wrong. Exactly what is it that these graphs are showing, in your opinion?

 

[Meles]

How do you have an Elo for a match?

I simply can't follow what you are doing.

So are these rankings from ALL HC matches? Are these rankings then used to rate the players at the USO?

If I take these numbers seriously, the 2011 match between Nadal and Djokovic marks the absolute high-water mark of the USO. Well, that and the 2008 match between Murray and Fed.

To be honest, I'm not yet following the logic yet. Pretend I am very VERY stupid. And you may not be wrong. Exactly what is it that these graphs are showing, in your opinion?

This is the same format as Wimbledon. The ELOs are the average of the eight quarterfinalists (the average of their peak ELOs.) These are hard court ELOs.

This has all the peak ELOs for various surfaces throught end of 2012. The Rank of the winner and finalist are the number on this list for hard courts:
http://simtheworld.blogspot.co.il/2012/10/tennis-elo-ratings-hard_19.html

Here's the top 30 hard court slam weighted for everyone to see:
Pos Name Ratings Date
1 Roger Federer 2673.55 March 5, 2007
2 Novak Djokovic 2663.61 November 7, 2011
3 Andre Agassi 2532.93 September 4, 1995
4 Pete Sampras 2525.69 March 3, 1997
5 Andy Murray 2504.61 August 24, 2009
6 Juan Martin Del Potro 2501.89 September 7, 2009
7 Rafael Nadal 2476.30 April 1, 2009
8 Ivan Lendl 2469.02 January 18, 1988
9 Stefan Edberg 2447.17 January 25, 1993
10 Jim Courier 2424.14 March 15, 1993
11 John Mcenroe 2414.51 September 2, 1985
12 Mats Wilander 2412.67 November 7, 1988
13 Nikolay Davydenko 2405.26 January 25, 2010
14 Andy Roddick 2377.91 October 20, 2003
15 Lleyton Hewitt 2367.33 January 24, 2005
16 Michael Chang 2361.25 January 20, 1997
17 Marat Safin 2360.57 January 24, 2005
18 Marin Cilic 2339.91 January 25, 2010
19 Jo Wilfried Tsonga 2333.66 January 23, 2012
20 Boris Becker 2326.56 March 18, 1991
21 Fernando Gonzalez 2326.10 January 22, 2007
22 David Ferrer 2324.45 January 23, 2012
23 Petr Korda 2304.76 February 23, 1998
24 Robin Soderling 2298.41 January 24, 2011
25 Tomas Berdych 2297.35 February 20, 2012
26 Tommy Haas 2295.76 January 21, 2002
27 David Nalbandian 2290.54 August 16, 2010
28 Stanislas Wawrinka 2287.76 March 17, 2011
29 Jimmy Connors 2285.45 September 19, 1983
30 Gael Monfils 2283.82 November 15, 2010

I question the earlier ratings once again and Connors looks a tad low for his number in 1983. McEnroe is low as these were built from 1978 onwards.

Here is a more recent list built from 1969 not slam weighted, but only top 10:
Player Year Hard Ct Elo
Roger Federer 2007 2453
Novak Djokovic 2014 2418
Ivan Lendl 1989 2370
Pete Sampras 1997 2356
Rafael Nadal 2014 2342
John McEnroe 1986 2332
Andy Murray 2009 2330
Andre Agassi 1995 2326
Stefan Edberg 1987 2285
Lleyton Hewitt 2002 2262

This one has McEnroe better rated.

The graphs show strength of field as discussed. My take on this versus grass is that Agassi was very up and down in the 90s so that is why the field is generally weak with some periods of strength. Good Agassi makes a lot of those fields respectable, but they are generally weak by this measure. I can tell you this was not perception coming into this, but it makes sense.

 

[Noleberic123]

No but dude its a weak era

 

[buscemi]

Davydenko 7 spots higher than Becker?

Becker was 37-10 at the U.S. Open. Davydenko was 26-13. Becker won the U.S. Open in 1989, beating world #1 Ivan Lendl, who was playing in his 8th straight U.S. Open final and who had won the Australian Open and Canada that year.

Davydenko's best U.S. Open result was reaching the SF twice. Beyond his win in 1989, Becker made 3 other SFs. He also made four other 4th round appearances vs. two for Davydenko.

I would again say that this ranking is unfair to those with later-career success. After his "peak" year in 2010, Davydenko went 4-3, with those results not hurting his ranking at all. Meanwhile, after his "peak" year in 1991, Becker went 11-4, making the SF in 1995 (losing a close 4 setter to Agassi), with those results not helping his ranking at all.

 

[buscemi]

Also, how in the world is Korda ranked so high? He must be getting A LOT of mileage out of that Round of 16 win over Sampras in 1997. Overall, Petr's record at the U.S. Open was 11-9, with 1st round losses 5/9 times he played and 2nd round, 3rd round, and two QF losses the other year. Picking a name out of the air, Aaron Krickstein was 26-10 at the U.S. Open, with a SF appearance, 2 QF appearances, and 3 fourth round appearances as well as several solid wins (e.g., over Agassi in 1991). I would easily rank him higher.

[Edit: Now I see that these ratings include both AO and U.S. Open results. This helps explain the Korda ranking, but it makes the Becker ranking even less explicable given that Becker won two titles at the AO on Rebound Ace.].

 

[Meles]

No but dude its a weak era

Well according to the QF average peak ELO 2008 to 2012 was a strong as death era on hard courts. Think about it. Davydenko, Djokovic, Del Potro, Nadal, Federer, and Murray. Murray was beating peak Fed lol during this period. Murray and Del Potro had great years in 2009. Its Murray's peak statistical year by a long shot on hard courts. LOL, Roddick was still floating around and dangerous. Those were some loaded fields.

@Gary Duane has been beating me up in conversation about McEnroe having a peak hard court ELO in 1986 weighted for three sets....
"ELO shows McEnroe beating Edberg twice and Connors with many wins in 1986. Djokovic just lost early at Wimbledon. I'd say McEnroe's win rate was quite good in 1985 and 1986. Curren was a God at Wimbledon in 1985 and Lendl was better in 1985 at US Open. No shame losing to Wilander at the French. 1984 did look like a weaker field with Connors the major hurdle. The 7 months off in 1986 had to hurt and his game was becoming outmoded by graphite rackets. He won nothing in 1987 and I'd say that was when age started really catching up to him. It will be interesting to see Fed comeback after a 7 month layoff. With three wins in 1986 it doesn't suprise me that ELO says that is McEnroe's peak. Mary Carillo calls him back in form against Edberg:

Series covering the whole Connor's Final:

Some of the best tennis I've seen McEnroe play. Don't think ELO is much off the mark rating his late 1986 comeback."

Looking at McEnroe's 1984 and the weaker field, you really wonder about matchups and how things play out. With Federer's abscence this year we truly do have a weakening field. 2013-2015 have still been very strong on hard courts. ELO just shows the greatness of the big 4 and you wonder what some of these players might have down with the fields in the 1990s that were up and down.

 

[Meles]

Davydenko 7 spots higher than Becker?

Becker was 37-10 at the U.S. Open. Davydenko was 26-13. Becker won the U.S. Open in 1989, beating world #1 Ivan Lendl, who was playing in his 8th straight U.S. Open final and who had won the Australian Open and Canada that year.

Davydenko's best U.S. Open result was reaching the SF twice. Beyond his win in 1989, Becker made 3 other SFs. He also made four other 4th round appearances vs. two for Davydenko.

I would again say that this ranking is unfair to those with later-career success. After his "peak" year in 2010, Davydenko went 4-3, with those results not hurting his ranking at all. Meanwhile, after his "peak" year in 1991, Becker went 11-4, making the SF in 1995 (losing a close 4 setter to Agassi), with those results not helping his ranking at all.

Sorry. I wasn't watching this thread.

We are talking peak ELO ratings. I'm not a Becker expert, but his record at the US Open had a lot of earlier losses. Davydenko was very strong on hard courts. His hard court ELO peaked in 2010 and its little surprise:
Doha
Doha, Qatar
2010.01.04 - 2010.01.09
Rafael Nadal
W 06 76(8) 64
Roger Federer
W 64 64


Barclays ATP World Tour Finals
London, England
2009.11.23 - 2009.11.29
Finals 5
Juan Martin del Potro
W 63 64
Semi-Finals 1
Roger Federer
W 62 46 75
Round Robin 9
Robin Soderling
W 764 46 63
Round Robin 2
Rafael Nadal
W 61 764

ATP World Tour Masters 1000 Shanghai
Shanghai, China
2009.10.12 - 2009.10.18
Finals 2
Rafael Nadal
W 763 63
Semi-Finals 4
Novak Djokovic
W 46 64 761

That is an incredible collection of scalps and hence peak Davydenko is over peak Becker. Becker knocked off Lendl and a bunch of nobodies for his 1989 US Open. Davydenko was also very strong at 2006 and 2007 US Open. ELO does not lie.

