Stats for AO semis (Federer-Roddick '07, Edberg-Wilander '90)

[krosero]

Federer d. Roddick 6-4, 6-0, 6-2
Edberg d. Wilander 6-1, 6-1, 6-2

These matches have the highest and the third-highest rates of winners per game that I have counted myself or read about, so far.

Let's start with Fed-Roddick.

This is the first time, since I started this project, that there have been so few discrepancies in stats. There is no discrepancy between the published stats (e.g., the ATP stats and those published by Reuters at http://uk.reuters.com/article/UK_TENNIS/idUKL2501342520070125); and there is only a minor discrepancy with my own counts of aces, service winners, doubles and other winners when compared the stats displayed throughout the match by ESPN2 (there is one missing point in the coverage).

In short, all but one service or non-service winner is accounted for. True, the announcers (Enberg and Carillo) did not keep a running count of aces and service winners; but they provided a running count of total winners, and it was easy to look at my own sheet and see what the ESPN statisticians were adding up, until the missing point in the set 3, game 5 (the opening point, won by Federer).

I'm hoping that my counts line up this well with available stats for recent matches generally, though if you've been following my posts you know that so far this is the only match I've done more recent than 1996.

Here are some published stats that I did not count myself:

Federer served at 50%, Roddick at 63%.

There were 128 points. Federer won 83 of them (or 65%).


And these I counted myself:

Federer had 10 aces, 2 service winners, and 2 doubles.
Roddick had 4 aces, 2 service winners, and 2 doubles.

(Note that one of Roddick's service winners was shown to be out, by Shot Spot).

Federer made 32 winners: 15 FH, 14 BH, 2 FHV, 1 BHV, 0 smashes.

Roddick made 5 winners: 2 FH, 0 BH, 0 FHV, 2 BHV, 1 smash.


Federer's winners by set: 9, 10, 13
Roddick's winners by set: 2, 1, 2


Federer's ground stroke winners are impressive, but especially for such a brief match.

On the other hand, between them Federer and Roddick had just 5 volley winners, and none at all in the second set.

Andy had not a single backhand winner, and though that looks terrible, he also had just 2 forehand winners.

There were only two service return winners, both forehand returns by Federer of Roddick’s first serve, both in the first set. The first was a baseline-to-baseline winner; the second was a pass.

Federer passed Roddick 15 other times, with 5 FH's and 10 BH's. None were lobs.

Roddick never passed Federer.

(Per Reuters, Federer won 10 of 11 approaches. Roddick won just 9 of 31, or 29%).

On his serve at love-2 in the second, down a set and a break, Roddick started coming in like gangbusters, approaching on four straight points and losing them all. (That’s the game where he smashes a forehand from near the net post, back toward the middle of the court, and got a backhand pass from Roger). He then lost two more approaches in the next game. In all but one of these approaches, he got to hit a ground stroke or volley at the net but could not put away the point.

I often felt that Roddick was coming in on approaches that bounced too high: hard-struck approaches that might have worked elsewhere but that looked suicidal against Federer.

And most of all I felt that Roddick needed to trade some power for better placement on his serve.

EDBERG-WILANDER

I have only read about this one and would gladly buy the match from anyone who has it.

It lasted 1 hour 22 minutes.

Edberg served at 80% and “put away 39 volley winners to Wilander’s one” (per the St. Louis Dispatch, LA Times, etc.)

To put this in perspective, Federer's rate of non-service winners per game against Roddick is 1.33 -- the same, for example, as Agassi's rate against Wilander at the 1988 FO.

Edberg, with 39 winners just from volleys, already has 1.77. That the highest rate of non-service winners per game that I know about, even before his ground strokes are counted.

(Edberg's match had the same score at the 1984 Wimbledon final, in which McEnroe hit 25 non-service winners, for a rate of 1.14).

I would very much like to confirm Edberg's winner count.

If you added aces, then Federer's rate of winners per game goes up to 1.75. Among the matches with stats counted by myself, MooseMalloy and Urban, and other matches with published stats for aces and non-service winners, the only higher rates I know about belong to Edberg-Wilander and the Sampras-Wilander USO match of 1989. The New York Times gave Sampras 12 aces and 72 "placement winners". If that term means only non-service winners (and his aces were all clean aces) then his rate there was 1.87.

That's another match I'm looking for.

In some of his USO matches later against Agassi (1990, 1995, 2001), Sampras has very high rates, going by published figures, but not higher than Fed's 1.75. And in any case those figures still need service winners subtracted from the total. For that I would still need to watch the matches and do my own stats. Also I would need to note the any discrepancies between my stats and published figures.

