2016 Clay Court Handbook

[Meles]

The 2016 Clay Court Season may be the best ever. Federer seems set to return in Monte Carlo and will likely now have to play all the major clay events to get back into form and maintain his ranking. Will Djokovic finally win the French Open? Will Murray put in another season as Clayray? Will Thiem continue his winning ways? (4 clay wins in 10 months)

The following table looks at the key statistics for all time great players on clay along with 2015's top players and some data from 2016:

*Call it the Bolelli effect, but Simone's numbers are a bit inflated from playing some challenger events. This is the case for some others perhaps.

Update 5/23/2016: 2016 up to French Open!
Update 5/9/2016: 2016 through Madrid. Also updated a week ago.
Update 4/7/2016: New tables. These are all detailed points stats computated from the raw points totals. Dominance Ratio is now in decimal form. Points Won is percent of points won in matches. Weighted Dominance Ratio weights Break Point performance as 25% to 75% raw points performance. Its a similar number that tries to factor clutch play. These tables now use lookup data so the 2016 numbers will be very easy to update (check back on Monday evening after major clay events for updates.)
Update 4/4/2016: New table. These are all points based stats. The clutch serve and return shows the differential between regular points and break points. Its normal for the clutch serve to be a little negative as can be seen by all time greats. The Dominance Ratio is the players return points won divided by their serve points lost, which is simply 100% minus the players serve points won. The Dominance Ratio is multiplied by 100 for readability.

Nadal's career domination in these stats is easily seen with the top numbers in all but % First Serve Points Won. Nadal's domination ratio is off the charts.

The numbers for the top players last year (2015) were very strong as a group with some world class numbers for Murray, Nishikori and even Berdych. Djokovic's serve numbers were unbelievable in 2015, but his Break Points Converted number was 5% below his career average! Murray's Break Point Saved and Converted rates were better than Nadal's career average! Federer did not appear to have slipped much in 2015 except for his lowish Break Points Saved Number.

Nadal and Ferrer's numbers show a disturbing downward trend in their return games in 2015 and 2016. Nadal despite all of his serve and return numbers slipping still has retained a stellar Break Point Conversion % in 2016. Nadal's 2015 Break Points Saved was very weak. Almagro's return is also poor this year, but his serving numbers are better than ever so far.

Thiem has emerged as a top clay courter. For 2015 his 1st Serve Points Won % is only eclipsed by Kuerten's and Almagro's career stat. Thiem's 2nd Serve Points Won was also top drawer. 2nd Serve Return Points is another strong Thiem stat. In 2016 Thiem has improved his 1st Serve Points Won. HIs Break Points Converted despite a poorer 2nd Serve Return Points Won %, indicates clutch play on Thiem's part for 2016.
Update 4/7/2016:
Thiem in the last three hard courts tournaments has upped his number to a tour leading 39% on 1st Serve Return points (up from 26% last year. ). If this transfers to the clay courts when team returns to them in Monte Carlo it will elimate his major weakness.

Cuevas has achieved stellar serving stats with two victories on clay so far this year.

These are all clay court stats and the serving numbers for the all time clay greats. The serving numbers are not the best ever on clay; those belong to the Servebots.

Edit: I've added Points Dominance ratio in the center column. It shows that we are in a great clay court era with Federer and Djokovic also having some of the best all time clay court performance. (Ferrer is not bad too at 53% of points won on clay for his career.)

Federer's overall game was above his average last year and its very obvious that the changes to his game are working for his serve games (his return game is down 2% though.) Nishikori had slam worthy numbers last year as did Ferrer who was substantially above his career average for clay.

For 2016 the numbers for Ferrer and Nadal are shockingly low so far.

Thanks to @Gary Duane and @falstaff78 for their ideas and introduction to dominance ratios.

 

[Liam Fitzpatrick]

Great Analysis hopefully someone can stop Novak on clay. IT seems pushing works against him as he played awfully against Simon in Melbourne.

 

[Meles]

Great Analysis hopefully someone can stop Novak on clay. IT seems pushing works against him as he played awfully against Simon in Melbourne.

Yes I think these stats are great since they really tell the tale about the great clay court players of the past. By only showing the stats for these top clay courters it really shows what is a good enough serve on clay. Gulbis is not on this but his breakpoint numbers for his career are 62/42 which matches Ferrero and Moya, so hopefully Gulbis can bring back his top form this year and add even more spice to the season. Gulbis was 63/49 and 78% first serve won in 2014 and 63/48 and 76% in 2013. Gulbis had 203 aces in 22 matches on clay in 2014 which was 2nd though Raonic, Isner, and Karlovic averaged more aces per match. Prime Gulbis would be a danger to anyone in a given match.

Djokovic's First Serve Return Points Won was weak last year. This is what probably lowered his break point conversion rate so much (down a whopping 5% from his career average.) Thiem has a strong serve game on clay and his giant super kicker first serve might allow him to hold serve at a good rate against Djokovic even on hard courts. I'd be very happy to see them draw each other in Indian Wells round of 16. The first serve return on clay appears to be Nole's Achilles heel.

 

[Liam Fitzpatrick]

Wow finally someone who can leave me stunned by a reply on a tennis or football(soccer). forum. Brilliant knowledge and in training to go into the industry of sports journalism in the uk

 

[Chanwan]

Cheers,
however, I tend to think too much data can muddle the picture. Personally I like @Gary Duane 's percentage of games won quite a bit (or expressed as in how many games out of every 200, you'll win). And @falstaff78 's dominance ratios.
Nevertheless, good job on gathering the data!

Thiem vs. Wawa in the FO-final would be great. But I feel Thiem still has some growing to do, before he emerges as one of the true top dogs.

 

[I Am Finnish]

Great Analysis hopefully someone can stop Novak on clay. IT seems pushing works against him as he played awfully against Simon in Melbourne.

Wawrinka stopped him last year only at rg

 

[Meles]

Wawrinka stopped him last year only at rg

Djokovic is stoppable if you can dominate his first serve return. Wawrinka has a strong service game on clay. Wawrinka's typical return game is another matter. How that got through for breaks against Djoko is the amazing thing. I suppose this is how Fed derailed him in 2011. Kygrios made a final at Estoril last year on clay. His stats stink, but with a red hot serve and his new return he might be able to make inroads against Djokovic.

 

[Meles]

Cheers,
however, I tend to think too much data can muddle the picture. Personally I like @Gary Duane 's percentage of games won quite a bit (or expressed as in how many games out of every 200, you'll win). And @falstaff78 's dominance ratios.
Nevertheless, good job on gathering the data!

Thiem vs. Wawa in the FO-final would be great. But I feel Thiem still has some growing to do, before he emerges as one of the true top dogs.

