Better Server on Grass: Federer or Sampras ?

[NonP]

btw, Federers career % of service games won is actually 88.4% and Sampras' is 88.7%. So it's not even a 1% difference.

Remember, too, that Federers % is brought down by his early days on tour. If you compare their best 10 year run, Federer has won a higher % of service games.

Not to say his serve his better than Pete's, it isn't, but the service games are close. I can guarantee you that there exists a 984 match stretch in Federers career (significant because Sampras PLAYED 984 matches in his career) where his % of service games won is higher than 88.7. Well, I mean, of course there is.

Misleading on just about all counts. A couple of things:

- As I've pointed out time and again, and contrary to the talking points you hear about the supposedly slower surfaces and better returns, players today are holding serve more frequently than they did in the '90s and probably ever. According to the ATP's stats (which BTW aren't very reliable with regards to the '90s--more on this below) the top 10 servers of the '00s actually won a higher % of service games than their 1990s predecessors, 87.1% vs. 86.0%, and though I don't have the breakdown for this decade yet I suspect the %s have if anything gone up since then. And get this: players won 90% or more of their service games only seven times in the '90s (the total number of players is 3, with only Sampras achieving the feat more than once), as opposed to a whopping 20 times in the '00s.

The reasons for this unexpected improvement in service stats are complex but in short, players are winning more service games these days because today's racquets allow for greater spin than ever, which in turn allows for more 1st serves and heavier spin on 2nd. Take a gander at this:

http://www.atpworldtour.com/en/stats/1st-serve/all/all/all/

As you can see a vast majority of the top 100 players here are from the '00s and beyond. That's no accident, and also despite the quite possible inflation of many of the '90s players' %s (again more on this in a bit).

Put another way Pete's % of service games won would likely go up in this era and Fed's down in the '90s. If you really think a mere 2-3% difference is so significant (which you apparently do, as you put up nary a fuss in response to the poster that made the claim) the % of service games won isn't a stat you want to focus on.

- This by-decade comparison doesn't give you a full idea of just how much the game has shifted in favor of the server (again contrary to the talking point about players returning better than ever). For example even Goran and Krajicek had fewer than 10 aces per match in 1991, which is just unfathomable these days (I remember mattennis expressing his shock when I pointed this out). Something you might want to keep in mind before you cherry-pick particular years to suit your purpose.

BTW Fed's % of return games won from '03-'10 is 28.7% as opposed to Pete's 26.3% from '91-'98, this despite taking the latter's non-Slam years in '91 and '92 and anomaly in '96 when he won a mere 21.1% (obviously a lot going on both on and off court that year). But unlike you I won't be so selective in my dissection. And one thing you didn't consider when you mentioned Fed's superior TB record is the fact that Pete actually played fewer TBs per match in his major final wins than Fed in his own. In other words, and contrary to expectations, Pete actually broke more often than Fed when it counted. But again unlike you I fully acknowledge that this is cherry-picking and won't put up a pretense of impartiality by praising Fed as a GOAT returner after declaring my guy better.

Do you realize just how "money" a serve has to be when a player is wholly dependent on it? Karlovic has been known to go weeks without breaking serve. And that part is kind of important to the whole "winning matches" thing. It doesn't matter how great a serve is, if given a near-infinite amount of chances to implode, it eventually will. And that's what happens with Karlovic in the rare case that he gets broken. As a standalone shot, it blows Pete's out of the water.

Karlovic does have the better serve, but you'd be surprised at how small the gap is between him and Pete (or almost any other GOAT server, for that matter). People like to look at ace counts for serve comparisons and indeed the likes of Ivo, Goran and Raonic often post ridiculous numbers in this department. What they don't look at, though, and a better measure of serving prowess, is the % of serves that their opponents fail to put back in play. As one can expect these guys serve plenty of these freebies as well, usually over 40% of unreturned serves, often 50%, sometimes even 60%. Another guy that regularly posted these numbers? You guessed it: Sampras, and also Roddick to a somewhat lesser extent, despite their considerably lower ace frequency. (Curiously Isner is underwhelming in this stat, though the available sample size for him so far is rather small.) There are several possible reasons for this surprising comparison but the biggest one I suspect is the high topspin component of Sampras' serve that Yandell once analyzed in his oft-cited article.

Also your breakdown of serve vs. return games, while popular and valid on paper, is too simplistic. In short a player's overall game has a much bigger impact on his return game than on his service game (which makes sense, when you consider that players overall hold about three times more often than break serve). Go look up the service/return stats of Pete, Fed, Agassi and almost anyone else, and you'll see that a 4-5% drop in their return games won hurt their results much more than the same decline in their service games, which in fact don't see such fluctuation for much of their career.

And you're not even considering the difference between Pete's and Ivo's 2nd serve, and their different approach to the game, which brings us to this:

Karlovic has a better serve in almost every respect. Pace, accuracy, placement and kick. It's just better. So so so so so so much better. Denying it is just being contrarian.