 

[bluetrain4]

This is the AO where Edberg was playing SO great. He very well could have 7 slams instead of 6, but he had to retire in the third set during the final vs. Lendl after pulling an abdominal muscle. He had led a set and a break when he pulled the muscle. Lendl won the second set 7-6 and led in the third. Not saying Lendl could not have came back anyway, but Edberg really had the upper hand when the injury occurred.

 

[krosero]

This is the AO where Edberg was playing SO great. He very well could have 7 slams instead of 6, but he had to retire in the third set during the final vs. Lendl after pulling an abdominal muscle. He had led a set and a break when he pulled the muscle. Lendl won the second set 7-6 and led in the third. Not saying Lendl could not have came back anyway, but Edberg really had the upper hand when the injury occurred.

sometimes you see a great performance in the semis, when the winner is called unbeatable, and then in the final there's an upset (eg, Ashe-Connors in 1975). But generally when someone like Edberg is in the zone, playing in his best year (which I think 1990 was), you'd expect him to win.

 

[andreh]

This is the AO where Edberg was playing SO great. He very well could have 7 slams instead of 6, but he had to retire in the third set during the final vs. Lendl after pulling an abdominal muscle. He had led a set and a break when he pulled the muscle. Lendl won the second set 7-6 and led in the third. Not saying Lendl could not have came back anyway, but Edberg really had the upper hand when the injury occurred.

Actually, Edberg pulled his abdominal muscle already in the SF against Wilander and was injured from the start in the final against Lendl. He started getting treatment a few games into the first set and had to play from the baseline for much of the 2nd set and on.

 

[lambielspins]

Those stats say it all. Federer is far superior to Roddick in every fact of the game.

Edberg's performance vs Wilander was extremely impressive though, even if Wilander was on the down already. I dont think there is anyway Lendl would have won that final if Edberg had been healthy, I respect Lendl alot, Edberg was just too on fire that tournament.

 

[krosero]

2007 USO quarterfinal

Federer d. Roddick, 7-6 (5), 7-6 (4), 6-2

From the USO website:

Federer served at 61%, with 15 aces, 48 winners (including service), 18 unforced errors, winning 14 of 22 approaches, and 110 points overall.

Roddick served at 71%, with 14 aces, 42 winners (including service), 24 unforced errors, winning 18 of 36 approaches, and 94 points overall.


Roddick served at 81% in the second set.

There were no breaks until Roddick was broken at 2-3 in the third set. In fact there had not been any break points before 3-4 in the second, when Andy earned his first but failed to put Roger's second serve back in play.

Roddick had exactly a 50% success rate coming in against Federer, a great improvement over the match in Australia but still almost a losing strategy.

Roddick came close to matching Federer in winners if aces and service winners are included, but in non-service winners (my own count) Federer was solidly ahead:


Federer had 27 winners: 13 FH, 11 BH, 2 FHV, 0 BHV, 1 overhead.

Roddick had 16 winners: 6 FH, 2 BH, 4 FHV, 2 BHV, 2 overheads.


Federer hit 7 passing shots (and 1 lob winner), while Roddick hit only 1.

But Roddick had just as many winners at net as in the backcourt.

 

[Moose Malloy]

Going back to the '90 AO, I think its a bit of a stretch to say that Edberg was a lock to beat Lendl if he wasn't injured.

There was still a huge gap between the players at that time, esp in terms of ranking points(big lead for Lendl) & hardcourt slams(Lendl had 4 at that point & Edberg had 0)

And Lendl was still up 9-6 in the head to head at the time of the match.

And I wouldn't read too much into Edberg's win over Wilander(which I admit I haven't seen), Mats was pretty bad in '89 & was a joke the rest of '90, he was a shadow of his former self at that point(I know he beat Becker that event, but Becker was pretty horrible at the AO up to that point in his career, I think he was capable of losing to anyone there)

And we shouldn't forget that Lendl & Edberg played at the '91 AO, when Edberg was finally ranked #1(& Lendl was #3 & finally showing signs of decline/age) Yet Lendl won in 5 sets, saving match points.

sometimes you see a great performance in the semis, when the winner is called unbeatable, and then in the final there's an upset (eg, Ashe-Connors in 1975). But generally when someone like Edberg is in the zone, playing in his best year (which I think 1990 was), you'd expect him to win.

I think '91 was a better year for Edberg than '90. And I've seen him in the zone many times, then falter('91 W comes to mind, he played some of the best grasscourt tennis I've seen in the 1st 5 rounds, then played 3 horrible tiebreaks vs Stich. That's a match I'd like to do stats on, from what I recall Edberg really outplayed him in all categories, he had no business losing that match)

He also played some of the best tennis of his career in summer of '90, yet lost in the 1st round of the USO. And then there was his horrible breakpoint conversion rate vs Chang at the '89 FO. I think there was still a lot of doubt in Edberg's mind in big matches at the time of the '90 AO(having lost 2 major finals in '89), & Lendl was more than capable of taking advantage of that.

