Been playing around with the new MatchFacts calculator on the ATP website. You can now compare stats by career, year, all matches and surface. There are some interesting revelations. Sampras is 5th all-time for points won on first service and 17th for points won on second serve, bringing some reality to the legend surrounding his delivery. Unquestionably a great serve, but not quite unrivaled as some would suggest. Remember he played on quicker courts back then too. Ljubicic averages more aces per match at 11, than Sampras did at 10. Nadal comes out on top for points won on 2nd serve at 58%, Federer follows closely behind at 57%. Interesting to see how the stats change so dramatically when you compare them across different surfaces. Coria comes out on top for return games won across all surfaces, but Chang is best on hardcourts, Nadal on clay and Kucera on grass. It's fun to play around with. Interesting to see how success in tennis can be achieved with a careful balance of strengths and weaknesses, as opposed to being at or near the top in each category. http://www.atpworldtour.com/Matchfacts/Matchfacts-Landing.aspx

Thanks for posting the link. Cool indeed. 1st serve points won... (over career) #1 - Ivo Karlovic #2 - Goran Ivanisevic #3 - Richard Krajicek #4 - Wayne Arthurs #5 - Pete Sampras ... #18 - Federer (just behind Sam Querry, Mardy Fish and J-W Tsonga) #48 - Murray (just behind Tommy Haas, Marin Cilic) #88 - Djokovic (just behind Scott Draper, James Blake) #106 - Nadal (just behind Michael Chang, Petr Korda, Syan Wawrinka) 1st serve points won (in 2011) 1 - Milos Raonic 79% 2 - Roger Federer 79% 3 - Jo-Wilfried Tsonga 78% 4 - Tomas Berdych 78% 5 - Feliciano Lopez 78* ... 19 - Murray 75% (just behind Tipsarevic, Petzschner, Soderling) 22 - Djokovic 74% (just behind Malisse, Gulbis) 29 - Nadal 71% (just behind Kohlschreiber, Monfils, Belluci)

My mistake, from 1991 then. Still pretty good. I wonder if they have the data to go back further. Hopefully they will be able to add more years later.

They will not.. they run this from 1991 and don't have data for years before. (and still there is lot of mistakes)

We'll see, because the rankings now go back to 1973, but when the ATP originally updated to this website they did not go back that far.

ATP can't manage to tell us surfaces of some tournaments from 70's, do you really think they can provide match statistics (no one can) They started this in 1991. and that's showed in match facts

There are serious problems with the ATP stats for matches from the 1990s. The most extensive problem is explained here: http://tt.tennis-warehouse.com/showthread.php?p=2812426#post2812426 Aces and double-faults are counted twice, and that distorts the Total Points Won, but not just that. The Points Won on 1st Serve are also distorted, because most aces are first serves and they're getting counted twice in that category. The figures for Points Won on 2nd serve are likewise distorted, because all the double-faults are getting counted twice in that category. So one thing that happens is that a player like Ivanisevic who averages a lot of aces per match will get an edge in Points Won on 1st serve because his aces are being counted twice. Similarly someone who double-faults a lot will end up too low on Points Won on 2nd serve. In my survey of the ATP stats I have not found this problem in any matches from 1991, or in any matches from 2000 onward. The problem seems restricted to the period 1992-99, as far as I know.

Another problem with the ATP stats in the '90s is that they don't include all the service games in a match, if the match goes beyond 6-all in the fifth set. For example, Norman d. Ivanisevic in 1997 at Wimbledon, 6-3, 2-6, 7-6, 4-6, 14-12. Not including the tiebreaks, that comes to 65 service games. The ATP boxscore reports only 39 service games: http://www.atpworldtour.com/Share/Match-Facts-Pop-Up.aspx?t=540&y=1997&r=2&p=N250 It's as if the 26 games of the fifth set got dropped altogether. That problem is documented in the thread I linked to, here: http://tt.tennis-warehouse.com/showthread.php?p=3200690#post3200690