These are your impressions and I won't bother to dispute them, except two:

- If you really think pace matters so much (it doesn't--once you get to the 120-130 mph range which both Pete and Ivo certainly did/do regularly disguise and placement are much more important) then you should acknowledge that Pete served with significantly faster pace than Fed. But on this very thread you argue that Pete's and Fed's serves aren't so much apart after all while at the same time claiming (without much basis, I should point out) that Ivo's "blows Pete's out of the water." This irony flew over your head.

-I don't know what you mean by "kick," but if you meant the usual topspin kicker I'd like to see what data if any you have to support this claim, not only because of Pete's high topspin component I just mentioned but because Ivo's career % of points won on 2nd serve (which of course is usually a kicker) is a good but not great 53%. Now before you or anyone else points out that Pete has the same career % it's important to keep in mind a couple of things. One, the higher spin afforded by today's racquets I talked about earlier. And two, a lot of the '90s match stats on the ATP site are wildly off, because their statisticians somehow counted aces and double faults twice for most Slam matches up to 2000. (Now there are some non-Slam events like the '92 YEC at Frankfurt that were subjected to the same bastardization, but such instances should be rare.) krosero and slice serve ace have explained this in more detail but let me take a brief stab at it.

Say a certain player posted these actual stats:

60 of 100 1st serves in (60%)
48 of 60 1st-serve points won (80%)
20 of 40 2nd-serve points won (50%)

And let's say he served 10 aces and 5 DFs in this same match, which would show up like this on the ATP site:

70 of 115 (after adding 10 & 5 to 100) 1st serves in (61%)
58 of 70 1st-serve points won (83%)
20 of 45 2nd-serve points won (44%)

Notice how the % of 1st-serve points won went up by 3% but the 2nd-serve % dropped twice as much by 6%. Check the ATP stats against those in the press reports, and you'll see that the former are off for just about every pre-'00 Slam match. Given such discrepancies it's hard to say what Pete's career % of 2nd-serve points won actually was, but make no mistake, it was certainly higher than the 53% shown on the ATP site. slice serve ace has provided a (partial) correction of the '90s top 10 players' 2nd-serve stats here:

http://tt.tennis-warehouse.com/inde...pts-won-for-1992-1999-corrected-stats.469808/

And you're not taking into account the fact that Pete S&Ved a lot more often than Ivo or nearly anyone else in this era even on 2nd serves, which makes this comparison trickier. After all would Karlovic really rather use his own 2nd serve than Pete's or for that matter Roddick's or Isner's for S&V? I doubt it, but apparently some of the geniuses here think otherwise. Not to mention that Pete faced top-ranked players more often which as expected hurt his % (this particular comparison was between Pete and Roddick but the same logic applies here):

http://tt.tennis-warehouse.com/index.php?threads/roddick-vs-sampras-the-serve.451723/#post-7152066
http://tt.tennis-warehouse.com/inde...-sampras-the-serve.451723/page-2#post-7228245

Long story short, these cross-era comparisons aren't as simple as you and your ilk think, and when you do, you're wrong.

P.S. This is supposed to be one of the more "reasonable" posters in this joke of a forum. Good grief.

 

[TheFifthSet]

^lol. Talk about being selective.

Please, don't enter these threads with a preconceived notion about people. I wasn't arguing that Petes and Feds serves "weren't so much apart". I was talking about their.... AHEM.

WAIT FOR IT.

SERVICE GAMES.

And they're not. But their serves? Oh, no question. Pete's is way, way better. So, try again. I would never try to argue that Federers serve is better than Pete's. I'm not TMF. All-around game? Yes. Serve? No. Service game? Well, its complemented by a players all-around game, where in my opinion Federer is superior, so its close.


"Not to say his serve is better than Pete's, it isn't, but the service games are close"

That's what I said. I seriously don't think I could have been any clearer. For you to misread THAT is bizarre.


So, I'll state it again in plain English, in case it flies over YOUR head: Pete's serve is way better than Feds. Karlovic's serve is way better than Pete's. Nope, no irony, and no hypocrisy to be found in my post.


Maybe next time, when you try to dissect someone's post, read between the lines a little better. Plz, thnx, and bai bai.

 

[TheFifthSet]

It's hilarious, because I JUST was lamenting how people can't seem to make the distinction between a serve and a service game.

PS: if you think Karlovic's serve wasn't better than Pete's by a significant margin, then count me out of this thread. I'm bailing. Much fuss is made about Ivo supposedly stat-padding, so to speak, against lower ranked players, yet his % of service games won against the top 10 is astonishingly high, despite literally only possessing one tour-caliber stroke. A serve. And I guess he's good at hitting put-away volleys off weak returns. That's it.

PPS: I don't recall ever having been called "one of the more reasonable posters on this board". If you don't like this place, you can show yourself the door. Not that I hope you do, you seem like an intelligent fellow (other than the reading comprehension glitch there, but I'll assume you just skimmed through my post), but hell, no ones forcing you to post here.