 

[CYGS]

Does that prove Fed's era is the weakest of all time?

 

[Meles]

Also, how in the world is Korda ranked so high? He must be getting A LOT of mileage out of that Round of 16 win over Sampras in 1997. Overall, Petr's record at the U.S. Open was 11-9, with 1st round losses 5/9 times he played and 2nd round, 3rd round, and two QF losses the other year. Picking a name out of the air, Aaron Krickstein was 26-10 at the U.S. Open, with a SF appearance, 2 QF appearances, and 3 fourth round appearances as well as several solid wins (e.g., over Agassi in 1991). I would easily rank him higher.

[Edit: Now I see that these ratings include both AO and U.S. Open results. This helps explain the Korda ranking, but it makes the Becker ranking even less explicable given that Becker won two titles at the AO on Rebound Ace.].

Korda won Auz in 2008 and knocked Sampras out of US Open in 2007. Becker was a bit low, but perhaps he was a bit of an opportunist on hard courts. 1991 was knocking off Lendl in the final who wasn't exactly Mr. Clutch in slam finals. 1996 was Kafelnikov and Chang as the big scalps at Auz. Davydenko earned his peak rating with a streak of 3 set domination. The rating in question is weighted for five set matches, but we see the reason for Davydenko's regard on hard courts. I doubt BB would have had such a run. Looking at this peak Davydenko was about 4 years ending in early 2010. Davy lost to the 2010 Auz Open champ Federer in the quarterfinal, 6–2, 3–6, 0–6, 5–7.

 

[Meles]

Does that prove Fed's era is the weakest of all time?

On hard courts afraid not. I would have agreed whole hardedly with you before ELO, but there was depth to the field and that is a big deal at majors in 5 sets formats where some tough, long matches can ruin a players chances. Federer had a healthy field when he was cleaning up, but as we know he didn't have Murray and two other ATGs to deal with. I rate Federer's level of play very highly in 2015 and as you know his record against peak Djokovic was excellent last year. Peak Federer and peak Djokovic would have been very, very close on hard courts. I'd say Nadal had the worst deal, but when he was on he rarely faltered at the final stages of a slam.

For me ELO shows that all three of them would have done exceedingly well in other time periods or with less competition. Did some scenarios on the Big 4 minus one of the other three (probably minus two would have been the fair test.) Everybody gets a few more. Old Feddy springs up to 23-24. Nole needs to get to 16 slams to be in the same rarified air. Federer is the ELO longevity GOAT of all time hands down.

Without Federer and Nadal to complicate things, Djokovic is now facing a normal level of competition on hard courts. Its been far from weak on hard courts.

 

[Meles]

Well according to the QF average peak ELO 2008 to 2012 was a strong as death era on hard courts. Think about it. Davydenko, Djokovic, Del Potro, Nadal, Federer, and Murray. Murray was beating peak Fed lol during this period. Murray and Del Potro had great years in 2009. Its Murray's peak statistical year by a long shot on hard courts. LOL, Roddick was still floating around and dangerous. Those were some loaded fields.

@Gary Duane has been beating me up in conversation about McEnroe having a peak hard court ELO in 1986 weighted for three sets....
"ELO shows McEnroe beating Edberg twice and Connors with many wins in 1986. Djokovic just lost early at Wimbledon. I'd say McEnroe's win rate was quite good in 1985 and 1986. Curren was a God at Wimbledon in 1985 and Lendl was better in 1985 at US Open. No shame losing to Wilander at the French. 1984 did look like a weaker field with Connors the major hurdle. The 7 months off in 1986 had to hurt and his game was becoming outmoded by graphite rackets. He won nothing in 1987 and I'd say that was when age started really catching up to him. It will be interesting to see Fed comeback after a 7 month layoff. With three wins in 1986 it doesn't suprise me that ELO says that is McEnroe's peak. Mary Carillo calls him back in form against Edberg:

Series covering the whole Connor's Final:

Some of the best tennis I've seen McEnroe play. Don't think ELO is much off the mark rating his late 1986 comeback."

Looking at McEnroe's 1984 and the weaker field, you really wonder about matchups and how things play out. With Federer's abscence this year we truly do have a weakening field. 2013-2015 have still been very strong on hard courts. ELO just shows the greatness of the big 4 and you wonder what some of these players might have down with the fields in the 1990s that were up and down.

@Noleberic123 videos fixed in this one. Old Mac.

 

[buscemi]

Becker knocked off Lendl and a bunch of nobodies for his 1989 US Open.

Becker beat:

2R: Rostagno. A dangerous player who reached a career high ranking of #13 in 1991. Had made the U.S. Open QF the prior year. Had a knack for hanging with and beating top players.

3R: Mecir. Two-time Major finalist, including at the Australian Open earlier in the year. Had won Indian Wells earlier in the year. Career-high #4 in 1988.

4R: Pernfors. Career-high #10 in 1986. At the next hard court Major -- 1990 Australian Open -- he would make the QF. Ended up winning Canada in his career.

QF: Noah. Career-high #3 in 1986. French Open champion. Had made the Indian Wells final earlier in the year and would make the SF of the next hard court Major -- 1990 Australian Open

SF: Krickstein. Ended the year #8. Would soon rise to a career-high #6 early in 1990. U.S. Open QF in 1988 and 1990.

F: Lendl. #1 player in the world who won the Australian Open in 1989 and 1990 and was playing in his 8th straight U.S. Open final, having won 3 of them.​

This wasn't the toughest Major draw in the world, but he beat some very good players who were playing some solid hard court tennis in 1989/1990.

 

[Mainad]

Series covering the whole Connor's Final:

Some of the best tennis I've seen McEnroe play. Don't think ELO is much off the mark rating his late 1986 comeback."

Loved the name of that event, "The Transamerica Open". Why don't they have events with names like that in the US anymore?

 

[Noleberic123]

@Noleberic123 videos fixed in this one. Old Mac.

Mac was younger in 86 than Federer this year

 

[Mainad]

Mac was younger in 86 than Federer this year

Certainly was. He was only 27.

 

[buscemi]

Korda won Auz in 2008 and knocked Sampras out of US Open in 2007. Becker was a bit low, but perhaps he was a bit of an opportunist on hard courts. 1991 was knocking off Lendl in the final who wasn't exactly Mr. Clutch in slam finals. 1996 was Kafelnikov and Chang as the big scalps at Auz. Davydenko earned his peak rating with a streak of 3 set domination. The rating in question is weighted for five set matches, but we see the reason for Davydenko's regard on hard courts. I doubt BB would have had such a run. Looking at this peak Davydenko was about 4 years ending in early 2010. Davy lost to the 2010 Auz Open champ Federer in the quarterfinal, 6–2, 3–6, 0–6, 5–7.

In 1988, Becker won Stockholm, including winning a best-of-5 final, then smoked Edberg in the Davis Cup final 6-3, 6-1, 6-4. At the 1988 WTF, he won the title, beating Lendl in a 5 set final.

In 1989, he won Philadelphia, including a best-of-5 final, Paris ("Masters Series"), including a 6-4, 6-4, 6-3 win over Edberg in the final. At the U.S. Open, he won the title, beating Lendl in the final. He also went undefeated in Davis Cup play, smoking Edberg (6-2, 6-2, 6-4) and Wilander (6-2, 6-0, 6-2) in the final. At WTF, he went undefeated before losing in 4 sets to Edberg in the final.

In 1990, at the Australian Open, he beat Mecir in the 4R before losing to Wilander in the QF. He won Belgium, including a best-of-5 final, Indianapolis, Sydney, including a 7-6, 6-4, 6-4 win over Edberg in the final, and Stockholm (Masters Series), including a 6-4, 6-0, 6-3 win over Edberg in the final. At the U.S. Open, he made the SF before losing to Agassi. At WTF, he went undefeated in pool play before losing to Agassi in the SF. (He also made the final of Paris (Masters Series), but had to retire in the first set at 3-3 against Edberg).

In 1990, he won the Australian Open over Lendl.

I would take that version of Becker any day of the week over peak Davydenko.

 

[Meles]

Loved the name of that event, "The Transamerica Open". Why don't they have events with names like that in the US anymore?

