Incorrect stats
In all these, the ATP appears to have counted aces and df's twice.
 For IvanisevicEdberg, 1992 Wimbledon quarterfinal, Moose Malloy counted 316 points in the match; the ATP has 315 if aces and df’s are subtracted once from their figures for Total Points Won.
 For AgassiRafter, 1995 AO, R16, I have Agassi leading in total points by 8253, and so does the ATP if aces and df’s are subtracted once.
 For PhilippoussisSampras, 1996 Australian Open, R32, I have Philippoussis leading in total points by 119109, and so does the ATP if they don’t count their aces and df’s twice.
 For KafelnikovSampras, 1996 Roland Garros SF, I have Kafelnikov leading in total points by 10571, and so does the ATP if they don’t count their aces and df’s twice – but note that they have two fewer Kafelnikov aces and two fewer Sampras doublefaults than I do.
 For WashingtonRadulescu, 1996 Wimbledon quarterfinal, the Boston Globe has Washington leading in total points by 185182, and so does the ATP if aces and df’s are subtracted once.
 For RafterRusedski, 1997 USO final, a boxscore in the Hartford Courant has Rafter leading in total points by 127106; the ATP has the same if their aces and df’s are subtracted once.
Rafter Rusedski
Firstserve pct. 62 60
Winners 49 31
Aces 7 9
Unforced errors 23 26
Double faults 6 4
Breakpoint conv. 617 39
Pct. 1stserve pts 82 63
Net approaches 6895 54100
Pct. 2ndserve pts 46 51
Total points won 127 106
 For PhilippoussisMoya, 1998 U.S. Open semifinal, the New York Times has Philippoussis winning 63 of 74 points on first serve. The ATP has 83 of 94, which looks as if they counted nearly all of Mark’s aces twice (the Times and the ATP agree that he had 21 aces).
 For RafterSampras, 1998 USO semifinal, I have Rafter leading in overall points by 152138. The ATP has Rafter leading by 153135, if the aces and df’s are subtracted once.
