Greatest Serves of All Time

[pc1]

Originally Posted by Datacipher

I was going to point this out. I'm not sure the gross measure of serve games won is much of a measure of return at this level, but, when you compound that even further by correlating serve ability with ranking....I think the validity flies right out the window.

And that's another thing I'm worried about. I don't want to be rigid so I'll give in to some points. It seems to me that point of the article was whether Agassi is the default choice for GOAT of service return and whether there is debate on his status as the best returner in his time. I think we'll muddling in too many different scenarios and examining whether it's top ten, top twenty. We now should go whether the serves of Chang's opponents are superior to Agassi's opponent. It's getting too complicated for a simple point and it's very simple. Can anyone say with 100% or 90% or even 5% confidence that Agassi is the best serve returner of his time? I think it's debatable.

It's not quite like Sampras who led the ATP in percentage of holding serve in most of the ten years and even when he wasn't leading he was very close. Sampras is legit in his argument for greatest server.

Agassi led the ATP only a few times in percentage of breaking serve during the ten years examined and some years he wasn't in the top ten. The GOAT of returner should be better than that.

I guess we just have to disagree. I don't want to keep beating a dead drum. My point is Agassi doesn't have enough when we examine his record to be call the return GOAT. We don't have all the information for the amount of games played. It's just not substantial enough information for me.

I'm just going to stop on this topic. I understand everyone's reasoning and it is all logical and does make sense. Below is the information again. If the returners are in the top ten, it is listed. If they are out of the top ten, the ranking is not shown.

1991
Agassi 29% service games won-56 matches
Chang 36% 67 matches-3rd in ATP
1992
Agassi 34% 57 matches-6th in ATP
Chang 37% 80 matches-1st in ATP
1993
Agassi 37% 44 matches-1st in ATP
Chang 33% 87 matches-5th in ATP
1994
Agassi 34% 66 matches-5th in ATP
Chang 33% 87 matches-6th in ATP
1995
Agassi 36% 82 matches-2nd in ATP
Chang 32% 84 matches- 8th in ATP
1996
Agassi 31% 52 matches-8th in ATP
Chang 35% 84 matches-1st in ATP
1997
Agassi 28% 24 matches
Chang 33% 68 matches-2nd in ATP
1998
Agassi 32% 86 matches 3rd in ATP
Chang 30% 52 matches
1999
Agassi 33% 77 matches-1st in ATP
Chang 26% 52 matches
2000
Agassi 28% 55 matches
Chang 29% 68 matches-5th in ATP

[krosero]

Originally Posted by pc1

It's not quite like Sampras who led the ATP in percentage of holding serve in most of the ten years and even when he wasn't leading he was very close. Sampras is legit in his argument for greatest server.

thing is with Sampras, he had to have a great total game, backing up his serve, in order to be leading the tour in service holds. A complete game, plus mental consistency, year in and year out (in that respect he resembles Chang, not Agassi). So I think that's what Pete's hold statistics reflect: his ability to hold serve. That's the argument for Pete: that he had the best total serve-and-volley game (or serve-and-stay-back game, when he chose to do that) of his time. Theoretically, with his hold numbers, you would think that part of his total hold game included a serve that was, at least, one of the best of its time. And everyone agrees Pete's serve was at least one of the best of its time.

In the same way I think that everyone agrees that Agassi had, at the least, one of the best returns of his time. But the best return game of his time? I don't know if Agassi has ever had that argument made for him. Maybe he has. But that's not what I think people are referring to when they say he had the best return.

I always assumed McEnroe's famous statement about Agassi having a better return than Connors referred to the return as a single stroke. Now, maybe I read that into his statement. It's possible my assumption was wrong. But that's how I read it.

Anyway that's what I think the argument for Agassi has always been -- the best return of his time, considered as a single stroke.

[krosero]

Originally Posted by Datacipher

I was going to point this out. I'm not sure the gross measure of serve games won is much of a measure of return at this level, but, when you compound that even further by correlating serve ability with ranking....I think the validity flies right out the window.

I'm not really clear what you mean. So let me make this point generally.

If it's a bad idea to correlate serve ability with ranking -- if it's not justified to assume that higher ranked players have better serving ability -- then it is equally unwarranted to assume that higher ranked players don't have better serving ability.

Just speaking generally, I would think the latter assumption carries a higher burden of proof.

So it may be possible that Agassi, while facing top players more frequently than Chang, did not face better serving ability. But is it likely? I would say no.