I like all of those stats and I might add dominance ratio to this table since I think I can easily calculate it with a formula

I don't tire or find it muddling when the data does seem to correlate with the past greats on clay. The detailed data here seems to at least identify signs of weakness or strength though we see that the 2016 data shows that break point stats don't always follow in lock step with the points won data. The Djokovic weakness last year on 1st serve returns is a nice piece of information that does not follow from the ratios.

Something about Gary''s games won stuff seems to almost state the obvious. Points won is a little more obtuse and that just might make it better. Djokovic won almost everything he played on clay last year, so of course his won numbers look great.

 

[MichaelNadal]

Lol, the OP makes Ultron sweeping the clay season seem mighty exciting.

 

[Chanwan]

I like all of those stats and I might add dominance ratio to this table since I think I can easily calculate it with a formula

I don't tire or find it muddling when the data does seem to correlate with the past greats on clay. The detailed data here seems to at least identify signs of weakness or strength though we see that the 2016 data shows that break point stats don't always follow in lock step with the points won data. The Djokovic weakness last year on 1st serve returns is a nice piece of information that does not follow from the ratios.

Something about Gary''s games won stuff seems to almost state the obvious. Points won is a little more obtuse and that just might make it better. Djokovic won almost everything he played on clay last year, so of course his won numbers look great.

I just think that % of winning games - or dominance ratio – is a better predictor, because you get a sense of what's going on.
For instance, Thiem on clay this year: 107 out of every 200 games won on clay: http://www.atpworldtour.com/en/players/dominic-thiem/tb69/player-stats?year=2016&surfaceType=clay (last year, 110).
That tells us, he hasn't been blowing opposition off the court and I think it gives us a truer picture of his abilities right now than data indicating he's closing in on all time great stats.
Over the course of his career, Birdman has 109 in this category across surface - and 111 for last year (108 on clay, 109 in 2014).

In other words - Thiem is around Birdman clay territory, not all time clay greats, as of now.

I like dominance-ratio as well, but I don't have the same understanding of it in terms of what's good, great or moderately great as I have on this more simple stat.

 

[vive le beau jeu !]

cuevas' stats are indeed impressive...

 

[Meles]

Lol, the OP makes Ultron sweeping the clay season seem mighty exciting.

Well looking at last year your might think that this will 2015 with Djoko winning the French, but I see at least two moving targets. Murray who almost beat him in the French semi last year might up his game on clay. His overall stats were superior to Djokovic last year on clay. The other moving target is Thiem who might have the kind of first serve to take advantage of Djokovic first serve return. Thiem making later rounds of these tournaments would be very interesting since he's likely to keep improving for several years.

 

[Meles]

cuevas' stats are indeed impressive...

Thanks. It was fun gathering them and looking at them. We'll see if they end up being insightful for some upcoming matches on clay between these players.

 

[Meles]

Substantially updated the handbook. It now has the very cool points dominance ratio!

This shows that the numbers last year from Federer, Murray, and Djokovic were at the highest levels. Even Ferrer and Nishikori had strong years. The ratio shows very poor performance for Nadal and Ferrer this year on clay, while Cuevas looks set to do some damage. Poor Dominic's solid numbers still leave him below 4 of these players. Fortunately, Thiem appears to be a statistical baby Stanimal so he still might be able to break through this extremely high level of play.

Get ready for some great clay court tennis!

 

[MichaelNadal]

Well looking at last year your might think that this will 2015 with Djoko winning the French, but I see at least two moving targets. Murray who almost beat him in the French semi last year might up his game on clay. His overall stats were superior to Djokovic last year on clay. The other moving target is Thiem who might have the kind of first serve to take advantage of Djokovic first serve return. Thiem making later rounds of these tournaments would be very interesting since he's likely to keep improving for several years.

I really think Thiem has a lot of potential!

 

[Sysyphus]

Well looking at last year your might think that this will 2015 with Djoko winning the French, but I see at least two moving targets. Murray who almost beat him in the French semi last year might up his game on clay. His overall stats were superior to Djokovic last year on clay. The other moving target is Thiem who might have the kind of first serve to take advantage of Djokovic first serve return. Thiem making later rounds of these tournaments would be very interesting since he's likely to keep improving for several years.

Yes!

Embrace the Clay-Muzziah!

This shall be the year, I am sure.

 

[LinePainter]

Happy to see Thiem is improving on his 30% 1st serve return won, to have the same returning statistics as Berdych is no bueno. Especially, considering that his quality of opponent in 2015 was much lower than Berdych's. It seems that 35% would be a good goal to reach.

 

[3fees]

Great Analysis hopefully someone can stop Novak on clay. IT seems pushing works against him as he played awfully against Simon in Melbourne.

I dont think pushing was it, Djokovic for some reason played like Simon and strayed from usual playing style..

Cheers
3fees

 

[Meles]

I dont think pushing was it, Djokovic for some reason played like Simon and strayed from usual playing style..

Cheers
3fees

Tomic is a flat hitting player and gave Djokovic a good run for a set in Shanghai. Both have been called pushers (but that hardly does either justice.) I rememeber when Tomic drove Djoko nuts way back at Wimbledon. Don't koow of anyone else who can match the Simon formula. I think the best surface for that tactic happens to be Djokovic's best surface.

 

[falstaff78]

Awesome work! @Meles

Hard to believe that Murray's 2015 was the equivalent of Rafa's career!

 

[3fees]

Tomic is a flat hitting player and gave Djokovic a good run for a set in Shanghai. Both have been called pushers (but that hardly does either justice.) I rememeber when Tomic drove Djoko nuts way back at Wimbledon. Don't koow of anyone else who can match the Simon formula. I think the best surface for that tactic happens to be Djokovic's best surface.

definition:

In tennis, a pusher is a defensive player who "pushes" back any shot they can chase down, without deliberately hitting a winner. They can angle shots, aim deep, as well as produce effective lobs. Pushers are extremely quick and consistent, rarely making errors.

Djokovic hits winners and plenty of them, at times he is a pusher, yet so is ever tennis player,,pusher -backboard,ect,lol

Djokovic game has changed as he has played, earlier he outlasted them and ran them out of gas, now he hits a variety of shots so they cant predict what he'll do next, he out thinks his opponents and plain and simple outskills them. he has combined physical and mental aspects of tennis thats way above the field.

Cheers
3Fees

 

[Meles]

Awesome work! @Meles

Hard to believe that Murray's 2015 was the equivalent of Rafa's career!

He'd probably win the French this year if @Mainad would back him in the match threads.

Everybody has been down on Murray, but another AO final with nice win over incredible Raonic. He should maintain or up his level on clay this year. Last year, Djoko's first serve return was very suspect on clay. If Djoko's stats slips some Murray really has a shot at even bigger clay court titles. I suspect that Murray will have a hard time maintaining those numbers as his career numbers don't support such lofty heights. Its hard to believe that Djoko's first return will worsen this year on clay; Nole just looks so good this year and late last year.