 

[urban]

Yes, Moose you are right. Edbergs play in the first 5 rounds at Wim 1991 was one of the best performances i have seen. It was maybe the best tournament, he ever played. Same goes for Connors in 1975, it was the best tournament, i have ever seen him play.

 

[krosero]

Those are good points about Edberg, though I would not have said that he was a lock. I still would have expected him to win, not in straight sets, but simply as a favorite. He had beaten Lendl a month earlier at the Masters, in two close sets.

Much would depend on seeing these 1990 matches (including the Becker match; I don't disagree that Becker was in poor form but Wilander was thought to have played a great match, maybe by Drysdale??).

Also haven't seen the 1991 match you mentioned; I always wondered if Edberg choked on the match points.

Stich match: sorry to say I also missed that one, but yeah, Edberg held his serve throughout. Just on paper, he played the big points badly.

 

[Moose Malloy]

Those are good points about Edberg, though I would not have said that he was a lock.

That wasn't really directed to you, its just that I've heard that from others a lot over the years here & just wanted to give my 2 cents.

I still would have expected him to win, not in straight sets, but simply as a favorite.

I wonder if he would have been favored with the oddsmakers for that match though. A #3 vs a #1, who was the defending champ? And Lendl was pretty flawless in his semi as well. all just on the basis of the Wilander match? Maybe it would have been a pick 'em.

Thought of another match that many picked Edberg to win: '87 W SF vs Lendl.

Also haven't seen the 1991 match you mentioned; I always wondered if Edberg choked on the match points.

I have the dvd, will check it out. Maybe I'll do stats. I think some of the Edberg-Lendl matches were among the most entertaining matches I've seen.

He had beaten Lendl a month earlier at the Masters, in two close sets.

I'm doing stats on that one right now. In the 2nd set, espn flashes 'pts at net' stats for the 1st set. I'm curious if they match up with mine. Lendl had a pretty big edge in winners, despite losing the 1st set.

Apparently it was considered a huge upset at the time. Lendl had been in 8 straight Masters finals. Press was already asking Becker about playing Lendl in the final before Edberg & Lendl had taken the court.
Edberg sounded pretty surprised that he won as well. Less than 24 hours later he beat Becker. And he had a cold.

 

[krosero]

I wonder if he would have been favored with the oddsmakers for that match though. A #3 vs a #1, who was the defending champ?

I asked myself about the oddsmakers after my post, and maybe they would have picked Lendl, for his reliability and his prior record against Edberg overall; and his prior record at the AO (def. champ). Basically though, my pick at the time was Edberg, and not especially on the basis of the Wilander match. Yes, Edberg played a great match in the semis (because even if Wilander played poorly, Edberg still served at 80%), but I'd seen him fall before under pressure. What I was going on was that Edberg had just beaten Lendl at the Masters, on arguably Lendl's favorite surface. He had, by then, the game to beat Lendl -- if he was not injured and played his best.

I'm doing stats on that one right now. In the 2nd set, espn flashes 'pts at net' stats for the 1st set. I'm curious if they match up with mine. Lendl had a pretty big edge in winners, despite losing the 1st set.

Nice surprise that you're doing this one. Right now, before I get to more matches, I'm taking net stats in the Roddick matches and comparing against available figures. I think I know how they're counting in the AO match but I need to do the USO match too.

 

[krosero]

In the Mecir thread you gave me some advice on how to tally net stats. I did what you suggested(SVW, SVL, APW, APL)

But I'm not counting SV points where the server got unreturned serves(which I think you disagree with) since guys like Becker & Sampras would get too much credit for SV when they really had no intention of the V part. But even in the Mac-Lendl W match I just did, Mac was really going for aces on his 1st serve a lot, so maybe I did it the right way. But if a player gets passed on the return I count it in the SVL column. Maybe that seems contradictory, but I don't think including it has affected the final stats since clean passes don't seem that common.

As far as approaches, a player needs to make the approach for it to count in my stats, so I'm not counting errors at all. But if a player makes the approach & his opponent misses the pass I count it in the APW column.

So far its been pretty straightforward in the grasscourt matches I did so far. More questions come up in the claycourt matches, but the total numbers are pretty low anyway, so I doubt it would change things that much if I did it differently.

I still haven't got a chance to compare with available net stats I have on some matches to see if I'm doing it right(or at least am on the right track)
have you done net stats of your own on many matches? does it differ from the way I'm doing it? You said it was straightforward in the Mecir-Wilander match.