 

[TheFifthSet]

BTW Fed's % of return games won from '03-'10 is 28.7% as opposed to Pete's 26.3% from '91-'98, this despite taking the latter's non-Slam years in '91 and '92 and anomaly in '96 when he won a mere 21.1% (obviously a lot going on both on and off court that year). But unlike you I won't be so selective in my dissection. And one thing you didn't consider when you mentioned Fed's superior TB record is the fact that Pete actually played fewer TBs per match in his major final wins than Fed in his own. In other words, and contrary to expectations, Pete actually broke more often than Fed when it counted. But again unlike you I fully acknowledge that this is cherry-picking and won't put up a pretense of impartiality by praising Fed as a GOAT returner after declaring my guy better.

Yes, this is quite a dissection.

You probably weren't going to address at any point that Pete's return stats in his non-slam winning years remained roughly the same, and he broke serve around 26.5% of the time in '91-'92. Or that Federer ALSO had an anomalous returning year, in 2009, where he won 24% of his return games. Either way, a 2.4% edge in return games over an 8 year span is not something to gloss over.

I'm aware of the 90's statistical discrepancy, btw. Even if we bump up Pete's second serve an extra 2%, all this does is demonstrate that a man with one of the most impressive all-around games of all time wins a slightly higher % of his second serve points than a man that arguably possesses the worst ground game out of any top 30 player EVER. Chip away further by bringing up Pete S&V'ing on his second serve so often. That's a fair point. Still doesn't come close to explaining why their second serve stats are so close, IMO. Ivo's supporting game is just THAT bad. And even once the 2000's commenced, Pete's second serve didn't really improve that dramatically, if even at all. He won 53% of his second serve points those last 3 years. If these statistical errors are enough for you to conclude that their serves are "closer than you'd think", then I don't know what to tell you. How's this for a statistic: Karlovic lands his first serve in a whopping 7% of the time more, and still wins a higher percentage of first serve points that 90's Pete, whose first serve statistics were bumped up by that very same statistical miscalculation you mentioned. It's just not close. The man has so much more margin on his serve than Sampras. It's just flat-out better, and it sure as hell IS NOT "surprisingly closer". The Roddick-Sampras comparison is irrelevant here. Go compare KARLOVIC to SAMPRAS. Compare their % of service games won against top players. Go ahead. I don't think it'll yield the result you're hoping for.

And I mentioned the % of tiebreaks won over their WHOLE CAREERS, saying that the serve/return games won over their WHOLE CAREERS don't fully articulate Federer's superiority. Of course, this isn't what one would call a smoking gun, but it's another reason to believe Federer was superior as a player. So no, this is no façade. There is no "pretense of impartiality". I'm VERY partial to Federer, and I was indeed trying to prop up his career and compare him favourably to Sampras. I believe he's a greater player than Sampras, and I also believe that Sampras quite clearly had the better serve.

 

[World Beater]

does anyone have stats on speed numbers of first serves over their careers.

much has been said about pete serving much harder than fed. I think over the course of their careers, pete probably served ~5 mph faster on the first serve on average.

But federer in his prime was serving 125-130 and has been clocked at 137. So i don't think the difference is really that big.

However, i am open to changing my mind if the facts do not back it up.

 

[jhhachamp]

Give Sampras Karlovic's serve and he wins 20-25 majors. Give Federer Karlovic's serve and he wins over 30 majors. Give Karlovic Sampras' or Federer's serve and he likely does not break the top 200. Give Karlovic an average serve (something like Robredo's serve), and he likely never breaks the top 500.

Karlovic's serve is that good and the rest of his game is that bad. I've seen up close in matches and in practice.

 

[Chadillac]

Its hard to tell because of the slow surfaces now a days. But feds serve on the glass of the 90's would be equally as devistating.

Technically pete had less to back his up so id give him the nod and the fact fed in the 90's is speculation

 

[TheFifthSet]

does anyone have stats on speed numbers of first serves over their careers.

much has been said about pete serving much harder than fed. I think over the course of their careers, pete probably served ~5 mph faster on the first serve on average.

But federer in his prime was serving 125-130 and has been clocked at 137. So i don't think the difference is really that big.

However, i am open to changing my mind if the facts do not back it up.

Pete's serve was way pacier.

Even in their 2001 match, his avg first serve speed was 125, Feds was at 114. I've made it a habit to look at post-match stats over the years, and Federers first serve speed has always been, without fail, ranging from 114-120. He can clock a few serves in the 130 range, yes, but Sampras AVERAGED 122-128 on his first serve for years.

 

[TheFifthSet]

Give Sampras Karlovic's serve and he wins 20-25 majors. Give Federer Karlovic's serve and he wins over 30 majors. Give Karlovic Sampras' or Federer's serve and he likely does not break the top 200. Give Karlovic an average serve (something like Robredo's serve), and he likely never breaks the top 500.

Karlovic's serve is that good and the rest of his game is that bad. I've seen up close in matches and in practice.

QFT

 

[Chanwan]

Is this thread a joke? Federer isn't even on Roddick's level, nor was he ever on the level that Krajicek showed in his Wimbledon run. Nor was his level of serving ever close to Goran. Sampras was the best server of the 90's and the others mentioned are the best of the rest, in recent times. Why? Because these guys produced serving displays which changed the was the game is played today, and they served at a level which many ignorant posters seem to label as boring or simply 'not tennis'. Federer's serve is very good, but his serve was never as dominant to the point tennis fans called his game boring.