 

[stringertom]

Euroclay '16 is going to be tough for anyone to totally dominate. It's quite possible a #1 or #2 seed could face both Cuevas and Thiem before the QFs of any of the MS1000's or RG. Imagine that gauntlet with another couple of toughies from the likes of Nadal or Ferrer and then Fed or Stanimal before lining up against the top seed from the other half of the draw.

It would be great if the three 1000's get split by two or three different champions. Madrid might be watered down a bit by Djokovic's expected absence but MC and Rome could be wide open too.

 

[Meles]

Euroclay '16 is going to be tough for anyone to totally dominate. It's quite possible a #1 or #2 seed could face both Cuevas and Thiem before the QFs of any of the MS1000's or RG. Imagine that gauntlet with another couple of toughies from the likes of Nadal or Ferrer and then Fed or Stanimal before lining up against the top seed from the other half of the draw.

It would be great if the three 1000's get split by two or three different champions. Madrid might be watered down a bit by Djokovic's expected absence but MC and Rome could be wide open too.

Thiem get's a top seed in the round of 16 since he will be 13th seed!

Updated the table with some more state of the art points stats. (The raw data is the same.) King Rafa's dominance ratio was off the charts! I plan to update the 2016 numbers with the results before Madrid, and then Rome, and then the French Open. For now most of the players have not played on clay this year.

 

[Meles]

Massive update to stats in original post. Will update the 2016 numbers after Monte Carlo and every major tournament thereafter (easily done with six pastes of the data from the ATP site and saving over previous images.)

 

[Gary Duane]

The 2016 Clay Court Season may be the best ever. Federer seems set to return in Monte Carlo and will likely now have to play all the major clay events to get back into form and maintain his ranking. Will Djokovic finally win the French Open? Will Murray put in another season as Clayray? Will Thiem continue his winning ways? (4 clay wins in 10 months)

The following table looks at the key statistics for all time great players on clay along with 2015's top players and some data from 2016:

*Call it the Bolelli effect, but Simone's numbers are a bit inflated from playing some challenger events. This is the case for some others perhaps.

Update 4/7/2016: New tables. These are all detailed points stats computated from the raw points totals. Dominance Ratio is now in decimal form. Points Won is percent of points won in matches. Weighted Dominance Ratio weights Break Point performance as 25% to 75% raw points performance. Its a similar number that tries to factor clutch play. These tables now use lookup data so the 2016 numbers will be very easy to update (check back on Monday evening after major clay events for updates.)
Update 4/4/2016: New table. These are all points based stats. The clutch serve and return shows the differential between regular points and break points. Its normal for the clutch serve to be a little negative as can be seen by all time greats. The Dominance Ratio is the players return points won divided by their serve points lost, which is simply 100% minus the players serve points won. The Dominance Ratio is multiplied by 100 for readability.

Nadal's career domination in these stats is easily seen with the top numbers in all but % First Serve Points Won. Nadal's domination ratio is off the charts.

The numbers for the top players last year (2015) were very strong as a group with some world class numbers for Murray, Nishikori and even Berdych. Djokovic's serve numbers were unbelievable in 2015, but his Break Points Converted number was 5% below his career average! Murray's Break Point Saved and Converted rates were better than Nadal's career average! Federer did not appear to have slipped much in 2015 except for his lowish Break Points Saved Number.

Nadal and Ferrer's numbers show a disturbing downward trend in their return games in 2015 and 2016. Nadal despite all of his serve and return numbers slipping still has retained a stellar Break Point Conversion % in 2016. Nadal's 2015 Break Points Saved was very weak. Almagro's return is also poor this year, but his serving numbers are better than ever so far.

Thiem has emerged as a top clay courter. For 2015 his 1st Serve Points Won % is only eclipsed by Kuerten's and Almagro's career stat. Thiem's 2nd Serve Points Won was also top drawer. 2nd Serve Return Points is another strong Thiem stat. In 2016 Thiem has improved his 1st Serve Points Won. HIs Break Points Converted despite a poorer 2nd Serve Return Points Won %, indicates clutch play on Thiem's part for 2016.
Update 4/7/2016:
Thiem in the last three hard courts tournaments has upped his number to a tour leading 39% on 1st Serve Return points (up from 26% last year. ). If this transfers to the clay courts when team returns to them in Monte Carlo it will elimate his major weakness.

Cuevas has achieved stellar serving stats with two victories on clay so far this year.

These are all clay court stats and the serving numbers for the all time clay greats. The serving numbers are not the best ever on clay; those belong to the Servebots.

Edit: I've added Points Dominance ratio in the center column. It shows that we are in a great clay court era with Federer and Djokovic also having some of the best all time clay court performance. (Ferrer is not bad too at 53% of points won on clay for his career.)

Federer's overall game was above his average last year and its very obvious that the changes to his game are working for his serve games (his return game is down 2% though.) Nishikori had slam worthy numbers last year as did Ferrer who was substantially above his career average for clay.

For 2016 the numbers for Ferrer and Nadal are shockingly low so far.

Thanks to @Gary Duane and @falstaff78 for their ideas and introduction to dominance ratios.

Man, how in hell are you doing all this work? If you are going from raw data from the ATP it takes forever.

Or do you have some other source that is giving you some of these stats with one decimal?

Either way, great work!

 

[Gary Duane]

Cheers,
however, I tend to think too much data can muddle the picture. Personally I like @Gary Duane 's percentage of games won quite a bit (or expressed as in how many games out of every 200, you'll win). And @falstaff78 's dominance ratios.
Nevertheless, good job on gathering the data!

Thiem vs. Wawa in the FO-final would be great. But I feel Thiem still has some growing to do, before he emerges as one of the true top dogs.

This is really good because Meles is including a weighted DR. I'm not sure exactly what he did, but it looks about right. I assume he is sorting based on that.

The flaw with % of games is that it is only very relative when comparing two players who have the same % of service games.

In other words, it is good for comparing 89/32 to 89/30 (the better record is obvious) but not good for comparing 89/25 to 84/30. When the total is the same but there is a 5% difference in serving and return, the guy with the higher % on serve will have a much better record and higher % of games.

But I still don't now why...

And here is an example, career:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37-----54.64% +4
Murray--------Grass: 89.20/26.10 115.30-----DR 1.28-----53.79% +2

It appears that Murray won more games per match, right? But he did not. His % of points is almost 1% down, which is HUGE because the range is rather small from low to high. The DR is in line with % of games. Clutch is to the advantage of Sampras, but not hugely so.