Partly I'm doing it to compare how agressive some players were(like Mac vs Edberg), & already I'm noticing a difference on surfaces. Players are getting more returns in play on AO grasss vs Wimbledon grass, more on indoor/hardcourt than grass, etc.

That last point is interesting, considering all the debates here about the differences between types of grass.

One reason I’ve come to value net stats is that the more I count winners, the more I suspect that net-rushing produces greater numbers of winners: more volleys, more passing shots.

It would need to be taken into account, because if you have one victor with 50 winners, and in another match the victor has 20, you can’t just say that the first match was more “powerful.” Maybe in the first match the victor faced a net-rusher, while in the second he grounded it out powerfully with a baseliner.

A good example: the two 1988 USO semifinals (Wilander-Cahill had more winners than Lendl-Agassi).

Hopefully no one reads these counts of winners as merely an index of power. The counts do reflect power but I think they also reflect net-rushing.

On the other hand, I think we both found the winner counts in the two Hanover matches of 1996 (Sampras-Becker) to be low. There we had two net-rushers. So there’s still more to think about, as the project progresses.

When I started counting winners, I wanted also to count net stats, merely because I was interested in them. I had not yet thought that the two stats would be linked.

We haven’t got net stats for every match, and often I think it’s enough simply to count the passing shot winners; then you get a sense what the net-rushing did to the total winner count.

And I think we’ve got a lot of net stats already, even if our definition of net-rushing is still unclear.

In the two Roddick matches, I compared my own counts, with available figures.


AO (per Reuters):
Federer – 10 of 11 “net approaches” won
Roddick – 9 of 31

AO (my count):
Federer – 10 of 11
Roddick – 9 of 30



USO (official website):
Federer – 14 of 22 “net approaches” won
Roddick – 18 of 36

USO (my count):
Federer – 12 of 18
Roddick – 17 of 35


So I came very close to the “official” figures – a little closer in the brief match, which you would expect.

Where I’m off, I’ve under-counted, not over-counted.

However let me say something about the TV stats. At the USO, CBS provided figures after 12 games. They had Roddick at 6 of 8. Up to that point, though (and I went through it twice), I saw Roddick getting passed cleanly at the net three times. The most recent pass had been two games before the figures went up, so there shouldn’t be a question of the figures not being updated quickly. Simply, 6 of 8 is wrong.

Now, whether the match is reviewed for the print media and official websites to publish proofed stats, I have no idea.

CBS put Federer at 4 of 7, by the way. I had him at 3 of 6. I didn’t find the missing approach that he won. This concerns me more, because it suggests I just missed something, or that they’re counting in some unknown way. I did see Federer make two FH putaways where he moved forward but did not get past the service line.

This is the rule I’m using: if someone gets to the service line while the ball is in play, a net point has been played.

I found that this eliminates errors on the approach, which you and I actually agreed about. But now I’m not thinking in terms of the shot; I’m thinking in terms of the player, and where he is while the ball is in play.

I find that this would eliminate any complicating questions like what type of approach is hit, whether the other player has seen them coming and feels the pressure, whether the player is rushing forward intentionally or trying to get a drop shot, whether they intend to put a ball away without volleying, etc. It basically comes down to where their feet are: do they get beyond playing from the baseline and from no-man’s land, while the ball is still in play?

But you brought up a good point about service. There are tons of points where the serve forces an error. On some points, if you simply went by my rule literally, the receiver shanks the ball up almost vertically, and while we wait for the ball to come back down (while we wait for the ball to not be in play anymore), the server could have sauntered up to the net, back to the baseline, and had a chat or two with the linesmen.

So we can’t eliminate human judgment altogether. We have to allow that service – and some ground strokes – can force errors that have nothing to do with whether the opponent sees the player coming.

I suggested that we not count aces and service winners as net points, even if the server is charging forward. The purpose is to identify certain points where the serve would not have been returnable no matter whether the server is coming in.

You’ve made it simpler by not counting any unreturned serve as a net point. That may work. I haven’t counted net stats for any of the big-serving attacking players; I’ve only done it for baseliners, so I don’t know that I could tell you anything useful here. My suggestion is to get some available stats for such players, compare against your own, mark down everything that might be counted, and see what gets you closest to the available stats.

Speaking for myself, I think so long as we get a good general sense of what should be counted, minor discrepancies in judgment calls is no big deal; personally I'm only looking for approximate numbers of net-rushes in any given match, to get an idea of how that affected the winners.

 

[krosero]

Stats for the Edberg-Wilander match are here: http://tt.tennis-warehouse.com/showthread.php?t=179739