A lot of the US events have died off. I suppose the mandatory participation in masters 1000 events has a lot to do with it. Seems like there are less events on tour now. It is amazing how many American events have gone away. The masters events are mighty in the US, but Newport, Memphis, Del Ray, and Houston are weak events these days.

 

[Meles]

In 1988, Becker won Stockholm, including winning a best-of-5 final, then smoked Edberg in the Davis Cup final 6-3, 6-1, 6-4. At the 1988 WTF, he won the title, beating Lendl in a 5 set final.

In 1989, he won Philadelphia, including a best-of-5 final, Paris ("Masters Series"), including a 6-4, 6-4, 6-3 win over Edberg in the final. At the U.S. Open, he won the title, beating Lendl in the final. He also went undefeated in Davis Cup play, smoking Edberg (6-2, 6-2, 6-4) and Wilander (6-2, 6-0, 6-2) in the final. At WTF, he went undefeated before losing in 4 sets to Edberg in the final.

In 1990, at the Australian Open, he beat Mecir in the 4R before losing to Wilander in the QF. He won Belgium, including a best-of-5 final, Indianapolis, Sydney, including a 7-6, 6-4, 6-4 win over Edberg in the final, and Stockholm (Masters Series), including a 6-4, 6-0, 6-3 win over Edberg in the final. At the U.S. Open, he made the SF before losing to Agassi. At WTF, he went undefeated in pool play before losing to Agassi in the SF. (He also made the final of Paris (Masters Series), but had to retire in the first set at 3-3 against Edberg).

In 1990, he won the Australian Open over Lendl.

I would take that version of Becker any day of the week over peak Davydenko.

Becker has 3 YECs on carpet. There is little doubt that on grass and carpet he was vastly superior to Davydenko. Carpet is pretty much a defunct surface these days, but here are the top players slam weighted:
Pos Name Ratings Date
1 John Mcenroe 2597.61 April 8, 1985
2 Ivan Lendl 2479.36 March 31, 1986
3 Boris Becker 2475.17 November 5, 1990
4 Stefan Edberg 2452.24 November 19, 1990
5 Pete Sampras 2394.36 November 2, 1998
6 Marc Rosset 2351.77 November 7, 1994
7 Magnus Larsson 2338.86 February 27, 1995
8 Michael Stich 2325.83 December 13, 1993
9 Greg Rusedski 2323.10 February 22, 1999
10 Goran Ivanisevic 2321.48 November 9, 1992
11 Bjorn Borg 2318.70 March 16, 1981
12 Andre Agassi 2317.20 November 21, 1994
13 Richard Krajicek 2310.38 October 18, 1999
14 Yevgeny Kafelnikov 2275.97 November 30, 1998
15 Jimmy Connors 2227.61 November 12, 1984

I was unaware of Davydenko's late 2009 run. Becker as 16 titles to Davydenko's 8 on hard courts.
Notable Davydenko final wins:
12. April 6, 2008 Miami Masters, Miami, United States Hard Spain Rafael Nadal 6–4, 6–2
18. October 18, 2009 Shanghai ATP Masters 1000, Shanghai, China Hard Spain Rafael Nadal 7–6(7–3), 6–3
19. November 29, 2009 ATP World Tour Finals, London, United Kingdom Hard (i) Argentina Juan Martín del Potro 6–3, 6–4
20. January 9, 2010 Qatar Open, Doha, Qatar Hard Spain Rafael Nadal 0–6, 7–6(10–8), 6–4

Becker's:
3. 26 August 1985 Cincinnati, US Hard Sweden Mats Wilander 6–4, 6–2
6. 18 August 1986 Canada, Canada Hard Sweden Stefan Edberg 6–4, 3–6, 6–3
7. 20 October 1986 Sydney, Australia Hard (i) Czechoslovakia Ivan Lendl 3–6, 7–6(7–2), 6–2, 6–0
10. 23 February 1987 Indian Wells, US Hard Sweden Stefan Edberg 6–4, 6–4, 7–5
23. 10 September 1989 US Open, New York City, US Hard Czechoslovakia Ivan Lendl 7–6(7–2), 1–6, 6–3, 7–6(7–4)
28. 8 October 1990 Sydney, Australia Hard (i) Sweden Stefan Edberg 7–6(7–4), 6–4, 6–4
30. 27 January 1991 Australian Open, Melbourne, Australia Hard Czechoslovakia Ivan Lendl 1–6, 6–4, 6–4, 6–4
45. 28 January 1996 Australian Open, Melbourne, Australia Hard United States Michael Chang 6–2, 6–4, 2–6, 6–2

I'd take Becker narrowly over Davydenko at his best at the US Open from what I've seen (Davydenko was a machine cruising through the early rounds). This late 2009, into early 2010 Davydenko I'm not familiar with, but I'd love to catch some of those matches. You don't beat that many top hard court players without being a very high level. ELO has served its purpose and we see why Davydenko was able to get that peak rating. Becker had some great opportunities. Davydenko was dealing with some very rough hard court players and had injury issues. I'd also say he was more than suspect in the five set format and under major pressure. Boris seized his opportunities. For the purposes of rating the fields at the US Open, a case could be made for altering Davydenko's ELO, but his runs to the SF in 2006 and 2007 are the only fields affected and his level was high for those.

 

[Meles]

Certainly was. He was only 27.

Mac was considered over the hill at that point, but from the looks of these videos I'd say not much. Big mistake on his part taking 7 months off in 1986.

 

[Mainad]

Mac was considered over the hill at that point, but from the looks of these videos I'd say not much. Big mistake on his part taking 7 months off in 1986.

I often wonder about his motivation at that point. He was adopting a party lifestyle and doing drugs and stuff (he had just got married to Tatum). His fitness took a hit and tennis didn't seem to be uppermost in his mind anymore. He needed to get that all out of his system. By the time he did, tennis had largely passed him by with new stars taking centre stage. I don't think he ever paid as much attention to the game again until he went into commentary.

 

[Meles]

I often wonder about his motivation at that point. He was adopting a party lifestyle and doing drugs and stuff. His marriage to Tatum was on the rocks. His fitness took a hit and tennis didn't seem to be uppermost in his mind anymore. He needed to get that all out of his system. By the time he did, tennis had largely passed him by with new stars taking centre stage. I don't think he ever paid as much attention to the game again until he went into commentary.

I was amazed by his form in the youtube clips. I think he and Jimmy were not helped by the technology change. I wonder really how good Connors, Borg, and McEnroe really were.
Here is a more recent hard court list built from 1969 not slam weighted, but only top 10:
Player Year Hard Ct Elo
Roger Federer 2007 2453
Novak Djokovic 2014 2418
Ivan Lendl 1989 2370
Pete Sampras 1997 2356
Rafael Nadal 2014 2342
John McEnroe 1986 2332
Andy Murray 2009 2330
Andre Agassi 1995 2326
Stefan Edberg 1987 2285
Lleyton Hewitt 2002 2262

Borg and Connors are conspicuous in their abscence. Hard courts were a new surface in the 1970s and it may be that the top players just didn't have games built for the surface. Certinaly the early pros that started the open era would have been grass court players. Borg only had 5 titles on hard courts. I don't think either is ATG on hard courts despite their reputation. Maybe I'm off base, but the tour dropped off a lot in level after those 3.

On clay these same ratings have Borg at 2:
Player Year Clay Ct Elo
Rafael Nadal 2009 2550
Bjorn Borg 1982 2475
Novak Djokovic 2015 2421
Ivan Lendl 1988 2408
Mats Wilander 1984 2386
Roger Federer 2009 2343
Jose Luis Clerc 1981 2318
Guillermo Vilas 1982 2316
Thomas Muster 1996 2313
Jimmy Connors 1980 2307

Its really great to see some of these earlier players rated in with Nadal, Djokovic, and Federer on clay.