In contrast:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37------54.64% +4
Federer-------Grass: 92.42/24.55 116.97-----DR 1.36------54.62% +3

Now they are REALLY close because they both were neck and neck on serving. ATP show 93% of Pete, 92% for Fed, and this is a perfect example of why the rounding is horrible. For these two, on the basis of stats, I would have to judge them equal. They are head and shoulders over everyone else in the past 26 years, including Novak - on grass.

 

[Meles]

Man, how in hell are you doing all this work? If you are going from raw data from the ATP it takes forever.

Or do you have some other source that is giving you some of these stats with one decimal?

Either way, great work!

Haha. Excel just copy and paste special as text and it goes right in with a little cleanup. I create a sheet for each statistic and then us vlookup to get the data (sometimes quite a few in an equation.) The neat trick will be that for 2016 I'll be able to update all the data periodically with relative ease as all the work is done. Just copy and paste a little bit for the six stats and its done. I host on dropbox for free and literally just save over the old file to update the images.

For the all time greats data some of the players like Berrastigue aren't in the top 200 (First serve for him). Then it doesn't work. For annual data I can do all of the players just like 2016. Got to read the rest of the posts, but we can play around with DR formulas etc. very easily.

 

[Meles]

This is really good because Meles is including a weighted DR. I'm not sure exactly what he did, but it looks about right. I assume he is sorting based on that.

The flaw with % of games is that it is only very relative when comparing two players who have the same % of service games.

In other words, it is good for comparing 89/32 to 89/30 (the better record is obvious) but not good for comparing 89/25 to 84/30. When the total is the same but there is a 5% difference in serving and return, the guy with the higher % on serve will have a much better record and higher % of games.

But I still don't now why...

And here is an example, career:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37-----54.64% +4
Murray--------Grass: 89.20/26.10 115.30-----DR 1.28-----53.79% +2

It appears that Murray won more games per match, right? But he did not. His % of points is almost 1% down, which is HUGE because the range is rather small from low to high. The DR is in line with % of games. Clutch is to the advantage of Sampras, but not hugely so.

In contrast:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37------54.64% +4
Federer-------Grass: 92.42/24.55 116.97-----DR 1.36------54.62% +3

Now they are REALLY close because they both were neck and neck on serving. ATP show 93% of Pete, 92% for Fed, and this is a perfect example of why the rounding is horrible. For these two, on the basis of stats, I would have to judge them equal. They are head and shoulders over everyone else in the past 26 years, including Novak - on grass.

I liked the look of the weighting too. Its the same calculation. Easier to just show the formula:
((0.25*M2+J2*0.75)
(1-(0.25*E2+0.75*F2)))
M2 and E2 are the respective break point numbers. J2 and F2 are the normal serve and return numbers.

I'll respond to the rest in a bit.

 

[Gary Duane]

I liked the look of the weighting too. Its the same calculation. Easier to just show the formula:
((0.25*M2+J2*0.75)
(1-(0.25*E2+0.75*F2)))
M2 and E2 are the respective break point numbers. J2 and F2 are the normal serve and return numbers.

I'll respond to the rest in a bit.

It seems to work.

A refinement would be to not treat service and return +/1 the same way, but that will usually not make any difference.

For instance, Fed is this on HCs this year: -16/+2

The reason why this is so bad is that the choking is on serves. It would also be bad on return, but here it is even worse.

 

[every7]

Djokovic is stoppable if you can dominate his first serve return. Wawrinka has a strong service game on clay. Wawrinka's typical return game is another matter. How that got through for breaks against Djoko is the amazing thing. I suppose this is how Fed derailed him in 2011. Kygrios made a final at Estoril last year on clay. His stats stink, but with a red hot serve and his new return he might be able to make inroads against Djokovic.

Thanks @Meles for this. It's great to see someone supply interesting numbers and also not just add to the crowd of Djokovic doomsayers who just assume he is going to win everything. Strange things happen in tennis (particularly the FO) and it is great that you don't lose that perspective.

 

[Gary Duane]

What has happened is that@Meles has become the GOAT of statistical computations here.

Either that or he has Deep Throat status at the ATP and gets the proper figures, while the rest of us only see the rounded data!

 

[Meles]

This is really good because Meles is including a weighted DR. I'm not sure exactly what he did, but it looks about right. I assume he is sorting based on that.

The flaw with % of games is that it is only very relative when comparing two players who have the same % of service games.

In other words, it is good for comparing 89/32 to 89/30 (the better record is obvious) but not good for comparing 89/25 to 84/30. When the total is the same but there is a 5% difference in serving and return, the guy with the higher % on serve will have a much better record and higher % of games.

But I still don't now why...

And here is an example, career:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37-----54.64% +4
Murray--------Grass: 89.20/26.10 115.30-----DR 1.28-----53.79% +2

It appears that Murray won more games per match, right? But he did not. His % of points is almost 1% down, which is HUGE because the range is rather small from low to high. The DR is in line with % of games. Clutch is to the advantage of Sampras, but not hugely so.

In contrast:

Sampras------Grass: 92.66/21.63 114.29-----DR 1.37------54.64% +4
Federer-------Grass: 92.42/24.55 116.97-----DR 1.36------54.62% +3

Now they are REALLY close because they both were neck and neck on serving. ATP show 93% of Pete, 92% for Fed, and this is a perfect example of why the rounding is horrible. For these two, on the basis of stats, I would have to judge them equal. They are head and shoulders over everyone else in the past 26 years, including Novak - on grass.

When you think of big servers you thing of close sets on games, but often going the servers way. Big servers would have easier holds where they rack up the points too. So without really anaylzing the numbers, a big server might have less games dominance and still win more matches. Their points dominance might be overstated at the same time.

On clay, your big serve might be Kuerten and Berasategui the opposite or Muster has a weak first serve points won (see the table for field order, sorry.):

Thomas Muster 67.8% 54.1% 0.3% 63.1% 62.8% 53.4% 1.19 1.21 44.4% 35.6% 57.0% 46.6% 2.2%
Alberto Berasategui 63.0% 49.6% -1.4% 58.6% 60.0% 52.0% 1.09 1.09 43.5% 35.3% 57.4% 44.5% 1.0% (The numbers that have .0% are the crude numbers from the players stat page.)
Gustavo Kuerten.....73.8% 50.8% -0.5% 62.9% 63.4% 52.4% 1.14 1.14 41.7% 34.5% 51.7% 42.9% 1.3%

Muster had a lot of titles, but really was a one hit wonder at the French Open, with only one SF and a QF to go with his amazing 42 clay court titles at lesser tournaments (3 Monte Carlo and 3 Rome with no Hamburg). Alberto B had one Final and 15 clay court titles and no big ones. Kuerten had 14 titles, but 3 were French Opens and 4 atp1000 equivalents on clay. Kuerten's career was much shorter and I suspect he was not trying to play every clay court event he could.