Edit: Just realised TMF created this thread. This thread was clearly started as a joke. My bad.

I like you quite a bit as a poster, but I seriously don't get how you can maintain Sampras as the best server ever (is that a joke?), when there's so much evidence against it (not that Fed is at his level or better, but others certainly are - Roddick and Goran included).

Edit, I see my point has been made very well above:

Karlovic and Isner hold serve 85-90% of the time against Federer, Nadal, Djokovic and Murray. Problem is, they break them less than 10% of the time.

Karlovic in his career has won an unfathomable % of first serve points - 82%. Both Karlovic and Isner have held serve at a much higher rate, despite playing in this era of slow courts.

People just want to sound like informed skeptics by saying Sampras had a better serve, when it is so demonstrably inferior. He didn't. It's not just pace. Karlovic and Isners serves are perfect in every single way. Pace, placement, and accuracy. Karlovic lands his first serve in 5% of the time more than Sampras did, and Isner 7%, and they still win a higher pct% of first serve points. And second serve points, too. They're also so tall that the kick they get on their second serve is enormous. The pace on their second serves are better too, and they double fault less.

If Karlovic had Sampras' all-around game, he would have won everything in sight and would be the undisputed, unanimous GOAT of tennis. Does anybody seriously think that with a Sampras-like game, ANYBODY could break Karlovic? I sure don't, because it wouldn't happen. This isn't hyperbole. Karlovic has had SEASONS where he holds serve 95% of the time. Now, regardless of whether or not his competition was on par, that's just incomprehensible for a player that has a ground game like his. Does anybody actually break Karlovic without him flubbing several sitter volleys? The guy breaks himself.

Give Sampras Karlovic's serve and he wins 20-25 majors. Give Federer Karlovic's serve and he wins over 30 majors. Give Karlovic Sampras' or Federer's serve and he likely does not break the top 200. Give Karlovic an average serve (something like Robredo's serve), and he likely never breaks the top 500.

Karlovic's serve is that good and the rest of his game is that bad. I've seen up close in matches and in practice.

you're right except that Karlo would still break top-100 and possibly top-50 too with Sampras/Fed serve. They are that good too, but still a huge notch below his own.

 

[NonP]

It's hilarious, because I JUST was lamenting how people can't seem to make the distinction between a serve and a service game.

I like how you completely ignored the part about Pete's and Ivo's %s of unreturned serves not being so far apart and about a player's overall game having a significantly more effect on his return game than on his service game. I say you're still making too much out of this distinction, but fair enough about your intentions. I apologize for that part.

PS: if you think Karlovic's serve wasn't better than Pete's by a significant margin, then count me out of this thread. I'm bailing. Much fuss is made about Ivo supposedly stat-padding, so to speak, against lower ranked players, yet his % of service games won against the top 10 is astonishingly high, despite literally only possessing one tour-caliber stroke. A serve. And I guess he's good at hitting put-away volleys off weak returns. That's it.

Again I do think Ivo's serve is clearly better, and in fact he and Goran are the only ones I put above Sampras without much hesitation in my own ranking. And FYI I recently added Fed myself to the GSOAT list (which in case you don't know has been a longtime work in progress), so I do rate Fed's serve pretty damn high. See here:

http://tt.tennis-warehouse.com/inde...erves-of-all-time.306579/page-19#post-9180812

But yes, I do think the gap between Ivo and Pete (or again ANY other great server) isn't as great as it's made out to be. Again the % of unreturned serves is a more accurate barometer of serving prowess than ace frequency, and trust me when I say Pete's stats here are quite comparable to those of Ivo, Goran, Krajicek, Isner, you name it. You're talking about a guy that often had over 50%, sometimes even 60% of his serves unreturned by his opponent, and you won't find many matches by Karlovic or anyone else where he had a much higher share of freebies. (FWIW Fed's numbers here generally are decent but not that crazy good. In fact his highest % of freebies I know of is from none other than his lone match against Pete where he had 49.2% of his serves unreturned, though I wouldn't be surprised to see a few instances of higher % from his recent serving streak.)

And you should check out the last thinks I provided above. There's no question servers hold less often against superior competition with presumably better return games. Now I do often take the time to point out that the competition factor is exaggerated, but when you're talking like a mere 2-3% difference is a big deal the quality of one's opponents shouldn't be overlooked.

(I see that you just replied to the part about the return games. Will get to it in a minute.)

PPS: I don't recall ever having been called "one of the more reasonable posters on this board". If you don't like this place, you can show yourself the door. Not that I hope you do, you seem like an intelligent fellow (other than the reading comprehension glitch there, but I'll assume you just skimmed through my post), but hell, no ones forcing you to post here.

FWIW the only reason why I responded to your posts in particular is because I thought you'd be more open to reason than most. It's really more a compliment than an insult, though I admit I came across as crankier than I liked.