Here is the slam weighted clay, but only built from 1978 so the earlier players aren't represented properly:
http://simtheworld.blogspot.co.il/2012/10/tennis-elo-ratings-clay.html
Pos Name Ratings Date
1 Rafael Nadal 3142.25 May 18, 2009
2 Roger Federer 2953.74 September 21, 2009
3 Ivan Lendl 2879.27 May 30, 1988
4 Novak Djokovic 2807.51 May 30, 2011
5 Mats Wilander 2800.23 September 23, 1985
6 Gustavo Kuerten 2794.66 July 23, 2001
7 Thomas Muster 2792.64 June 26, 1995
8 Juan Carlos Ferrero 2777.80 July 28, 2003
9 Sergi Bruguera 2775.01 July 12, 1993
10 Jim Courier 2768.41 June 1, 1992
11 Robin Soderling 2763.97 May 31, 2009
12 Guillermo Coria 2749.76 May 31, 2004
13 Gaston Gaudio 2743.86 July 12, 2004
14 Yannick Noah 2736.26 April 23, 1984
15 Carlos Moya 2730.34 July 27, 1998
16 Albert Costa 2716.11 June 3, 2002
17 Yevgeny Kafelnikov 2713.52 July 15, 1996
18 Bjorn Borg 2696.46 April 12, 1982
19 Juan Martin Del Potro 2692.96 May 30, 2011
20 Andres Gomez 2685.94 July 16, 1990
21 David Ferrer 2675.26 April 25, 2011
22 Fernando Verdasco 2665.89 May 3, 2010
23 Fernando Gonzalez 2664.58 May 31, 2009
24 Michael Chang 2661.90 June 5, 1989
25 Guillermo Vilas 2654.95 May 31, 1982
26 Nicolas Almagro 2639.60 February 28, 2011
27 Henri Leconte 2630.14 May 30, 1988
28 Alex Corretja 2622.45 June 1, 1998
29 Mariano Puerta 2621.19 May 30, 2005
30 Michael Stich 2619.86 June 3, 1996
31 Andre Agassi 2618.42 April 17, 2000

These generally match in well with my points leaders since 1991:
http://tt.tennis-warehouse.com/index.php?threads/2016-clay-court-handbook.556878/

I really like the look of these lists and simpler than the handbook.

 

[buscemi]

I'd take Becker narrowly over Davydenko at his best at the US Open from what I've seen (Davydenko was a machine cruising through the early rounds).

I agree that Davydenko was generally a machine when he played weaker opponents at Majors, but he was nothing special against better opponents. Outside of the losses to Federer during his hard court peak, here were Davydenko's matches against his six highest rated opponents:

-Beat #9 Verdasco
-Beat #10 Haas
-Lost to #12 Sodeling
-Lost to #12 Haas
-Beat #13 Berdych
-Lost to #14 Youzhny​

That gels with my memory of Davydenko: In best-of-five matches, he could usually crush lesser opposition, he would always lose to top players, and it was a coin toss whether he would win over solid players.

Meanwhile, 2 best players/hard court players (other than Becker) during Becker's hard court peak were Lendl and Edberg, and he beat both of them in best-of-5 set matches, including beating Lendl in 2 Major finals. Of course, Becker also beat Kafelnikov (#6) and Chang (#5) to win the 1996 Australian Open.

Peak Davydenko might have had a shot to beat peak Becker in a best-of-3 hard court match, but I don't see him having any real shot to beat peak Becker in a best-of-5 match at the U.S. Open

 

[buscemi]

Here is the slam weighted clay, but only built from 1978 so the earlier players aren't represented properly:
http://simtheworld.blogspot.co.il/2012/10/tennis-elo-ratings-clay.html
Pos Name Ratings Date...
13 Gaston Gaudio 2743.86 July 12, 2004...
31 Andre Agassi 2618.42 April 17, 2000...

Yeah, I guess I understand why ELO comes up with this result, but it still doesn't compute for me.

 

[Meles]

I agree that Davydenko was generally a machine when he played weaker opponents at Majors, but he was nothing special against better opponents. Outside of the losses to Federer during his hard court peak, here were Davydenko's matches against his six highest rated opponents:

-Beat #9 Verdasco
-Beat #10 Haas
-Lost to #12 Sodeling
-Lost to #12 Haas
-Beat #13 Berdych
-Lost to #14 Youzhny​

That gels with my memory of Davydenko: In best-of-five matches, he could usually crush lesser opposition, he would always lose to top players, and it was a coin toss whether he would win over solid players.

Meanwhile, 2 best players/hard court players (other than Becker) during Becker's hard court peak were Lendl and Edberg, and he beat both of them in best-of-5 set matches, including beating Lendl in 2 Major finals. Of course, Becker also beat Kafelnikov (#6) and Chang (#5) to win the 1996 Australian Open.

Peak Davydenko might have had a shot to beat peak Becker in a best-of-3 hard court match, but I don't see him having any real shot to beat peak Becker in a best-of-5 match at the U.S. Open

Well. We've established the reason for Davydenko's over rating in the slam weighted ELO stats; an incredible peak outside of the slams. Becker is not in the top ten in either list, great credit to him for getting those hard court majors and it shows what a champ he was and also that being a big server made him most dangerous in slams.

 

[Gary Duane]

I was amazed by his form in the youtube clips. I think he and Jimmy were not helped by the technology change. I wonder really how good Connors, Borg, and McEnroe really were.

Hold on. You are now saying that these players were probably "not very good" because they made it to the top using equipment of the time?

Of course technology changed the game. It always does. But now you can say that players in the 80s and 90s were "not that good" because they were playing without modern strings and rackets.

Its really great to see some of these earlier players rated in with Nadal, Djokovic, and Federer on clay.

As long as you do not take Borg's numbers seriously.

Here is the slam weighted clay, but only built from 1978 so the earlier players aren't represented properly:
http://simtheworld.blogspot.co.il/2012/10/tennis-elo-ratings-clay.html
Pos Name Ratings Date
1 Rafael Nadal 3142.25 May 18, 2009
2 Roger Federer 2953.74 September 21, 2009
3 Ivan Lendl 2879.27 May 30, 1988
4 Novak Djokovic 2807.51 May 30, 2011
5 Mats Wilander 2800.23 September 23, 1985
6 Gustavo Kuerten 2794.66 July 23, 2001
7 Thomas Muster 2792.64 June 26, 1995
8 Juan Carlos Ferrero 2777.80 July 28, 2003
9 Sergi Bruguera 2775.01 July 12, 1993
10 Jim Courier 2768.41 June 1, 1992

I don't have data on Lendl and Wilander because they peaked way before 1991. But the others on my list are very close to what you have there. This works pretty well for 1991 on forward.

 

[Gary Duane]

@Meles

@Gary Duane has been beating me up in conversation about McEnroe having a peak hard court ELO in 1986 weighted for three sets....

I'm not "beating up" you or anyone else.

What I am saying:

Go here:

https://en.wikipedia.org/wiki/John_McEnroe_career_statistics

Go to Singles performance timeline:

A picture is worth a million words.

Look at slam results.

Look at YECs.

Look at Grand Prix tournaments.

Look at year end rankings.

Then look at doubles.

You are analyzing his fall due to better competition and changes in technology. Those may be factors, but you are using Elo to make one of the most dominant years in tennis history (1984) and perhaps years leading up to that as nothing but a "weak era".

I'm saying this: Elo is a tool, and I think it shows things. But it doesn't show everything.

 

[Gary Duane]

@Meles

Borg and Connors are conspicuous in their abscence. Hard courts were a new surface in the 1970s and it may be that the top players just didn't have games built for the surface.

New surface?

At slams, yes. In tennis? Absolutely not. in fact, check back to the time of Pancho Gonzales and look how he dominated HC tennis.

It is ridiculous comparing the 70s with the 90s and 2000s if you are weighting tournaments as they are weighted today.

You are vastly over-weighting slams in an era when they were often skipped. It was not the same as today: there were not clearly marked differences between the Grand Prix/Super 9 tournaments and slams. Pros go where the money is. It's fine to give 2000 points for slams now. They pay the most money.

Back then players were fighting for more money in non-slams. It was like the Wild West compared to now.

Let's put things into perspective: In 1968 the winner got $12,000 for winning the USO. In fact, Ashe only got $20 because he was still technically an amateur.

Top players went where the money was, and it was not at slams.

It took decades to get where we are now. Even in the 90s the AO was frequently skipped and had no where near the rep it has now.

 

[Meles]

Yeah, I guess I understand why ELO comes up with this result, but it still doesn't compute for me.

Pos Name Ratings Date...
13 Gaston Gaudio 2743.86 July 12, 2004...
31 Andre Agassi 2618.42 April 17, 2000...