[abmk]

Originally Posted by pc1

And that's another thing I'm worried about. I don't want to be rigid so I'll give in to some points. It seems to me that point of the article was whether Agassi is the default choice for GOAT of service return and whether there is debate on his status as the best returner in his time. I think we'll muddling in too many different scenarios and examining whether it's top ten, top twenty. We now should go whether the serves of Chang's opponents are superior to Agassi's opponent. It's getting too complicated for a simple point and it's very simple.

now you are mixing 2 things up.

a) whether agassi was the best returner of his time
b) whether agassi was the best returner of all time

you ( and the author ) are trying to dispute b by disputing a , which I don't agree with . I think the article takes a wrong approach .... with insufficient data .

Originally Posted by pc1

Can anyone say with 100% or 90% or even 5% confidence that Agassi is the best serve returner of his time? I think it's debatable.

yes, he WAS the best returner of his time and most would say that with confidence !

Originally Posted by pc1

Agassi led the ATP only a few times in percentage of breaking serve during the ten years examined and some years he wasn't in the top ten. The GOAT of returner should be better than that.

he wasn't in the top 10 in returning in 3 years, 91, 97 and 2000 ... 97 was his worst year in that decade by far, he was riddled with injuries as well, so I guess its reasonable to give a pass on that one

Originally Posted by pc1

I guess we just have to disagree. I don't want to keep beating a dead drum. My point is Agassi doesn't have enough when we examine his record to be call the return GOAT. We don't have all the information for the amount of games played. It's just not substantial enough information for me.

yeah, there isn't sufficient info to prove he is the best returner of all time, but there is more than enough to prove he was the best of his era ..

[abmk]

Going by the best servers of their era :

vs sampras : agassi leads by a mile

vs goran : I checked this and agassi leads this as well

vs becker : agassi owned him ( both returning and results wise ) and becker owned chang , don't even have to bother to check in this case

vs stich : agassi owned him ( both returning and results wise ), 5-0 and chang went 3-3 vs him. Is there really a need to check here ?

vs krajicek : I checked this and agassi leads this as well if we include their 2003 match , otherwise he is just marginally behind

as he played richard more on the faster surfaces grass,faster HCs and carpet than chang did

[krosero]

During the Agassi-Sampras final, ESPN said that Sampras led the tour in 1994 in service games held, with 88% won. At the 1995 AO, coming into the final, he stood at 94.5%.

Then Cliff Drysdale gave some numbers for Agassi, but it wasn't his break percentage. He said Agassi had great return percentages, but the graphic they put up said "Return of 2nd Serve." Agassi was at 59%, for all of 1994 (marked down as leading the tour). In his first six rounds at the 95AO, he stood at 62%.

So I think ESPN was providing a different stat for Agassi, about his success returning second serve -- not the number of times he broke. The Lee article has Agassi breaking serve only 34% of the time in 1994 (fifth in the ATP).

And besides, you're not going to find anyone breaking 59% of the time. I don't know, but has anyone ever broken more than 40% of the time? Coria had 39% in 2003, does anyone know of something higher?

But Drysdale did say that Agassi had the best return in the game.

And Pete did too, in an article in The Age (Melbourne):

From the first few points it was clear that Sampras, the best serve-volleyer in the business, would only occasionally be attacking at the net. He said later that this was because of the quality of Agassi's returns, "the best in the world by far".

The Age, incidentally, had picked Agassi to win in four.

[Datacipher]

Originally Posted by krosero

I'm not really clear what you mean. So let me make this point generally.

If it's a bad idea to correlate serve ability with ranking -- if it's not justified to assume that higher ranked players have better serving ability -- then it is equally unwarranted to assume that higher ranked players don't have better serving ability.

Just speaking generally, I would think the latter assumption carries a higher burden of proof.

So it may be possible that Agassi, while facing top players more frequently than Chang, did not face better serving ability. But is it likely? I would say no.

I agree that it is equally unjustified to assume the opposite. If I HAD to pick one guess as to the correlation, I'd choose, top players have better serves, although, I think it at best a very weak correlation. Traditionally, big servers and lesser servers have been all over the tour. My guess would be the top 10 would have a slightly higher average serving ability overall, but about the same "range" as other segments of the top 100. And in fact, if you look at the sheer volume of opponents Chang and Agassi faced somewhat "randomly" from the same pool of opponents, my guess is, they probably faced essentially exactly equal serving.

In any case, I was simply saying that in terms of measuring THE return, return games won is already a slightly tenuous measure. To then draw conclusions about the return based on another measure thrown in between....well it's so shaky, I just can't validate it.