Of the all time greats on clay Kuerten and Courier did very well on first serve points. They also did very well on 1st serve return points considering they fall back in the pack on 2nd serve returns. If you sort by first serve return. Those players that are below average on first serve return points on clay are generally the clay wanna be's with the exception of Kafelnikov, who has by far the worst stats of any French Open Champion on clay. So the first serve dynamic seems to be important on clay and I suspect it is on other surfaces for winning big titles. 2nd return points seem to have some importance, and 2nd serve points performance is firmly in 4th place for the all time greats. Serve points won in total correlates well with the best players on clay, but not so much return points won. The all time greats that don't win a lot of serve points on clay don't have any FO titles on clay with the exception of Gaudio who beat Coria and his weak serve in 2004. So maybe the best weighting for big clay court tournament success would be for serve points won (or both serves) and 1st serve return points.

With the two grass examples you mention there is no doubt that adding the games percents is folly. The points DR does a great job of seperating Sampras from Murray (my table is points DR). Federer's better return numbers on games don't add up to more points which is interesting.... Its very interesting when you consider that Sampras had a reputation for loafing on return games once he had a break. You would think that would hurt his points totals. When you look at their first serve points their is a huge gulf with Sampras winning 85% of his first serve points to Federer's 79%. That must be the difference. I suspect Sampras had a lot of lopsided, easy service holds. Federer reverses this on the grass return points and is a far better at winning first return points. Sampras is much better on 2nd serve points on grass. Sampras won more points on serve by roughly 1% and Federer on returns by 1%. I suspect Federer was able to break more on returns because of that, but even more so with his vastly superior play on first serve points.

The difference in their return games says a lot and Pete is very, very impressive on the 2nd serve return on grass.
13 Andre Agassi 56.1%
14 David Ferrer 55.6%
15 Stefan Edberg 56.1%
17 Michael Chang 55.5%
18 Petr Korda 55.6%
22 Andy Murray 55.2%
24 Todd Martin 54.6%
32 Lleyton Hewitt 53.7%
33 Pete Sampras 53.5%
35 Boris Becker 53.7%
43 Tim Henman 53.0%
44 Patrick Rafter 52.8%
46 Rafael Nadal 53.2%
63 Novak Djokovic 52.3%
72 Michael Stich 51.5%
94 Roberto Bautista Agut 51.7%
96 Roger Federer 51.4%
97 Goran Ivanisevic 50.6%
162 Richard Krajicek 49.0%
164 Mark Philippoussis 49.0%
200 Andy Roddick 47.7%

The above numbers filtered down show some pretty fine grass court players winning a lot of 2nd serve points on grass. I'm amazed at how poor Federer is on that statistic. Roddick did reasonably well at Wimbledon, but he had a fantastic first serve that probably helped him compete at Wimbledon. The story is much the same as the more Servebot like players are down on this list. Edberg, Sampras, Becker, Rafter, and Pete are amazingly high. Djokovic is suprsingly low too. Murray has dominated this number in recent years clocking in at 59% to Djokoivc at 52% in 2015.

The numbers for Murray are quite good and its rather amazing he's not won more Wimbledons. I'd say Djokovic was very, very lucky last year that Federer took out Murray in the semifinals and that Murray was not fully recovered from his back operation in 2014.

What this says to me is that on grass the first serve is extremely important. The second serve is meaningless (Boris Becker is not in the top 200). First serve return does not matter so much (think of playing Sampras.) 2nd serve return points are huge on grass. Murray would wry havoc at Wimbledon if he had a better first serve game. Its little surprise he's been SF or better since 2009 (2014 injury year the exception).

Given these numbers and trends I'd say Pete Sampras was the best of all time on grass and Djokovic is fortunate to have the titles that he has. I can see why he's never taken a set from Murray at Wimbledon. Federer has amassed his titles with a major hole in his game with the mediocre performance on 2nd serve return points.

This certainly gives me an idea on how to do a weighted points DR on grass and maybe I'll have to work on a grass court handbook around the French Open or just before. I think this shows that the basic points DR you've been using on grass is very good and correctly identifies Sampras as marginally superior to Federer by the numbers. He dominates Fed in the two big stats of 1st serve and 2nd serve return. Federer makes up for this with stellar numbers in the lesser stats.

Djokovic last year was 38/(100-73) on points for a crude points DR of 1.41. He won 95% of his service games and was exceptionally crafty with his second serve getting a whopping 64% on 2nd serve points; a jaw dropping number that got him another title. Murray held his opponents to 41% on 2nd serve. Hope they play this year.

On clay I tried some weighting based on 1st serve return and 1st serve numbers, but thought it caused some oddities. It may be for grass and clay that weight with the break point numbers may encompass some of serve number differences inherently. Its way to luck influenced to use break point numbers exclusively especially when just using the past years data or current years data to predict performance.

 

[Meles]

Thanks @Meles for this. It's great to see someone supply interesting numbers and also not just add to the crowd of Djokovic doomsayers who just assume he is going to win everything. Strange things happen in tennis (particularly the FO) and it is great that you don't lose that perspective.

See the discussion of Djokovic on grass above. He's a hard nut to crack, but I'd say Wimbledon is less assured than the French Open.

 

[Meles]

What has happened is that@Meles has become the GOAT of statistical computations here.

Either that or he has Deep Throat status at the ATP and gets the proper figures, while the rest of us only see the rounded data!

I love doing formulas in excel.

Haha. If I had database access I'd have to become a lot better with queries.

Blowing the ****s away at the ATP site and on TV; that's what we do here at TTW!

 

[Gary Duane]

I love doing formulas in excel.

Haha. If I had database access I'd have to become a lot better with queries.

Blowing the ****s away at the ATP site and on TV; that's what we do here at TTW!

Seriously, the ATP site is laughable. Most of the data is actually there, but it is so clumsy. For all match stats you have to click on "go" each time or the page will not refresh. What a pain.

As for the others, not supplying one decimal point, that would take nothing but a simply change in programming.

The stat that bothers me most is % of total games because the range between out of this world and totally average is between 50% and 58%. 58% is for some bizarre year, on one surface. No one gets 50% who is any good. Most people range between 51 and 54%.

 

[Gary Duane]

An example: Both Kuerten and Verdasco are listed as winning 52% of points. One is 51.5, the other 52.4. A world of difference.

 

[Gary Duane]

When you think of big servers you thing of close sets on games, but often going the servers way. Big servers would have easier holds where they rack up the points too. So without really anaylzing the numbers, a big server might have less games dominance and still win more matches. Their points dominance might be overstated at the same time.