Agassi is amazingly low is the issue. I'm sure Guadio got a bump from beating Coria in the 2004 final and Coria was hurt greatly by that loss. Coria is surely top ten just based on his points stats. One can't quite argue with ELO on ranking Guadio's peak over Coria's at slams especially when one considers Coria had a shakey serve that crumbled under pressure. The kiss of death in any Slam. My points stats have Guadio 30th for his career, so for peak he's not too high. Another interesting one is Kafelnikov; he has horrible point stats for his career on clay and yet won a French Open and checks in at 17. He beat some shakey clay court players:
Finals 16
Michael Stich
W 764 75 764
Semi-Finals 1
Pete Sampras
W 764 60 62
Quarter-Finals 14
Richard Krajicek
W 63 64 674 62

I suppose these ELOs really aren't that far off. Nadal is a whopping 400 points higher than Kafelnikov and would have obliterated him and his 1996 opponents. Agassi knocked off the 4 seed (Moya) early in 1999 draw and then did not play any other seeds the rest of the way. It really was shocking that he won the tournament. My points stats on Agassi have him at 10. The peak date for Agassi really stands out and I'd say the peak dates are late and that a player can get a boost from decent later play building on previous results (winning French in 1999). McEnroe sees a similar affect and it shows that ELO cannot be instantly responsive nor can any surface specific ranking/rating. I suspect Agassi's high points stats (probably games won too) on clay come from playing a lot on green clay against lesser fields. Agassi is only 100 points or so behind Coria and the others. The fact that he didn't do squat on clay coming into the French might hurt his peak ELO some. I'd say the rating is reasonable for Agassi and perhaps our other Agassi clay stats are misleading. @Gary Duane ?

 

[Meles]

@Meles

I'm not "beating up" you or anyone else.

What I am saying:

Go here:

https://en.wikipedia.org/wiki/John_McEnroe_career_statistics

Go to Singles performance timeline:

A picture is worth a million words.

Look at slam results.

Look at YECs.

Look at Grand Prix tournaments.

Look at year end rankings.

Then look at doubles.

You are analyzing his fall due to better competition and changes in technology. Those may be factors, but you are using Elo to make one of the most dominant years in tennis history (1984) and perhaps years leading up to that as nothing but a "weak era".

I'm saying this: Elo is a tool, and I think it shows things. But it doesn't show everything.

I love looking at Singles performance timelines, but its not the whole story.
If you look at singles titles McEnroe has 9 in 1985:
https://en.wikipedia.org/wiki/John_McEnroe_career_statistics
He won something like 3 of 7 tournaments played after his 7 month layoff in 1986. 1984 is what it is. I think a player's level of play is greatly influenced by what the opponent is doing on the other side of the net. Connors was not a top 10 open era great on hard courts and was surely aging some by 1984. Lendl was not quite prime, but very close. McEnroe feasted on carpet and cleaned up in 1984. I'm amazed at his level in the links from 1986 I sent and it looks as good as 1984. Perhaps Mac didn't have the slam stamina and consistent high level of play in 1986, but he was bringing it both those years (85 and 86).

A slam only ELO would not work and I think ELO still captures McEnroe's potential those years. I think the field ratings for the late 70's and early 80's were quickly labelled very suspect in both the Wimbledon and US Open slam thread and I list the ELOs built from 1969 for top ten to show that McEnroe's rating was low (probably Borg and Connors too).

I remember Roscoe Tanner from the 1980s, but I don't have a great feal for the final 8 then to say that the fields at Wimbledon and the US Open are grossly underrated in 1983-1984.

 

[buscemi]

Pos Name Ratings Date...
13 Gaston Gaudio 2743.86 July 12, 2004...
31 Andre Agassi 2618.42 April 17, 2000...

Agassi is amazingly low is the issue. I'm sure Guadio got a bump from beating Coria in the 2004 final and Coria was hurt greatly by that loss. Coria is surely top ten just based on his points stats. One can't quite argue with ELO on ranking Guadio's peak over Coria's at slams especially when one considers Coria had a shakey serve that crumbled under pressure. The kiss of death in any Slam. My points stats have Guadio 30th for his career, so for peak he's not too high. Another interesting one is Kafelnikov; he has horrible point stats for his career on clay and yet won a French Open and checks in at 17. He beat some shakey clay court players:
Finals 16
Michael Stich
W 764 75 764
Semi-Finals 1
Pete Sampras
W 764 60 62
Quarter-Finals 14
Richard Krajicek
W 63 64 674 62

I suppose these ELOs really aren't that far off. Nadal is a whopping 400 points higher than Kafelnikov and would have obliterated him and his 1996 opponents. Agassi knocked off the 4 seed (Moya) early in 1999 draw and then did not play any other seeds the rest of the way. It really was shocking that he won the tournament. My points stats on Agassi have him at 10. The peak date for Agassi really stands out and I'd say the peak dates are late and that a player can get a boost from decent later play building on previous results (winning French in 1999). McEnroe sees a similar affect and it shows that ELO cannot be instantly responsive nor can any surface specific ranking/rating. I suspect Agassi's high points stats (probably games won too) on clay come from playing a lot on green clay against lesser fields. Agassi is only 100 points or so behind Coria and the others. The fact that he didn't do squat on clay coming into the French might hurt his peak ELO some. I'd say the rating is reasonable for Agassi and perhaps our other Agassi clay stats are misleading. @Gary Duane ?

I think the problem for Agassi under ELO is that his career had so many ups and downs. When you add up his clay court career, it looks pretty impressive: (1) 1 French Open title, 2 other finals (including a five set loss to Courier), 2 other SF, and 4 other QF; and (2) a Rome win, another final, another SF, and another QF. The problem for Agassi is that, in between his 1990-1992 FO run (F, F, SF) and his 1999 title, he had a 1st round loss and a 2nd round loss, dragging down his peak ELO.

The same goes for Kafelnikov, who you mentioned. His peak ELO is calculated in 1996, despite the fact that he was the 2nd/3rd best player at the French Open in 1997, 2000, 2001, losing to Kuerten in 5, 5, and 4 sets. But his poor results in 1998 and 1999 (second round losses), mean he gets no benefit from these matches.

 

[Meles]

@Meles

New surface?

At slams, yes. In tennis? Absolutely not. in fact, check back to the time of Pancho Gonzales and look how he dominated HC tennis.

It is ridiculous comparing the 70s with the 90s and 2000s if you are weighting tournaments as they are weighted today.

You are vastly over-weighting slams in an era when they were often skipped. It was not the same as today: there were not clearly marked differences between the Grand Prix/Super 9 tournaments and slams. Pros go where the money is. It's fine to give 2000 points for slams now. They pay the most money.

Back then players were fighting for more money in non-slams. It was like the Wild West compared to now.

Let's put things into perspective: In 1968 the winner got $12,000 for winning the USO. In fact, Ashe only got $20 because he was still technically an amateur.

Top players went where the money was, and it was not at slams.

It took decades to get where we are now. Even in the 90s the AO was frequently skipped and had no where near the rep it has now.

Again, the slam weighted ratings are built from 1978, so the under importance of slams was less of an issue. You raise a great point about the grass ratings in the other thread. Did the reduced field in Australia bias the rating? I suspect not, but its possible. The abscence of Australia after 1987 on grass would also have an impact.

The ELO numbers built from 1969 onwards are clearly suspect for their first five or six years. You can't rate Rosewall or Gonzales. Connors and Borg had a fair shake in those ratings. The 1969 built ELO top ten numbers on clay and hard are not slam weighted and would have treated all hard events the same. These Sackmann ELO ratings completely circumvent the purse concerns you mention and include all of those events.

I've not really been comparing the 1970s to anything in these threads. I just display the results to show the issues with ELO ratings being built.

 

[buscemi]

Definitely surprised by this result:

13 Gaston Gaudio 2743.86 July 12, 2004
14 Yannick Noah 2736.26 April 23, 1984​

Gaudio's best results at big clay court tournaments coming into the 2004 French Open were (1) a 4th round and two 3rd round losses at prior French Opens; and (2) a SF loss at Monte Carlo in 2000 (no other Masters Series clay results better than a 3rd round loss).

On the other hand, coming into the 1983 French Open, Noah had (1) a 4th round loss (to Connors) followed by two QF losses at the French Open (beating Vilas before losing to Pecci in 1981 and losing to Vilas in 1982); (2) a 1979 Madrid title; (3) a 1980 Rome final; (4) a 1983 Madrid title; and (5) a 1983 Hamburg title, beating Wilander and Higueras. Noah, of course, then went on to beat Lendl (6-0 in the fourth set) and Wilander (straight sets) to win the 1983 title. He then followed that up with a five set QF loss to Wilander in the 1984 French Open and a Rome title in 1985.

Given that Lendl and Wilander are #3 and #5 in peak clay ELO, I'm surprised to see Noah so low. I'm going to guess that guys like Vilas, Connors, and Higueras being underrated by ELO drags down Yannick's rating as well.

 

[Meles]

I think the problem for Agassi under ELO is that his career had so many ups and downs. When you add up his clay court career, it looks pretty impressive: (1) 1 French Open title, 2 other finals (including a five set loss to Courier), 2 other SF, and 4 other QF; and (2) a Rome win, another final, another SF, and another QF. The problem for Agassi is that, in between his 1990-1992 FO run (F, F, SF) and his 1999 title, he had a 1st round loss and a 2nd round loss, dragging down his peak ELO.