To me, it would be like saying: Chang is better at steering a car than Agassi. I know this because Chang and Agassi live in the same apartment and Chang has gotten to work about 10 seconds faster on average over the last 15 years. Ok...MAYBE....BUT, then we notice that Agassi has driven cheaper cars to work slightly more often then Chang. Cheaper cars have worse steering (again...MAYBE), therefore, Agassi probably actually steers better than Chang...(OK, I'm outta here!!!).

Now admittedly, there may indeed be more significant extraneous variables in the above analogy, but my point is simply that the ever-increasing chain of assumptions is too much for me at that point.

Now if one were trying to measure effectiveness of return GAMES, then, I think the top 10 measure would have more effect, since presumably the top players hold more often, regardless of the specific reason why. EVEN then though, I wonder if the percentage difference between Agassi and Chang could possibly allow for true differentiation given the relatively small difference in both measures....

[Datacipher]

Originally Posted by pc1

And that's another thing I'm worried about. I don't want to be rigid so I'll give in to some points. It seems to me that point of the article was whether Agassi is the default choice for GOAT of service return and whether there is debate on his status as the best returner in his time. I think we'll muddling in too many different scenarios and examining whether it's top ten, top twenty. We now should go whether the serves of Chang's opponents are superior to Agassi's opponent. It's getting too complicated for a simple point and it's very simple. Can anyone say with 100% or 90% or even 5% confidence that Agassi is the best serve returner of his time? I think it's debatable.

I agree, it's ENTIRELY debatable. There is no specific data for this, and even that which we have does not really show a clear trend. So going into even more tenuous rough measure, on rough measure, is futile IMO.

(as I also mentioned, there are many factors, I'd even suggest surface before I looked at the ranking of opponents....)

All we have left are appeals to authority. Indeed, many of Agassi's peers pick him. I'm inclined to give that weight, but as we discussed, you can't always take these at face value by any means. The players are just as influenced by flash, and reputation, and perhaps even more prone to personal bias than we are. I am sure they would all rank both high, but Agassi higher? why? Would any be aware that Chang has broken them 3% more over their career matches? Might they remember the 3 HUGE forehands Agassi once hit? Remember how heavy that one backhand Agassi hit was, while forgetting that Chang hit one to them like that to? Actually NOT SERVE AS WELL, because IT IS Agassi over there?? If you're Becker, do your opinion count considering Agassi had a special read on you? Who knows? Besides the fact, that the guy with the best return of all time, may have peaked at #53 in 1974, because he had a mediocre backhand, and his serve choked under pressure....or maybe it was that journeyman doubles player...who never played the big guys in huge matches. Maybe it's Todd Woodbridge! lol....not as crazy as it sounds!

[pc1]

Originally Posted by krosero

thing is with Sampras, he had to have a great total game, backing up his serve, in order to be leading the tour in service holds. A complete game, plus mental consistency, year in and year out (in that respect he resembles Chang, not Agassi). So I think that's what Pete's hold statistics reflect: his ability to hold serve. That's the argument for Pete: that he had the best total serve-and-volley game (or serve-and-stay-back game, when he chose to do that) of his time. Theoretically, with his hold numbers, you would think that part of his total hold game included a serve that was, at least, one of the best of its time. And everyone agrees Pete's serve was at least one of the best of its time.

In the same way I think that everyone agrees that Agassi had, at the least, one of the best returns of his time. But the best return game of his time? I don't know if Agassi has ever had that argument made for him. Maybe he has. But that's not what I think people are referring to when they say he had the best return.

I always assumed McEnroe's famous statement about Agassi having a better return than Connors referred to the return as a single stroke. Now, maybe I read that into his statement. It's possible my assumption was wrong. But that's how I read it.

Anyway that's what I think the argument for Agassi has always been -- the best return of his time, considered as a single stroke.

Excellent point. I think perhaps we should redefine it as Total Return Game which I would say is the overall skills to help a player break serve. Using Connors as an example he not only had a great offensive return but he was very quick and returned many serves that many players would not return at all. Connors was also great defensively, able to lob, had great mobility and had superb passing shots on the run. I don't have the statistics on Connors but subjectively I would rank Connors higher than Agassi in a number of these categories. Agassi had the great single stroke in the return and he had great passing shots, but not he was not that great on the run as Sampras pointed out.

So Sampras can be redefined as having a great Total Serving Game for the same reasons. Goran may have a better overall first serve but Sampras over the years backed it up better.

I think if you try to define a single stroke like the service return, then we have to look at just two things, one is how consistent is the return and how good is the return? Agassi was obviously superb at hitting the offensive return and perhaps he had the best single return overall if you account for everything. My guess and it's only a guess is that a few players got the serve back more consistently. Coria for example had some excellent years returning but I don't think his returns were quite as aggressive as Agassi's.