You probably weren't going to address at any point that Pete's return stats in his non-slam winning years remained roughly the same, and he broke serve around 26.5% of the time in '91-'92. Or that Federer ALSO had an anomalous returning year, in 2009, where he won 24% of his return games. Either way, a 2.4% edge in return games over an 8 year span is not something to gloss over.

My point here was that Fed's got a mere 2.4% edge over Pete in return games won despite factoring in Pete's two non-Slam years in his '91-'98 stretch and also his tough going in '96, and I could've cherry-picked only their best years to make the comparison even closer. To put it more bluntly, cherry-picking ain't no good.

I'm aware of the 90's statistical discrepancy, btw. Even if we bump up Pete's second serve an extra 2%, all this does is demonstrate that a man with one of the most impressive all-around games of all time wins a slightly higher % of his second serve points than a man that arguably possesses the worst ground game out of any top 30 player EVER.

That's quite possible, yes. After all you're talking about a guy with the lowest % of career return games won that I know of. But I still think you're exaggerating when you say all he's got is a serve. Ivo himself has talked about how people ignore his FH, and dude's actually got a pretty nasty slice, I'd say better than, for one, Rafa's.

Also you haven't addressed my point about one's overall game having a significantly bigger impact on his return game than on his service. I elaborated on this topic on the recent thread about Muster vs. Federer, you can go back and look if you're still interested.

Chip away further by bringing up Pete S&V'ing on his second serve so often. That's a fair point. Still doesn't come close to explaining why their second serve stats are so close, IMO.

Per the ATP the # of players with 51% or more of their 2nd-serve points won is 116, so if you insist on the 2.4% difference in Pete's & Fed's return games won being such a big deal (again this despite the stat very much geared to Fed's favor) then you must also concede that the 2% difference in 2nd-serve points won between Pete & Ivo can't be overlooked.

And even once the 2000's commenced, Pete's second serve didn't really improve that dramatically, if even at all. He won 53% of his second serve points those last 3 years.

But since you insist that a player's % of service games won doesn't depend so much on his serve shouldn't you also acknowledge that the last 3 years of Pete's career were his worst since the early '90s?

How's this for a statistic: Karlovic lands his first serve in a whopping 7% of the time more, and still wins a higher percentage of first serve points that 90's Pete, whose first serve statistics were bumped up by that very same statistical miscalculation you mentioned.

I do agree that Ivo's high 1st-serve % is indeed a point in his favor, but again you're not taking into account the higher spin afforded by today's racquets. It's all but certain that Pete would be posting higher 1st-serve %s in this era.

The Roddick-Sampras comparison is irrelevant here. Go compare KARLOVIC to SAMPRAS. Compare their % of service games won against top players. Go ahead. I don't think it'll yield the result you're hoping for.

I should sometime, but having already perused many of Karlovic's match stats I doubt the competition factor is as minor as you seem to think.

Pete's serve was way pacier.

Even in their 2001 match, his avg first serve speed was 125, Feds was at 114. I've made it a habit to look at post-match stats over the years, and Federers first serve speed has always been, without fail, ranging from 114-120. He can clock a few serves in the 130 range, yes, but Sampras AVERAGED 122-128 on his first serve for years.

No, Pete's averages for years actually didn't quite register that high, because of the radar. It's unclear exactly when radar guns began measuring serve speeds upon contact with the ball rather than at the net, but it almost certainly happened around the turn of the century as that's when average serve speeds for everyone saw a significant rise. Also Pete didn't begin averaging 110 mph average on 2nd serves until late in his career, largely because he had to (much like Fed is doing right now to make up for the loss in his ground game).

I'm not so sure about a previous poster's claim that Pete is still averaging 130 mph on his 1st and 110 mph on his 2nd, since last time I saw him he didn't seem to have near that pace and that was over a year or two ago. But it's true that his averages in the lone match (2nd one) of his exos against Fed with a radar gun were well above what he used to average in his prime years (hat tip to slice serve ace again):

http://tt.tennis-warehouse.com/inde...serves-of-all-time.306579/page-5#post-4456150

 

[zagor]

Count on TMF to overreact after a Fed win, Sampras' serve is better on the whole without question but Fed can reach Pete's level on very rare one-off serving performances like the one against Murray few days ago.

 

[Tennisanity]

Pete's serve probably wouldn't be as good in today's environment.

 

[Bobby Jr]

From my viewing experience, this is completely false. Karlovic and Isner, because of how tall they are, have much better placement than anyone else. Their height allows them to put balls in places Federer and Sampras are physically incapable of doing. Why would you even need accuracy when you have much bigger margins for error like these guys do?..

Accuracy and the ability to get angles are not the same thing. If a guy as tall as Isner or Karlovic had Federer or Sampras's accuracy they would be basically unbreakable. But they don't. That's a big part of why they're nobodies in the scheme of things (yes yes, in addition to all the other factors).

It's not like Djokovic ever faced the best version of Roddick, which was 2003-2005. And Roddick did beat Djokovic at the AO in 2009, when Djokovic was the defending champion.

Oh, so now the details do matter? That was my point.