The same goes for Kafelnikov, who you mentioned. His peak ELO is calculated in 1996, despite the fact that he was the 2nd/3rd best player at the French Open in 1997, 2000, 2001, losing to Kuerten in 5, 5, and 4 sets. But his poor results in 1998 and 1999 (second round losses), mean he gets no benefit from these matches.

ELO probably has a pretty short tail and once the rankings are built up for the entire tour (~1978-1984) a new player should pretty quickly move up and down based on their play versus the other players. Their is not tail set in the ratings it just goes off matches. Clay and grass would be especially susceptible to Agassi's downs and ups as the season ends and starts 9-11 months later for him. It might be harder for him to build up a peak ELO rating because of this on. On hard courts, I don't think his peak would be underrated because of the number of events on that surface. Your point on Agassi's clay ELO is an excellent one.

I had fun looking at his earlier success on clay from 1990-1992. He really did little on red clay until the French Open those years. I'd say the draw was favorable those years too. He won Stuttgart in 1988 on red clay over Gomez which was excellent. Rome final in 1989. And then 90-92 at the French. The we wait til 1998 for finals in Munich. Won FO 1999 and then Rome in 2002 after to changing to Poly at that tournament no less (beat Tommy Haas in final.) My guess is that if Agassi had been really serious about clay early in his career he might have done better and that would have been his peak time. Agassi was in a horrible slump in 1999 and almost did not play the French. I don't think his peak ELO is too far off the mark. I don't think he was a great clay court player and I suspect his events in lesser red clay tournaments over the years show that. If Agassi had applied himself more in his early years he really looked like he had tremendous potential. I was a huge fan and remember at the time that the results were actually a dissappointment.

1990s clay was a bit odd as the stronger (not strongest) often did not have a great first serve game. In the regular red clay events there was a much higher concentration of these players and they were tough to beat. At majors the draw could open up because the big servers might get hot and get by some of these guys, knocking them out early. Muster is a classic example of one of the strongest of these players. Despite the number of clay court events won his results at the biggest events are not overwhelming because his first serve percentage and points won were not strong assets.

Kafelnikov drives me crazy because his points stats are an abomination on clay for his career. Your listing of him in 3 more years as strong makes me want to look at what peak points Kafelenikov did those seasons:
1996: 82/32 games won and 65/42 points for serve and return (marginal slam winning numbers, but I do like K's first serve points won of 74%, dat dangerous and he proved it at French)
1997: 74/28 games won 60/39 points. The guy won less than 50% of his points on clay, LOL. Weak era on clay.

2000: 70/29 games won 59/40 points. Under 49.5% of points won on clay. Games won under 50% too. ????????
2001: 74/26 games won 60/39. Ugh horrible stats again.

ELO nailed Kafelnikov's peak. It certainly was 1996. I hope he was not 2nd or 3rd best player in the field in the later years. He must have done well in five set format and squeeked by in a lot of matches. I can see that his first serve game if on, might have been strong enough to really help him when his percentage also got hot. 74% first serve points won is very impressive back in 1996 and I'm sure it showed in his game especially if serve percentage got hot. Kuerten was good as squeeking by in a lot of matches and he certainly had a strong first serve game on clay and showed the power of the first serve game in the biggest events. ELO nails Kuerten's rating nicely.

 

[Meles]

Definitely surprised by this result:

13 Gaston Gaudio 2743.86 July 12, 2004
14 Yannick Noah 2736.26 April 23, 1984​

Gaudio's best results at big clay court tournaments coming into the 2004 French Open were (1) a 4th round and two 3rd round losses at prior French Opens; and (2) a SF loss at Monte Carlo in 2000 (no other Masters Series clay results better than a 3rd round loss).

On the other hand, coming into the 1983 French Open, Noah had (1) a 4th round loss (to Connors) followed by two QF losses at the French Open (beating Vilas before losing to Pecci in 1981 and losing to Vilas in 1982); (2) a 1979 Madrid title; (3) a 1980 Rome final; (4) a 1983 Madrid title; and (5) a 1983 Hamburg title, beating Wilander and Higueras. Noah, of course, then went on to beat Lendl (6-0 in the fourth set) and Wilander (straight sets) to win the 1983 title. He then followed that up with a five set QF loss to Wilander in the 1984 French Open and a Rome title in 1985.

Given that Lendl and Wilander are #3 and #5 in peak clay ELO, I'm surprised to see Noah so low. I'm going to guess that guys like Vilas, Connors, and Higueras being underrated by ELO drags down Yannick's rating as well.

Noah's probably a little low because these numbers start in 1978. He doesn't make the top 10 in the other list built from 1969, so I don't think its far from the mark. Again Guadio got a huge bump from beating Coria and winning in 2004. The ELOs are actually very tightly bunched except for Federer, but that is misleading since these are 2012 and back. 2015 and back top ten again for easy reference:
On clay these ratings have Borg at 2:
Player Year Clay Ct Elo
Rafael Nadal 2009 2550
Bjorn Borg 1982 2475
Novak Djokovic 2015 2421
Ivan Lendl 1988 2408
Mats Wilander 1984 2386
Roger Federer 2009 2343
Jose Luis Clerc 1981 2318
Guillermo Vilas 1982 2316
Thomas Muster 1996 2313
Jimmy Connors 1980 2307

Fed has a little seperation, but the top 5 clearly seperate. I prefer to guestimate the older codgers and Djoker in with the slam weighted rankings:
Pos Name Ratings Date
1 Rafael Nadal 3142.25 May 18, 2009
Bjorn Borg 1982 ~3020
Novak Djokovic 2015 ~3000
2 Roger Federer 2953.74 September 21, 2009
3 Ivan Lendl 2879.27 May 30, 1988
XXXXX 4 Novak Djokovic 2807.51 May 30, 2011
5 Mats Wilander 2800.23 September 23, 1985
6 Gustavo Kuerten 2794.66 July 23, 2001
Jose Luis Clerc 1981 ~2793
Guillermo Vilas 1982 ~2793
7 Thomas Muster 2792.64 June 26, 1995
Jimmy Connors ~2780
8 Juan Carlos Ferrero 2777.80 July 28, 2003
9 Sergi Bruguera 2775.01 July 12, 1993
10 Jim Courier 2768.41 June 1, 1992
Yannick Noah ~2766.26 April 23, 1984 (adjusted too 15 with the older codgers thrown in and just ahead of Soderling.)

Kuerten is quite respectable on this list with ELO and I like that.

For Noah, I'd say he's hurt some in the slam weighted ratings as you say, but still doesn't make top ten. His rating is surely a tad low. Another 30 points and he's in there with Courier and Soderling. For rating the fields this is not going to blow anything up too badly. It might be better if these ELO numbers were rounded to the nearest 25 or something and players shown as ties for rankings. 100 points is a significant gap when you do odds on current matches, 200 a wide gap, and 400 an ocean apart. I'm not too worried about Guadio being overrated, because he won't affect my field rating unless he makes the QF. I'd love to correct

I guess I need to do the French Open next. Keep talking about clay all you want here that will just spur me on. I think the field ratings for clay will be very interesting and I'd like to do more guestimates if we have the knowledge.

I have all the data, I wish I had the know how and technology to maintain surface specific ELOs. I have a friend who's worked with databases a lot in the past. I may ask him about the feasibility. I'd just like to read in the PDFs monthly and click a button to calculate the current ratings. Put in a time in the past and calculate ratings with the click of a button. That would be incredibly cool. It may take a computer program to accomplish. Not even sure if VBA code with Excel and access would be up to the task. If someone had the data for pre open era that would be a laugh riot to try with players moving into the pros from amateurs. ELO is the only number or ranking that would have a prayer of negotiating the "amateur" era.

 

[Meles]

Hold on. You are now saying that these players were probably "not very good" because they made it to the top using equipment of the time?

Of course technology changed the game. It always does. But now you can say that players in the 80s and 90s were "not that good" because they were playing without modern strings and rackets.

As long as you do not take Borg's numbers seriously.

I don't have data on Lendl and Wilander because they peaked way before 1991. But the others on my list are very close to what you have there. This works pretty well for 1991 on forward.

haha. In a way Gary I'm saying the opposite. I'm saying they are under rated and the younger guys had the equipment advantage (even Lendl really).