Originally Posted by Datacipher

I agree, it's ENTIRELY debatable. There is no specific data for this, and even that which we have does not really show a clear trend. So going into even more tenuous rough measure, on rough measure, is futile IMO.

(as I also mentioned, there are many factors, I'd even suggest surface before I looked at the ranking of opponents....)

All we have left are appeals to authority. Indeed, many of Agassi's peers pick him. I'm inclined to give that weight, but as we discussed, you can't always take these at face value by any means. The players are just as influenced by flash, and reputation, and perhaps even more prone to personal bias than we are. I am sure they would all rank both high, but Agassi higher? why? Would any be aware that Chang has broken them 3% more over their career matches? Might they remember the 3 HUGE forehands Agassi once hit? Remember how heavy that one backhand Agassi hit was, while forgetting that Chang hit one to them like that to? Actually NOT SERVE AS WELL, because IT IS Agassi over there?? If you're Becker, do your opinion count considering Agassi had a special read on you? Who knows? Besides the fact, that the guy with the best return of all time, may have peaked at #53 in 1974, because he had a mediocre backhand, and his serve choked under pressure....or maybe it was that journeyman doubles player...who never played the big guys in huge matches. Maybe it's Todd Woodbridge! lol....not as crazy as it sounds!

Again, excellent points.

Originally Posted by krosero

And besides, you're not going to find anyone breaking 59% of the time. I don't know, but has anyone ever broken more than 40% of the time? Coria had 39% in 2003, does anyone know of something higher?

Since records have been only kept since 1991 I would tend to think players have broken the 40% barrier but I can't prove it. Coria seems to be the highest from what's I've seen in the records.

[krosero]

Originally Posted by Datacipher

I agree that it is equally unjustified to assume the opposite. If I HAD to pick one guess as to the correlation, I'd choose, top players have better serves, although, I think it at best a very weak correlation. Traditionally, big servers and lesser servers have been all over the tour. My guess would be the top 10 would have a slightly higher average serving ability overall, but about the same "range" as other segments of the top 100. And in fact, if you look at the sheer volume of opponents Chang and Agassi faced somewhat "randomly" from the same pool of opponents, my guess is, they probably faced essentially exactly equal serving.

In any case, I was simply saying that in terms of measuring THE return, return games won is already a slightly tenuous measure. To then draw conclusions about the return based on another measure thrown in between....well it's so shaky, I just can't validate it.

To me, it would be like saying: Chang is better at steering a car than Agassi. I know this because Chang and Agassi live in the same apartment and Chang has gotten to work about 10 seconds faster on average over the last 15 years. Ok...MAYBE....BUT, then we notice that Agassi has driven cheaper cars to work slightly more often then Chang. Cheaper cars have worse steering (again...MAYBE), therefore, Agassi probably actually steers better than Chang...(OK, I'm outta here!!!).

Now admittedly, there may indeed be more significant extraneous variables in the above analogy, but my point is simply that the ever-increasing chain of assumptions is too much for me at that point.

Now if one were trying to measure effectiveness of return GAMES, then, I think the top 10 measure would have more effect, since presumably the top players hold more often, regardless of the specific reason why. EVEN then though, I wonder if the percentage difference between Agassi and Chang could possibly allow for true differentiation given the relatively small difference in both measures....

I'm following your point now. It is a slight edge in Agassi's favor, that he faced top players a little more frequently than Chang did. And all things being equal, that probably means that Agassi faced players who held a little more frequently than Chang's opponents did. And all things being equal, Agassi's opponents held more frequently probably because they were better servers. So that makes Agassi's break numbers a little more impressive -- and as the final step, that probably means Agassi's return itself played the key role in those breaks. All of these are reasonable assumptions (certainly better than the opposite assumptions), but I agree, it's a lot of steps, and at best the edge that Agassi has here is a slight one -- as far as rankings go. Statistically, it could easily be that Agassi has no significant edge at all in the rankings issue.

Now if Agassi had a massive edge in facing top players, that would be something else. But like you I'd look at surface as a bigger variable in this sort of comparison. And who knows if either player has any edge there.

[krosero]

Originally Posted by pc1

Excellent point. I think perhaps we should redefine it as Total Return Game which I would say is the overall skills to help a player break serve. Using Connors as an example he not only had a great offensive return but he was very quick and returned many serves that many players would return at all. Connors was also great defensively, able to lob, had great mobility and had superb passing shots on the run. I don't have the statistics on Connors but subjectively I would rank Connors higher than Agassi in a number of these categories. Agassi had the great single stroke in the return and he had great passing shots, but not he was not that great on the run as Sampras pointed out.