I read all your points. To add to the mystery, or explain it a bit:

Sampras------Grass: 92.66/21.63 114.29-----37.14/27.07 DR 1.37.20----+4
Federer-------Grass: 92.42/24.55 116.97-----38.39/28.05 DR 1.36.86----+3

I HATE doing this because it takes 4 numbers on serve and 4 numbers on return. For % of points you simply need to add the 4 numbers that were won over the 4 numbers that are all points played, but that means 8 numbers, and remove the stupid comas. All because the ATP won't provide 1 decimal point.

Then for DR, the same numbers, but manipulated differently.

In fact, Pete and Fed are closer than we thought. Using your system it woul be 92.7 to 92.4, but the actual difference is even smaller. Note that Sampras edges Federer ever so slightly in DR.

The kicker may be this:

Sampras: +1/+3

Fed: -1/+4

Remember, where the "clutch" is coming from makes a difference. Sampras is roughly +2 on serve to Fed, -1 on return. The weight is always on the serve, and that probably pushes the results up a notch.

In general:

Players play more points on return than on serve. The difference is not huge. I think Fed was 93%, which is about as low as it goes. In other words, for his career he only played 93 points on serve for every 100 points he returned. Sampras, strangely was at 97%. Novak 96%, Nadal 94%. I don't know what this means yet. The % is not linked to service dominance. Agassi was at 94%.

But no one plays as many points serving as returning.

More to the point: tennis is about sets, and in sets all sets with uneven scores have either the server or the returner play one extra game. 6/3 and 6/1 scores happen far more often when the winner serves first, and that means more service games in those sets. So for uneven sets the server has a huge advantage.

Those are some possible factors.

I don't think winning more on the 1st or 2nd serve is going to end up being the deciding factor for anything...

 

[Meles]

I read all your points. To add to the mystery, or explain it a bit:

Sampras------Grass: 92.66/21.63 114.29-----37.14/27.07 DR 1.37.20----+4
Federer-------Grass: 92.42/24.55 116.97-----38.39/28.05 DR 1.36.86----+3

I HATE doing this because it takes 4 numbers on serve and 4 numbers on return. For % of points you simply need to add the 4 numbers that were won over the 4 numbers that are all points played, but that means 8 numbers, and remove the stupid comas. All because the ATP won't provide 1 decimal point.

Then for DR, the same numbers, but manipulated differently.

In fact, Pete and Fed are closer than we thought. Using your system it woul be 92.7 to 92.4, but the actual difference is even smaller. Note that Sampras edges Federer ever so slightly in DR.

The kicker may be this:

Sampras: +1/+3

Fed: -1/+4

Remember, where the "clutch" is coming from makes a difference. Sampras is roughly +2 on serve to Fed, -1 on return. The weight is always on the serve, and that probably pushes the results up a notch.

In general:

Players play more points on return than on serve. The difference is not huge. I think Fed was 93%, which is about as low as it goes. In other words, for his career he only played 93 points on serve for every 100 points he returned. Sampras, strangely was at 97%. Novak 96%, Nadal 94%. I don't know what this means yet. The % is not linked to service dominance. Agassi was at 94%.

But no one plays as many points serving as returning.

More to the point: tennis is about sets, and in sets all sets with uneven scores have either the server or the returner play one extra game. 6/3 and 6/1 scores happen far more often when the winner serves first, and that means more service games in those sets. So for uneven sets the server has a huge advantage.

Those are some possible factors.

I don't think winning more on the 1st or 2nd serve is going to end up being the deciding factor for anything...

Surely winning players play more games on serve than return. More likely to serve out a set than to break on the final game. Which makes your stat more interesting on points. Also, somebody has to be playing more points on serve than return. It has to balance out. So the losers are playing more points on serve than return. The Djokovic vs Thiem match in Miami was insane. I believe @Chanwan noted that Djokovic played 72% more points than Thiem on serve. That was the rarest of exceptions. It makes sense that top player are troubling their opponents more so play less points on serve. Sampras was a horrible returner so its no suprise that he is close on that number. The real surprise is Djokovic, but maybe he gets himself into some Thiem like situations on occaision.

Despite what you say that winning more on first or second serve is not a deciding factor..., the approach of looking at the most successful players on a surface has validity. It may be encompassed in the break points numbers to some extent, but there may be results achieved by using a players first serve percentage and the average first serve percentage of opponents to generate numbers.

The approaches seem a bit crude, but in the final analysis the current weighted points DR on the career numbers differs by no more than .04. However, when evaluating these players it still seems important to have the most correct number. Over shorter seasons or partial seasons the number varies more, but then luck is more of a factor. I wonder at the whole approach as for career we are inevitably comparing players from different eras and who is to say whether David Ferrer is better than Juan Carlos Ferrero? They are close in these numbers and its enough that it pauses one to think. The real deal is trying to take the last few years of data or even the current season's data or perhaps historic tournament data and then trying to enlighten what might happen for these players. Ultimately luck of the draw is a huge factor, so I fear the simpler approaches have an appeal.

Its all very interesting when you look at something like grass and note how nobody who has been great on grass has relied on winning a bunch of second serve points and then just last year, team Becker/Djokovic turn the world on end and Djoko wins that 64% of second serve points and it carries him to victory. It may be as you say, that Djokovic's (a player's) net points on serve were all that mattered.

Too bad @falstaff78 hasn't spotted this.

 

[TrustTheNinja]

The numbers are quite heartening for Kei and I hope this year he makes the leap to at least SF at RG/winning Madrid or Rome. Its been said that clay suits him the best and hopefully we can see it transpire fully this year.

 

[Meles]

The numbers are quite heartening for Kei and I hope this year he makes the leap to at least SF at RG/winning Madrid or Rome. Its been said that clay suits him the best and hopefully we can see it transpire fully this year.

Murray rates above him and both have numbers from last year that match well with the greats (unfortunately Federer and Djokovic rate higher). The good think for Kei is he seems to be in great form except for the Djokovic final in Miami. Murray is mia right now, but a good chance for him to get in great form and perhaps take the French and Wimbledon. Murray actually has slighly better career stats at Wimbledon than Djokovic.

 

[Gary Duane]

Surely winning players play more games on serve than return. More likely to serve out a set than to break on the final game. Which makes your stat more interesting on points. Also, somebody has to be playing more points on serve than return. It has to balance out. So the losers are playing more points on serve than return. The Djokovic vs Thiem match in Miami was insane.

We talk about points LOST on serve and points WON on return. But of course we are merely talking about what player A makes returning vs player B.

And that's why it pretty much does not matter how points in either category are won.

I'm very interested in how Federer and Novak measure up against each other statistically on HCs, career. I think they are pretty even, but there is one huge difference. Federer has had ups and downs, but his career is not night and day, two periods. There is 2004-2007, but there are also some good years later on.