We are adjusting some numbers as Buscemi is doing a great of spotting irregulaties. French Open soon (before Olympics I hope.) Its a lot of fun digging up around the irregularities as you get some unique insights. Peak Kafelnikov and Davydenko no @SoBad. Kafelnikov had nice FO in 1996 with crushingly good 74% first serve points won. Davydenko in late 2009 and early 2010 beat so many ATGs and near ATGs.

It would be great to fully apply ELO to the "woolpants" era which is not as far back as you might think.

 

[buscemi]

ELO probably has a pretty short tail and once the rankings are built up for the entire tour (~1978-1984) a new player should pretty quickly move up and down based on their play versus the other players. Their is not tail set in the ratings it just goes off matches. Clay and grass would be especially susceptible to Agassi's downs and ups as the season ends and starts 9-11 months later for him. It might be harder for him to build up a peak ELO rating because of this on. On hard courts, I don't think his peak would be underrated because of the number of events on that surface. Your point on Agassi's clay ELO is an excellent one.

I had fun looking at his earlier success on clay from 1990-1992. He really did little on red clay until the French Open those years. I'd say the draw was favorable those years too. He won Stuttgart in 1988 on red clay over Gomez which was excellent. Rome final in 1989. And then 90-92 at the French. The we wait til 1998 for finals in Munich. Won FO 1999 and then Rome in 2002 after to changing to Poly at that tournament no less (beat Tommy Haas in final.) My guess is that if Agassi had been really serious about clay early in his career he might have done better and that would have been his peak time. Agassi was in a horrible slump in 1999 and almost did not play the French. I don't think his peak ELO is too far off the mark. I don't think he was a great clay court player and I suspect his events in lesser red clay tournaments over the years show that. If Agassi had applied himself more in his early years he really looked like he had tremendous potential. I was a huge fan and remember at the time that the results were actually a dissappointment.

1990s clay was a bit odd as the stronger (not strongest) often did not have a great first serve game. In the regular red clay events there was a much higher concentration of these players and they were tough to beat. At majors the draw could open up because the big servers might get hot and get by some of these guys, knocking them out early. Muster is a classic example of one of the strongest of these players. Despite the number of clay court events won his results at the biggest events are not overwhelming because his first serve percentage and points won were not strong assets.

Kafelnikov drives me crazy because his points stats are an abomination on clay for his career. Your listing of him in 3 more years as strong makes me want to look at what peak points Kafelenikov did those seasons:
1996: 82/32 games won and 65/42 points for serve and return (marginal slam winning numbers, but I do like K's first serve points won of 74%, dat dangerous and he proved it at French)
1997: 74/28 games won 60/39 points. The guy won less than 50% of his points on clay, LOL. Weak era on clay.

2000: 70/29 games won 59/40 points. Under 49.5% of points won on clay. Games won under 50% too. ????????
2001: 74/26 games won 60/39. Ugh horrible stats again.

ELO nailed Kafelnikov's peak. It certainly was 1996. I hope he was not 2nd or 3rd best player in the field in the later years. He must have done well in five set format and squeeked by in a lot of matches. I can see that his first serve game if on, might have been strong enough to really help him when his percentage also got hot. 74% first serve points won is very impressive back in 1996 and I'm sure it showed in his game especially if serve percentage got hot. Kuerten was good as squeeking by in a lot of matches and he certainly had a strong first serve game on clay and showed the power of the first serve game in the biggest events. ELO nails Kuerten's rating nicely.

At the 1996 French Open, Yevgeny was actually pretty dominant:

1R: Beat Blanco 6-1, 6-3, 6-3
2R: Beat ToJo 6-2, 7-5, 6-3
3R: Beat Mantilla 6-4, 6-2, 6-2
4R: Beat Clavet 6-4, 6-3, 6-3
QF: Beat Krajicek 6-3, 6-4, 6-7, 6-2
SF: Beat Sampras 7-6, 6-0, 6-2
F: Beat Stich 7-6, 7-5, 7-6​

That's 1 set dropped (in a tiebreaker) the whole tournament, with a lot of lopsided sets. And he did all of this while also winning the doubles title.

I suspect that Kafelnikov's FO title is undervalued by ELO because Krajicek (even though he made the Rome final in 1996), Sampras (even though he won Rome in 1994), and Stich (even though he won Hamburg in 1993) aren't regarded as great clay court players. But it's important to note that it was unusually hot at the 1996 French Open, which is why a bunch of fast court players made it so far. That doesn't show up in ELO, but it's a key factor that made those wins better than you might expect.

 

[Gary Duane]

haha. In a way Gary I'm saying the opposite. I'm saying they are under rated and the younger guys had the equipment advantage (even Lendl really).

OK. Thanks for clearing that up.

We are adjusting some numbers as Buscemi is doing a great of spotting irregulaties. French Open soon (before Olympics I hope.) Its a lot of fun digging up around the irregularities as you get some unique insights. Peak Kafelnikov and Davydenko no @SoBad. Kafelnikov had nice FO in 1996 with crushingly good 74% first serve points won. Davydenko in late 2009 and early 2010 beat so many ATGs and near ATGs.

Davydenko also peaked in winning games in 2006 on clay. He is one of many players who probably would have won a slam in an era without Nadal.

 

[Gary Duane]

Again, the slam weighted ratings are built from 1978, so the under importance of slams was less of an issue. You raise a great point about the grass ratings in the other thread. Did the reduced field in Australia bias the rating? I suspect not, but its possible. The abscence of Australia after 1987 on grass would also have an impact.

It may or may not change numbers. We know that even in the early 90s the AO was weak, with some top players not even playing. It's shocking to see Agassi already in QFs of RG and the USO in 1988 but not even playing in the AO until 1995.

The ELO numbers built from 1969 onwards are clearly suspect for their first five or six years. You can't rate Rosewall or Gonzales. Connors and Borg had a fair shake in those ratings. The 1969 built ELO top ten numbers on clay and hard are not slam weighted and would have treated all hard events the same. These Sackmann ELO ratings completely circumvent the purse concerns you mention and include all of those events.

That's good, but something should be weighted, and I suspect today no one knows which tournaments to weight.

It's a shame we do not have more stats before 1991. Before then we have to stumble around in the dark a lot more.

 

[Gary Duane]

On clay these ratings have Borg at 2:
Player Year Clay Ct Elo

Rafael Nadal 2009 2550
Bjorn Borg 1982 2475
Novak Djokovic 2015 2421
Ivan Lendl 1988 2408
Mats Wilander 1984 2386
Roger Federer 2009 2343
Jose Luis Clerc 1981 2318
Guillermo Vilas 1982 2316
Thomas Muster 1996 2313

You might be interested to know they are in this order on % of games won, career, for players since 1991,

Nadal
Djokovic
Muster
Federer

 

[SoBad]

haha. In a way Gary I'm saying the opposite. I'm saying they are under rated and the younger guys had the equipment advantage (even Lendl really).

We are adjusting some numbers as Buscemi is doing a great of spotting irregulaties. French Open soon (before Olympics I hope.) Its a lot of fun digging up around the irregularities as you get some unique insights. Peak Kafelnikov and Davydenko no @SoBad. Kafelnikov had nice FO in 1996 with crushingly good 74% first serve points won. Davydenko in late 2009 and early 2010 beat so many ATGs and near ATGs.

It would be great to fully apply ELO to the "woolpants" era which is not as far back as you might think.

By the way, what is the proper way to refer to pre-1990s tennis, “Woolpants Era” or “Wool Pants Era”? The two expressions seem to be used interchangeably in tennis debates these days.

 

[Meles]

By the way, what is the proper way to refer to pre-1990s tennis, “Woolpants Era” or “Wool Pants Era”? The two expressions seem to be used interchangeably in tennis debates these days.

Two words would be correct. Thank you for the clarification on Wool Pants Era, most would have ascribed this to tennis before World War 2 on seeing the term.

You place this later than the obsolesence of the wood racket. Were Agassi and the great American players of the 90s the end of the Wool Pants Era? This makes sense to me as players like Wilander, Edberg, Lendl, and even Becker would have been raised on wood rackets and only getting the new technolody in their teens (Becker early teens.) The Americans of course paved the way for the great Soviet block players.

 

[Meles]

It may or may not change numbers. We know that even in the early 90s the AO was weak, with some top players not even playing. It's shocking to see Agassi already in QFs of RG and the USO in 1988 but not even playing in the AO until 1995.
That's good, but something should be weighted, and I suspect today no one knows which tournaments to weight.

It's a shame we do not have more stats before 1991. Before then we have to stumble around in the dark a lot more.