So Sampras can be redefined as having a great Total Serving Game for the same reasons. Goran may have a better overall first serve but Sampras over the years backed it up better.

I think if you try to define a single stroke like the service return, then we have to look at just two things, one is how consistent is the return and how good is the return? Agassi was obviously superb at hitting the offensive return and perhaps he had the best single return overall if you account for everything. My guess and it's only a guess is that a few players got the serve back more consistently. Coria for example had some excellent years returning but I don't think his returns were quite as aggressive as Agassi's.

Yes, imo that would solve a lot of problems, redefining it that way. It's just more specific.

I think that the ESPN stats from the 95AO may be an example of how return differs from a return game. They wanted to show how Sampras and Agassi were each leading the tour, as best server and best returner. For Sampras that was easy -- he led the tour in hold percentage. So they just showed that stat. For Agassi, they showed him leading the tour in something specific to the return itself: something about returning second serves. I guess they showed that because Agassi was not leading the tour in return games won (he was only fifth).

Now, for someone like Connors, I don't see a big difference between return ability and the ability to break. Connors was good on the run and had great overall defense. Maybe not as good as Borg; maybe I would put Connors slightly higher if we just mean the return, and Borg higher if we mean ability to break serve overall. But I'd expect Connors to be higher than Agassi in break percentages. It's just a pity no such stats are available.

[krosero]

By the way, if ESPN in '95 could get the stats for who was leading the tour in return of second serves, then maybe the ATP kept records on unreturned serves. That stat would get you closer to judging the return itself, as distinct from the return game overall.

[pc1]

Originally Posted by krosero

By the way, if ESPN in '95 could get the stats for who was leading the tour in return of second serves, then maybe the ATP kept records on unreturned serves. That stat would get you closer to judging the return itself, as distinct from the return game overall.

There are stat sport services that research information for corporations. There is a company called STATS that used to publish books on statistics in baseball and I think they do research for companies like ESPN. I wonder if they can research information like this in tennis.

[abmk]

I calculated the nos for chang and agassi - surface wise breakup of matches played ..... had to do some manual work from the atp website for chang, grr, anyways here it goes



agassi of course played the higher percentage of matches on grass and hard, chang the higher percentage on carpet and clay ....

Can't get stats for distinguishing b/w slower and HCs though as it is subjective

[abmk]

pc1,

You have the stats ( and ranking ) for sampras holding serve , throughout the 90s,right ? Can you please post them here ? I'd like a comparision with roddick and even more interestingly federer !

[pc1]

Originally Posted by abmk

pc1,

You have the stats ( and ranking ) for sampras holding serve , throughout the 90s,right ? Can you please post them here ? I'd like a comparision with roddick and even more interestingly federer !

I'll try to get them to you later. It may have to wait until tomorrow but I'll get them out.

[pc1]

Originally Posted by abmk

pc1,

You have the stats ( and ranking ) for sampras holding serve , throughout the 90s,right ? Can you please post them here ? I'd like a comparision with roddick and even more interestingly federer !

1991-87%-Percentage of serve games held-2nd
1992-86%-4th
1993-90%-1st
1994-88%-1st
1995-89%-1st
1996-91%-1st
1997-91%-3rd
1998-89%-1st
1999-89%-2nd
2000-91%-1st

What's very impressive is that Sampras was the leader six out of ten years and was never lower than fourth. Very consistent. Clearly the best overall service game in tennis of those ten years.

[abmk]

Originally Posted by pc1

1991-87%-Percentage of serve games held-2nd
1992-86%-4th
1993-90%-1st
1994-88%-1st
1995-89%-1st
1996-91%-1st
1997-91%-3rd
1998-89%-1st
1999-89%-2nd
2000-91%-1st

What's very impressive is that Sampras was the leader six out of ten years and was never lower than fourth. Very consistent. Clearly the best overall service game in tennis of those ten years.

thanks for posting those ...

[Rick_Bremer]

Originally Posted by hoodjem

Here's mine:

1. Sampras
2. Ivanisevic
3. P. Gonzales
4. Roddick
5. Krajicek
6. Stich
7. Tanner
8. Newcombe
9. Tilden
10. Curren
11. Edberg
12. Federer
13. McEnroe
14. Becker
15. Dibley
16. Karlovic
17. Rusedski
18. Kramer
19. Fraser
20. Dent
21. Noah
22. Borg

I want sampras to be the number 1. like this list.

[hoodjem]

Does anyone know if Borg truly served at 135 mph at Wimbledon against Connors in 1978?

If so, that's pretty damn high for a wooden racquet and gut strings.