Novak, in contrast, was a much weaker player through 2010, and a totally different player from 2011 on. Again, night and day. That makes comparing them very difficult.

But here is something interesting, perhaps:

Fed is 31 and 51% on return, 1st and 2nd serves. 38% for all returns. If I'm right, that 38% is the only thing we care about.

In contrast, Novak is 33 and 56. Only 2% greater on 1st serves, but 5% greater on 2nd serves.

42% for all returns. Now, which is more important, 1st or 2nd serve returns? What if Fed were 33 and 49? Would that be important?

I don't know. But for now I will simply stick with the totals on both.

I believe @Chanwan noted that Djokovic played 72% more points than Thiem on serve. That was the rarest of exceptions. It makes sense that top player are troubling their opponents more so play less points on serve. Sampras was a horrible returner so its no suprise that he is close on that number. The real surprise is Djokovic, but maybe he gets himself into some Thiem like situations on occaision.

Fed last year at Cincinnati was weird. Perhaps not as weird, but very weird. So it can happen to anyone, for one match.

The same thing is true for games. Consider this: 7/6 7/6 0/6 0/6 7/6. I don't think BPs are counted as games by the ATP, so that would go down as 30 games for the losing player against 18 for the winner. I don't think anything that extreme has ever happened, but probably some 5 set match has been close to that bad. Weird things happen!

Here is a famous nightmare:

3-6 2-6 6-4 7-5 7-5

JMac won 26 games. Lendl won 21. RG. So stats don't work for individual matches!

Despite what you say that winning more on first or second serve is not a deciding factor..., the approach of looking at the most successful players on a surface has validity. It may be encompassed in the break points numbers to some extent, but there may be results achieved by using a players first serve percentage and the average first serve percentage of opponents to generate numbers.

We would have to look at the date very carefully. You might find two players who have the same % of all points on serve but differ on the % of 1st and 2nd serves. Then look at % of games won. But that could be very difficult to find.

We could start with Novak and Ivanisevic:

Both are listed at 86% of all service games for their careers. Let's see how the data breaks down:

Novak:
1st Serve 64%
1st Serve Points Won 74%
2nd Serve Points Won 56%
Break Points Saved 67%
Total Service Points Won 67%
Goran:
1st Serve 55%
1st Serve Points Won 82%
2nd Serve Points Won 50%
Break Points Faced 3,880
Break Points Saved 66%
Total Service Points Won 68%

To me it appears that they had very different service games. But total service points are really close, also % of BPs.

 

[kalpab]

Wow, this is crazy stuff! Take a bow, @Meles. You are surely an ATGA (all-time great analyst).

 

[Meles]

Wow, this is crazy stuff! Take a bow, @Meles. You are surely an ATGA (all-time great analyst).

Haha. No. Just putting up the numbers and ATG at that perhaps. Learning from Gary and others quickly.

 

[Meles]

We talk about points LOST on serve and points WON on return. But of course we are merely talking about what player A makes returning vs player B.

And that's why it pretty much does not matter how points in either category are won.

I'm very interested in how Federer and Novak measure up against each other statistically on HCs, career. I think they are pretty even, but there is one huge difference. Federer has had ups and downs, but his career is not night and day, two periods. There is 2004-2007, but there are also some good years later on.

Novak, in contrast, was a much weaker player through 2010, and a totally different player from 2011 on. Again, night and day. That makes comparing them very difficult.

But here is something interesting, perhaps:

Fed is 31 and 51% on return, 1st and 2nd serves. 38% for all returns. If I'm right, that 38% is the only thing we care about.

In contrast, Novak is 33 and 56. Only 2% greater on 1st serves, but 5% greater on 2nd serves.

42% for all returns. Now, which is more important, 1st or 2nd serve returns? What if Fed were 33 and 49? Would that be important?

I don't know. But for now I will simply stick with the totals on both.

Fed last year at Cincinnati was weird. Perhaps not as weird, but very weird. So it can happen to anyone, for one match.

The same thing is true for games. Consider this: 7/6 7/6 0/6 0/6 7/6. I don't think BPs are counted as games by the ATP, so that would go down as 30 games for the losing player against 18 for the winner. I don't think anything that extreme has ever happened, but probably some 5 set match has been close to that bad. Weird things happen!

Here is a famous nightmare:

3-6 2-6 6-4 7-5 7-5

JMac won 26 games. Lendl won 21. RG. So stats don't work for individual matches!

We would have to look at the date very carefully. You might find two players who have the same % of all points on serve but differ on the % of 1st and 2nd serves. Then look at % of games won. But that could be very difficult to find.

We could start with Novak and Ivanisevic:

Both are listed at 86% of all service games for their careers. Let's see how the data breaks down:

Novak:
1st Serve 64%
1st Serve Points Won 74%
2nd Serve Points Won 56%
Break Points Saved 67%
Total Service Points Won 67%
Goran:
1st Serve 55%
1st Serve Points Won 82%
2nd Serve Points Won 50%
Break Points Faced 3,880
Break Points Saved 66%
Total Service Points Won 68%

To me it appears that they had very different service games. But total service points are really close, also % of BPs.

I would contend that Goran was able to get to slam finals at Wimby because of big serve. A big serve can get very hot. Novak proved your point with his 2nd serve numbers at Wimby last year making up the difference. A similar performance no doubt this year, but a hot server can take him down. Novak can't peak in the same manner. So, first serve is a big deal, just not easy to crunch. I'll play with clay handbook a bit. Import first serve percentage for each player. Will have to swag average opponent service level. Then we can look and see if it does anything.

 

[Meles]

I would contend that Goran was able to get to slam finals at Wimby because of big serve. A big serve can get very hot. Novak proved your point with his 2nd serve numbers at Wimby last year making up the difference. A similar performance no doubt this year, but a hot server can take him down. Novak can't peak in the same manner. So, first serve is a big deal, just not easy to crunch. I'll play with clay handbook a bit. Import first serve percentage for each player. Will have to swag average opponent service level. Then we can look and see if it does anything.

Much to my chagrin first serve % is reported as a whole number, even for the annual data. Just using that in for the regular dominance formula in place of points won on serve creates up to a .01 difference in domiance 1 in every 8 times, so the error is not that bad on the whole. Oops. Just realized I did the simple dominance based on percentages and not actual points....

Wow doing simple dominance as return points won divided by service points lost is a big change versus the precise percentages. So that must stay (great one @Gary Duane !) Uh oh for 2016 data Rafa has blown this one up. Despite making semifinals in both tournaments. Rafa came out at 0.97 dominance (looks like Rafa had some long service games) vs 1.15 on the weighted dominance. Trouble in Denmark, but this over states Rafa's decline. Pure points dominance is interesting, but its better to use precise percentages for the return and serve points in the formula. The Thiem vs Djokovic match at Miami is a perfect example where raw points can get you in trouble. Djokovic played 72% more points on serve than Thiem in this match, but one. Thiem's DR looked great for the match and Djokovic looked horrible on true points. This match also shows the danger of break points where Djokovic save 14/15 break points, but he was no where near dominant in this match. Total points gives a nice picture as Djokovic one that by a fairly narrow margin. Over shorter ranges pure points dominance is not the most reliable statistic.