The point is that most of these details, on paper, should favor Sampras. Roddick has played in an era with slower surfaces, and has a worse overall game than Sampras...

...and poly strings and more modern, easy power/spin frames.

Give Sampras those advantages and test the theory again.

 

[The_18th_Slam]

Accuracy and the ability to get angles are not the same thing. If a guy as tall as Isner or Karlovic had Federer or Sampras's accuracy they would be basically unbreakable. But they don't. That's a big part of why they're nobodies in the scheme of things (yes yes, in addition to all the other factors).

I didn't say accuracy and angles were the same thing. I said accuracy doesn't matter when you have the margins they do. Again, not that they don't have accuracy, because they do, which is why they are in the top 32 and some other very tall ATP player isn't even heard of.

Oh, so now the details do matter? That was my point.

Nope, your point was the H2H. When I showed you how you were wrong, you jumped to details, ignoring the details which didn't favor the point you were making. But nice try.

...and poly strings and more modern, easy power/spin frames.

Which also favor returning a great deal.

Give Sampras those advantages and test the theory again.

Those aren't necessarily advantages. They favor returners, too. Again, you're only seeing things you want to see, and noticing details that favor you, without looking at the whole picture. Or maybe you're just being intellectually dishonest.

 

[World Beater]

Pete's serve was way pacier.

Even in their 2001 match, his avg first serve speed was 125, Feds was at 114. I've made it a habit to look at post-match stats over the years, and Federers first serve speed has always been, without fail, ranging from 114-120. He can clock a few serves in the 130 range, yes, but Sampras AVERAGED 122-128 on his first serve for years.

I don't think their 2001 match is a fair comparison.

Federer was hardly physically mature at 19yrs whereas sampras was 29 years old, and hitting the ball very hard.

Federer was serving much harder when he was 26/27 than when he was 19.

to say pete's serve is way pacier is great, but i am looking for stats and evidence to back that up.

 

[suppersready]

I tend to think Sampras has the slightly better serve because of the extra mph, but I also think federer holds back a bit to avoid injury.

 

[Deleted member 688153]

Sampras fans out in force.
Nervous for the final, eh?

 

[TheFifthSet]

I like how you completely ignored the part about Pete's and Ivo's %s of unreturned serves not being so far apart and about a player's overall game having a significantly more effect on his return game than on his service game. I say you're still making too much out of this distinction, but fair enough about your intentions. I apologize for that part.

No problem, simple misunderstanding.

Yes, % of unreturned serves ARE a better gauge of serving prowess. Absolutely. I should have addressed that. I was posting on my phone, and thus I really didn't have the patience to get to everything.

As far as Pete and Ivo are concerned, though, you didn't really make a direct comparison save for saying "they both have a historically high rate of unreturned serves". Well, duh. But, what ARE their respective rates of unreturned serves? For their careers, and not a match here and there. I'd wager that Ivo has a healthy edge.

If we had a comprehensive set of these statistics at our disposal, I'd use those instead of aces and % of service games won. If only there were a "percentage of meek returns a serve elicits". Any doubt that Karlovic would win THAT one too? I have none.

Again I do think Ivo's serve is clearly better, and in fact he and Goran are the only ones I put above Sampras without much hesitation in my own ranking. And FYI I recently added Fed myself to the GSOAT list (which in case you don't know has been a longtime work in progress), so I do rate Fed's serve pretty damn high. See here:

http://tt.tennis-warehouse.com/inde...erves-of-all-time.306579/page-19#post-9180812

That's one of the better threads on this board. Maybe I'll post my own list.

But yes, I do think the gap between Ivo and Pete (or again ANY other great server) isn't as great as it's made out to be. Again the % of unreturned serves is a more accurate barometer of serving prowess than ace frequency, and trust me when I say Pete's stats here are quite comparable to those of Ivo, Goran, Krajicek, Isner, you name it. You're talking about a guy that often had over 50%, sometimes even 60% of his serves unreturned by his opponent, and you won't find many matches by Karlovic or anyone else where he had a much higher share of freebies. (FWIW Fed's numbers here generally are decent but not that crazy good. In fact his highest % of freebies I know of is from none other than his lone match against Pete where he had 49.2% of his serves unreturned, though I wouldn't be surprised to see a few instances of higher % from his recent serving streak.)

I also do not believe that there is pronounced difference between most great servers. Karlovic is an exception. He doesn't have a top 1000 ground game yet holds serve 92% of the time in his career. Incomprehensible. And I am convinced that his unreturned serve rate for his career would stomp the yard. Whether the margin between Karlovic and Sampras is "huge", or just "very clear", there's not much to pin on me here. At worst I was being hyperbolic, according to you. Remember that you agree with my central premise that Karlovic is a better server. The original post that you were responding to was a response to a poster that claimed that Sampras has a much better serve, and insinuated that the reason Karlovic doesn't notch more scalps over top players is because his serve pulls a disappearing act in big moments.

Which is further from the truth?

And you should check out the last thinks I provided above. There's no question servers hold less often against superior competition with presumably better return games. Now I do often take the time to point out that the competition factor is exaggerated, but when you're talking like a mere 2-3% difference is a big deal the quality of one's opponents shouldn't be overlooked.