Honestly, I don't really care for the weighting. It probably gets Sampras rated correctly, but I think it causes mathematical issues for ELO. We won't be able to evaluate the merits until someone publishes weighted versus normal ELO over the same time period.

I'm not sure that the weighting of Auz as a great major when it was a minor one before 1995 really throws ELO all that much. You still have to beat top ranked players on hard courts or grass in Auz to get boosted to the top of ELO. ELO probably rewards the winner of Auz at the level of just a finalist in the other majors during this early period. One could argue it is far and away the most accurate rating as games and points stats would be thrown off by the lesser field.

ELO is a God send for the open era before 1991. Now we have a tool that covers most of the open era.

 

[Meles]

OK. Thanks for clearing that up.

Davydenko also peaked in winning games in 2006 on clay. He is one of many players who probably would have won a slam in an era without Nadal.

Wow. Davydenko is generally weak on clay like Kafelnikov. 2006? Yes. Holy moly @SoBad . Davydenko had slam winning stats on clay and much better than Kafelnikov:
80/39 on games and 64/54 on points. Wow. That's more than good enough to win a slam. @buscemi this Davydenko guy is not too bad. What a draw where Nalbandian took him out at 2006 French. I think Davy's stats were a bit pumped up from lesser tournaments. What was his ELO? ELO crys foul:
37 Nikolay Davydenko 2591.87 May 30, 2005

It likes Davy in 2005 when he made the SFs and went down to the insane (and likely drugged) Puerta in 5 sets. On the way, Davy defeated prime Coria. Would Davydenko have taken Nadal in the 2005 final?:
44/2008 QF Paris Masters Hard Nikolay Davydenko Rafael Nadal 6-1 0-0 Retired 3.10 - 1.35
17/2008 SF Monte Carlo Masters Clay Rafael Nadal Nikolay Davydenko 6-3 6-2 1.17 - 6.40
13/2008 F Miami Masters Hard Nikolay Davydenko Rafael Nadal 6-4 6-2 2.70 - 1.45
19/2007 SF Rome Masters Clay Rafael Nadal Nikolay Davydenko 7-6(3) 6-7(8) 6-4 1.08 - 11.50
45/2006 RR A Tennis Masters Cup Hard Rafael Nadal Nikolay Davydenko 5-7 6-4 6-4 1.62 - 2.51

One wonders.

 

[Meles]

At the 1996 French Open, Yevgeny was actually pretty dominant:

1R: Beat Blanco 6-1, 6-3, 6-3
2R: Beat ToJo 6-2, 7-5, 6-3
3R: Beat Mantilla 6-4, 6-2, 6-2
4R: Beat Clavet 6-4, 6-3, 6-3
QF: Beat Krajicek 6-3, 6-4, 6-7, 6-2
SF: Beat Sampras 7-6, 6-0, 6-2
F: Beat Stich 7-6, 7-5, 7-6​

That's 1 set dropped (in a tiebreaker) the whole tournament, with a lot of lopsided sets. And he did all of this while also winning the doubles title.

I suspect that Kafelnikov's FO title is undervalued by ELO because Krajicek (even though he made the Rome final in 1996), Sampras (even though he won Rome in 1994), and Stich (even though he won Hamburg in 1993) aren't regarded as great clay court players. But it's important to note that it was unusually hot at the 1996 French Open, which is why a bunch of fast court players made it so far. That doesn't show up in ELO, but it's a key factor that made those wins better than you might expect.

Brilliant. 100% hot temperatures make things play fast. Now Kafelnikov's competition at the French makes sense. I suspect the French can get a little hotter than the earlier clay court tournaments in the year and play faster. This may explain why some of the classic clay courters have a more spotty record at the French.

I believe ELO is spot on. For most years Kafelnikov faced similar issues to the big servers and was not able to do as well at the French. I have much more respect for Kafelnikov's achievement in 1996, but definitely don't think his peak rating is low in any way. I won't complain that it is too high anymore.

Wonderful discussion and analysis of conditions buscemi.

I must ask do you see major changes to the clay court game from ball and surface changes? I suspect Poly strings have had the most impact. I wonder at the differences from the 1990s with all the specialists (no Wilander, Lendl, or Borg to dominate). Why was clay so different in the 1990s? Was it just that the top player were American and just unable to make it their best surface?

 

[Gary Duane]

Honestly, I don't really care for the weighting. It probably gets Sampras rated correctly, but I think it causes mathematical issues for ELO. We won't be able to evaluate the merits until someone publishes weighted versus normal ELO over the same time period.

I'm not sure that the weighting of Auz as a great major when it was a minor one before 1995 really throws ELO all that much. You still have to beat top ranked players on hard courts or grass in Auz to get boosted to the top of ELO. ELO probably rewards the winner of Auz at the level of just a finalist in the other majors during this early period. One could argue it is far and away the most accurate rating as games and points stats would be thrown off by the lesser field.

ELO is a God send for the open era before 1991. Now we have a tool that covers most of the open era.

The reason I think that % of games is a valid indicator for any period is that it works so well from 1991 on. But of course we do not have that earlier.

However, from 1991 you can tell the best players pretty much by number of years of dominance, % of games and most of all from a combination of % of games and ranking. We know that Sampras and Agassi were the best of the 90s.

But for a larger comparison we can also look at slams since 68 and see how games were there. None of this proves anything, but it gives us some trends.

Now we can add Elo to this and see how it all goes together to "bake the cake".

 

[Gary Duane]

Wow. Davydenko is generally weak on clay like Kafelnikov. 2006? Yes. Holy moly @SoBad . Davydenko had slam winning stats on clay and much better than Kafelnikov:
80/39 on games and 64/54 on points. Wow. That's more than good enough to win a slam. @buscemi this Davydenko guy is not too bad. What a draw where Nalbandian took him out at 2006 French. I think Davy's stats were a bit pumped up from lesser tournaments. What was his ELO? ELO crys foul:
37 Nikolay Davydenko 2591.87 May 30, 2005

It likes Davy in 2005 when he made the SFs and went down to the insane (and likely drugged) Puerta in 5 sets. On the way, Davy defeated prime Coria. Would Davydenko have taken Nadal in the 2005 final?:
44/2008 QF Paris Masters Hard Nikolay Davydenko Rafael Nadal 6-1 0-0 Retired 3.10 - 1.35
17/2008 SF Monte Carlo Masters Clay Rafael Nadal Nikolay Davydenko 6-3 6-2 1.17 - 6.40
13/2008 F Miami Masters Hard Nikolay Davydenko Rafael Nadal 6-4 6-2 2.70 - 1.45
19/2007 SF Rome Masters Clay Rafael Nadal Nikolay Davydenko 7-6(3) 6-7(8) 6-4 1.08 - 11.50
45/2006 RR A Tennis Masters Cup Hard Rafael Nadal Nikolay Davydenko 5-7 6-4 6-4 1.62 - 2.51

One wonders.

Yes. I find him at his peak to be very impressive with an amazing return game. But he is in the Nadal period, and no once comes close to Nadal on clay.

 

[NatF]

On the subject of Davydenko his whole SF with Purta is on youtube in good quality..

 

[Meles]

On the subject of Davydenko his whole SF with Purta is on youtube in good quality..

Can't say I was in love with this match. Jumped around a bit to critical junctures and then bailed in the fifth. Gotta say it looked like wind was a factor some of the time and may have benefitted Puerta as Davydenko probably couldn't play his precision power game as well. Watching the final now and already Puerta looks much more impressive. Watching him fight back in the first set at 1-3. The power and drop shots were impressive. Not impressed with Davydenko in SF. Some late 2009 Davydenko next. I was enough of a fair weather fan that I'd never watched Puerta, probably righting him off. Really did not watch clay events as much until more of the big names started showing up.

 

[Meles]

Can't say I was in love with this match. Jumped around a bit to critical junctures and then bailed in the fifth. Gotta say it looked like wind was a factor some of the time and may have benefitted Puerta as Davydenko probably couldn't play his precision power game as well. Watching the final now and already Puerta looks much more impressive. Watching him fight back in the first set at 1-3. The power and drop shots were impressive. Not impressed with Davydenko in SF. Some late 2009 Davydenko next. I was enough of a fair weather fan that I'd never watched Puerta, probably righting him off. Really did not watch clay events as much until more of the big names started showing up.

Davydenko at WTF in 2009 was very, very impressive. Best I've ever seen him move and just added another dimension of toughness and precision to his game. I'd guess after the grind through at the slams he just wasn't able to be at his best later in the slams. In his deep US Open runs I believe he won most of his matches in three easy sets. Anyway, peak Davydenko impressive against near peak Delpo and Fed. Delpo really had no chance.