Now to bring in first serve points.... Haha. Stupid ATP data. I just figured out if I look at total number of first serve and second serve points I can get an exact first serve percentage and don't have to import any more tables which means easier up keep!... Hahahahaha and we have a new stat! First serve return points played percentage by the same method. Wow! Haha people served out of their mind against Murray and Fedberg (Federer and Edberg) while against Magnus Norman they served 6% lower. I'm posting this up now, but still HAVE NOT WORKED ANY SERVING DATA INTO THE DOMINANCE RATINGS.

 

[Gary Duane]

I would contend that Goran was able to get to slam finals at Wimby because of big serve. A big serve can get very hot. Novak proved your point with his 2nd serve numbers at Wimby last year making up the difference. A similar performance no doubt this year, but a hot server can take him down. Novak can't peak in the same manner. So, first serve is a big deal, just not easy to crunch. I'll play with clay handbook a bit. Import first serve percentage for each player. Will have to swag average opponent service level. Then we can look and see if it does anything.

There is no doubt that a huge serve is a tremendous weapon. We know what Isner can do when he gets hot.

As for Goran, in 1994 his stats speak for themselves, and with the stats you have been working on he actually choked horribly that year.

On grass was 94/19 113. That's a bit low. The reason is that although he won 56% of all his points (which is fantastic) and was around 1.5 DR (also fantastic) he was -18 on clutch return. No matter what the exact figures are, that is unbelievably horrible. -2 on serve is about average, nothing wrong there.

76% of service points is VERY high, right? So his great 1st serve is automatically accounted for in that number.

In comparison Sampras that year was 93/25 118. DR was 1.46 (a bit lower) but +4/+7 on serve and return. And Sampras also won 56% of his points.

The year Goran won Wimbledon was an aberration. He was 92/10 102, 51% of points on grass. DR 1.11, -11/+5, so down a net of -6 on BPs. If you look at his scores at Wimbledon in 2001 it was like God was on his side. He squeaked through match after match. I would say it was more about fate than serving, as if some force was making it up to him for all the times he was actually playing better but could not finish!

 

[Gary Duane]

Much to my chagrin first serve % is reported as a whole number, even for the annual data. Just using that in for the regular dominance formula in place of points won on serve creates up to a .01 difference in domiance 1 in every 8 times, so the error is not that bad on the whole. Oops. Just realized I did the simple dominance based on percentages and not actual points....

That's what I have been cursing about!

The ATP is awful that way. If you want the correct figure you have to go to all match stats, then divide the two figures they show. Again, when it says Sampras won 81% for his career on 1st serve, that can be 80.5 or 81.4999. It can be so close to a full % point off that it might as well be a full % point. And that REALLY sucks.

It can cause this:

36.5/24.5=1.489795918

37.49/25.49=1.470772852

In both cases the ATP will show 37% and 25%. But the first will be rounded to 1.49, and the second to 1.47. Now the error in DR will be .02.

The same thing happens in games. When you total, each figure can be off a full %. When you add, the error can be +/- 2 because you are doubling the percentage.

The worst is % of all points won, because that error can be a .999 of 1 %, and every % on that stat is HUGE.

That's why I have mostly stopped doing the calculations.

For % of all points I don't want to have to deal with 8 different numbers because the ATP is too stupid to show a decimal point. One decimal point is fine. Two is probably overkill. But no decimal point is horrible.

 

[Meles]

There is no doubt that a huge serve is a tremendous weapon. We know what Isner can do when he gets hot.

As for Goran, in 1994 his stats speak for themselves, and with the stats you have been working on he actually choked horribly that year.

On grass was 94/19 113. That's a bit low. The reason is that although he won 56% of all his points (which is fantastic) and was around 1.5 DR (also fantastic) he was -18 on clutch return. No matter what the exact figures are, that is unbelievably horrible. -2 on serve is about average, nothing wrong there.

76% of service points is VERY high, right? So his great 1st serve is automatically accounted for in that number.

In comparison Sampras that year was 93/25 118. DR was 1.46 (a bit lower) but +4/+7 on serve and return. And Sampras also won 56% of his points.

The year Goran won Wimbledon was an aberration. He was 92/10 102, 51% of points on grass. DR 1.11, -11/+5, so down a net of -6 on BPs. If you look at his scores at Wimbledon in 2001 it was like God was on his side. He squeaked through match after match. I would say it was more about fate than serving, as if some force was making it up to him for all the times he was actually playing better but could not finish!

Done I hope. The new serve based dominance rating is in place. It weights heavily for first serve and a bit for first serve return. The complex equations matches the typical points DR that is shown that uses the detailed percentages. By breaking out the first serve percentage and the new hilarious first return percentage and multiplying that by the appropriate win percentages it comes out exactly the same as the simple perecentage points equation. With this more complex equation it is easy to weight for first serve and first serve return (and different weightings can be done on different surfaces in the future.)

The first return percentage is a really hilarious stat when you see that Murray, Edberg, Federer, and Fernando Gonzalez dominating this stat. I don't know Fernando Gonzalez well, but perhaps opponents feared his might forehand on 2nd serves and got their first serves in more. I can see Murray's tirades filling the opponents with confidence on serve. Perhaps Federer and Edberg were too genteel on court. Not sure why Gaudio and Norman lead this stat. Gaudio's first serve return was pretty good.

I do like the bolded 1st Serve and 1st Return Weight Dominance Ratio. It does not use break point or clutch data. Some players do very well like Robredo and Ferrer, but they also played heavily on clay and faced lesser fields on average (same for Muster). So, you really have to have in your mind whether the player was a clay court specialist as that generally helps their numbers. It does a nice job of boosting Wawrinka up last year while Tsonga is penalized some for his poor first serve returning. It seems to capture the escence of what it takes to win on clay.

 

[Gary Duane]

Rafa came out at 0.97 dominance (looks like Rafa had some long service games) vs 1.15 on the weighted dominance. Trouble in Denmark, but this over states Rafa's decline. Pure points dominance is interesting, but its better to use precise percentages for the return and serve points in the formula.

Something sounds wrong here.

The ATP shows 40% return points, 68% service points. So that's 40/38. The answer has to be over 1. Even the worse case scenario, 39.5/38.49 is still over 1. DR can't be less than around 1.03. When you weight it you will be working with around +2/+10 or a net of +12, so obviously that will improve the ratio.