It isn't overlooked. It just doesn't offset everything else, especially when all the existing evidence indicates that Karlovic holds serve AGAINST THE TOP TEN at about as high/higher rate than Sampras did against the entire field, with a worse game to back up the serve.

FWIW the only reason why I responded to your posts in particular is because I thought you'd be more open to reason than most. It's really more a compliment than an insult, though I admit I came across as crankier than I liked.

Thanks, the sentiment is appreciated. Yeah, I also got testy with you, as I was taken aback by your disparaging comment. Which is also more of a compliment than an insult, as that's clearly not your M.O.

I sincerely am open to changing my mind on any topic given sufficient reason, and have many times even just this year. In this case, I believe there is a lot of overlap between our opinions, but not enough for us to totally be on the exact same page.

My point here was that Fed's got a mere 2.4% edge over Pete in return games won despite factoring in Pete's two non-Slam years in his '91-'98 stretch and also his tough going in '96, and I could've cherry-picked only their best years to make the comparison even closer. To put it more bluntly, cherry-picking ain't no good.

I wasn't intending to cherry-pick, though. Their return statistics and % of tiebreaks won relative to how many they are expected to win, provide very practical evidence that Federer is a better tennis player. The tiebreaks are a small proof that the % of total games won is further in Federers favour than the return/serve games won alone would have you think. This wasn't meant to be some conclusive proof. Just another link in the chain.

That's quite possible, yes. After all you're talking about a guy with the lowest % of career return games won that I know of. But I still think you're exaggerating when you say all he's got is a serve. Ivo himself has talked about how people ignore his FH, and dude's actually got a pretty nasty slice, I'd say better than, for one, Rafa's.

I would dispute that. His slice is a passable shot. He doesn't carve it that well and it has a tendency to sit up. Nadal isn't known for having a good slice, either.

His rally forehand is not a good shot. He can slap it well when he gets a hold of one, often at triple digits. But it's not a tour-average forehand. His serve is his ONLY weapon of note. Without it he wouldn't crack the top 1000. Again, I suspect we are engaging in semantics here. In a literal sense, of course Karlovic is not just a serve. He needs to win points on his opponents serve SOME of time, hit SOME knifing volleys and put away SOME forehands. But he doesn't do any of that with any regularity. So I believe I am justified in my assessment.

Also you haven't addressed my point about one's overall game having a significantly bigger impact on his return game than on his service. I elaborated on this topic on the recent thread about Muster vs. Federer, you can go back and look if you're still interested.

I don't disagree with this, I mean of course it does You don't have the game-changing edge of having the first strike on the return. That in itself is a weapon, good serve or not. So yeah, not one iota of disagreement there.

But a supporting game as horrendous as Ivo's doesn't do him any favours. Against a top player it REALLY is "unreturned serve/sitter return or bust".

Per the ATP the # of players with 51% or more of their 2nd-serve points won is 116, so if you insist on the 2.4% difference in Pete's & Fed's return games won being such a big deal (again this despite the stat very much geared to Fed's favor) then you must also concede that the 2% difference in 2nd-serve points won between Pete & Ivo can't be overlooked.

I wouldn't say it is "such" a big deal, other than to hit home the point that Federer is a better returner. It's a completely different set of circumstances.

I conceded a 2% edge to be diplomatic. There's no telling how much/little the numbers are fudged. What we DO know is, Karlovic DID win a higher % of total service points, a higher percentage of first serve points, and landed SIGNIFICANTLY more first serves in. Like, super-duper significantly more. So yes, in the context of this discussion, Karlovic winning a slightly lower percentage of second serve points (lets say 1.5-2% less), with everything else he has in his favour, is not enough to sway me.

We also know that, with the "correct" stats, from 2000-2002, Sampras won around 53% (I might be rounding up, mind you) of his second serve points. Yes, his game had deteriorated by that point, but....well...do we wanna go down that road? Pete's ground game TODAY would be better than Ivo's. So we can't use that.

You also omit that Federers edge on the return lies more in his defensive return. His first serve return is much better than Pete's. His second serve return is better, but not by a huge margin. So the difference in their second serve return points won is not the primary reason he is a better returner, and thus not a totally apt comparison.

 

[TheFifthSet]

No, Pete's averages for years actually didn't quite register that high, because of the radar. It's unclear exactly when radar guns began measuring serve speeds upon contact with the ball rather than at the net, but it almost certainly happened around the turn of the century as that's when average serve speeds for everyone saw a significant rise. Also Pete didn't begin averaging 110 mph average on 2nd serves until late in his career, largely because he had to (much like Fed is doing right now to make up for the loss in his ground game). I'm not so sure about a previous poster's claim that Pete is still averaging 130 mph on his 1st and 110 mph on his 2nd, since last time I saw him he didn't seem to have near that pace and that was over a year or two ago. But it's true that his averages in the lone match (2nd one) of his exos against Fed with a radar gun were well above what he used to average in his prime years (hat tip to slice serve ace again):

I'm aware of the history of how service speeds were registered, and of the spike in '99-'00. Keep in mind, I meant that Pete's first serve speeds were that high if we use the updated method. We can assume this by tracking his speeds in the early 2000's; they WERE that high. He routinely averaged in the mid-120s on the first serve, with the same frame he used in his prime. More importantly, his serve was much pacier than Feds, which is what I was responding to.