You can see that problem in the MatchFacts feature. It gives Ivanisevic 88 service games on grass in 1997: http://www.atpworldtour.com/Matchfacts/Matchfacts-List.aspx?c=4&s=2&y=1997 And 88 is the number you get if you add up his service games in the individual matches he played on grass that year. He only played two events on grass, Queens Club and Wimbledon. Look up the stats for each of those matches, and his service games add up to 88: http://www.atpworldtour.com/Finals/.../G/Goran-Ivanisevic.aspx?t=pa&y=1997&m=s&e=0# But if you just go by the scores of the matches, he must have served at least 99 games (and at most 105). The problem is resolved if you assume that the 26 games of his marathon 5th set with Norman are not being counted. Give him back his 13 service games from that set, add them to the 88 service games that the ATP gives him, and he comes out with 101 service games on grass in '97. Exactly as required by the scores.

Another problem you run into is calculating averages like this. The ATP lists the number of matches someone played -- but they do not have stats for all those matches. For example Sampras is at #1 for hitting aces in 1993: http://www.atpworldtour.com/Matchfacts/Matchfacts-List.aspx?c=0&s=0&y=1993 According to the ATP he hit 1,011 aces in 101 matches, but that is incorrect. They do not have Pete's ace count in 7 of his matches: 3 at the Grand Slam Cup and 4 in World Team Cup. If you try to pull up the Stats for those 7 matches, you get only empty boxscores: http://www.atpworldtour.com/Tennis/Players/Sa/P/Pete-Sampras.aspx?t=pa&y=1993&m=s&e=0# The same is true for all Davis Cup matches. As far as I know, the ATP does not have stats for them. In this example from 1993, that leaves the ATP counting Pete's aces only in 94 of his matches. (I used Chrome to count the matches on his 1993 page, did a search for the word "Stats"). So the ATP has him hitting 1,011 aces in 101 matches, and that would come out to 10.0 aces per match, but it's actually only 94 matches. His real average per match was 10.8. Not a small difference.

Yeah, for Isner this year they count 2 DC matches but not stats from them so we get wrong aces per match ratio (against Capdeville he fired 38 aces and they are not included)

He's got 62 matches on his page for last year, including two Davis Cup matches that the ATP does not have stats for: http://www.atpworldtour.com/Tennis/Players/Top-Players/John-Isner.aspx?t=pa&y=2010&m=s&e=0# And 62 is the number of matches he gets on the list of aces: http://www.atpworldtour.com/Matchfacts/Matchfacts-List.aspx?c=0&s=0&y=2010 1995 is way off for Sampras. He's got 88 matches listed on his page: http://www.atpworldtour.com/Tennis/Players/Sa/P/Pete-Sampras.aspx?t=pa&y=1995&m=s&e=0# And 88 is the number he gets on the list of aces: http://www.atpworldtour.com/Matchfacts/Matchfacts-List.aspx?c=0&s=0&y=1995. But 6 of those matches were Davis Cup, another was in Grand Slam Cup. So they really only have aces for 81 matches. So on the Matchfacts page it looks as if he averaged 11.1 aces in his ATP matches. But his real average was 12.0.

Data missing for entire rounds at GS events The ATP has no data for the first and third rounds at the French in 1991 (click on the scores to get the stats): http://www.atpworldtour.com/Share/Event-Draws.aspx?e=520&y=1991.

Data missing for other Grand Slam rounds The ATP has no data for the R32 at Wimbledon in 1996, except for one match (Krajicek-Steven): http://www.atpworldtour.com/Share/Event-Draws.aspx?e=540&y=1996. They have no data for the R32 at the AO in 1997: http://www.atpworldtour.com/Share/Event-Draws.aspx?e=580&y=1997. They also have no data for the R32 at the 1994 Paris Masters: http://www.atpworldtour.com/Share/Event-Draws.aspx?e=352&y=1994. (Data is otherwise available for that tournament in 1991-99).