But since you insist that a player's % of service games won doesn't depend so much on his serve shouldn't you also acknowledge that the last 3 years of Pete's career were his worst since the early '90s?

Actually, it depends more on the serve than on anything else (another "duh"), so your paraphrasing is a bit of a strawman. It does depend "so much" on the serve. That's why the server holds over 4/5 of the time. It's just not the only factor. Nadal, with a so-so serve, has had seasons where he held 87-90% of the time. Yes, Pete's serving numbers were at their worst those last 3 years, as you'd expect. But his game was still so good that I cannot envision a scenario where Karlovic is endowed with Pete's all-around game, even in its diminished state, and "only" holds 86-87% of the time.

I do agree that Ivo's high 1st-serve % is indeed a point in his favor, but again you're not taking into account the higher spin afforded by today's racquets. It's all but certain that Pete would be posting higher 1st-serve %s in this era.

Speaking of which, are there any figures on first serve % now compared to the 80s/90s? Not disagreeing with you here, just curious. It's not an unsound assumption, you're right. But Ivo lands 2/3's of his first serves in. That's insane. Ivo creates HIS OWN margin by being so freaking tall.

I should sometime, but having already perused many of Karlovic's match stats I doubt the competition factor is as minor as you seem to think.

It's minor only in the sense that it doesn't contradict my central premise.

 

[NonP]

Yes, % of unreturned serves ARE a better gauge of serving prowess. Absolutely. I should have addressed that. I was posting on my phone, and thus I really didn't have the patience to get to everything.

I missed most of the match today and will respond after the replay, but you posted all that text on your phone?! I'm just amazed at how some people can write so much on that tiny thing. I much prefer the old-fashioned keyboard. (Needless to say I hate texting.)

 

[Fiji]

Pistol Pete.

 

[Bobby Jr]

I didn't say accuracy and angles were the same thing. I said accuracy doesn't matter when you have the margins they do. Again, not that they don't have accuracy, because they do, which is why they are in the top 32 and some other very tall ATP player isn't even heard of.

No. Watch some flipping tennis for a change. The average accuracy of Isner and his ilk is rarely anything like that of Sampras and Federer.

Those aren't necessarily advantages. They favor returners, too. Again, you're only seeing things you want to see, and noticing details that favor you, without looking at the whole picture. Or maybe you're just being intellectually dishonest.

Seriously. How deluded must you be to think that poly strings which give servers an advantage somehow makes those serves more touchable because the returner has poly strings? If they cant even get a play on the ball because the serve was improved then the strings on the returners racquet are irrelevant.

Seriously, I put money on you not even being able to hit a ball given the notions you have given here.

 

[The_18th_Slam]

No. Watch some flipping tennis for a change. The average accuracy of Isner and his ilk is rarely anything like that of Sampras and Federer.

And yet, they do a much better job holding serve than either Federer or Sampras, despite a much worse groundgame. Use your flipping head for a change.

Seriously. How deluded must you be to think that poly strings which give servers an advantage somehow makes those serves more touchable because the returner has poly strings? If they cant even get a play on the ball because the serve was improved then the strings on the returners racquet are irrelevant.

You do realize there are more ways to win points than through aces, right? That there are these things called unreturned serves? You think it was as easy getting the ball back in play without poly strings?

Seriously, I put money on you not even being able to hit a ball given the notions you have given here.

You'd lose your money. All $50 of it. And then you'd be broke.

 

[jhhachamp]

No. Watch some flipping tennis for a change. The average accuracy of Isner and his ilk is rarely anything like that of Sampras and Federer.

I've watched tons of tennis and don't agree with you that Sampras and Federer have more accurate serves than Karlovic and Isner.

 

[HoyaPride]

I like you quite a bit as a poster, but I seriously don't get how you can maintain Sampras as the best server ever (is that a joke?), when there's so much evidence against it (not that Fed is at his level or better, but others certainly are - Roddick and Goran included).

I don't consider Roddick and Ivo to be better servers than Sampras. If your definition of "better" is hitting more aces, then I suppose you have a point.

I think people forget how shaky Goran's second serve could be. Sure, he was a sight to behold when he was firing off aces, but when you watched all of his matches during a season (which I did in HS and college when I had tons of free time to watch TV), you saw him serve a slew of DFs and in some cases some rather tentative 2nd serves.

To me, Roddick simply has a faster serve than Sampras. I would not say it's a better serve. There's more to the serve than the MPH. For placement, disguise, and the heaviness of the shot, I would take Sampras. There aren't too many guys whose serves I would take over Sampras.

 

[NonP]

I also do not believe that there is pronounced difference between most great servers. Karlovic is an exception. He doesn't have a top 1000 ground game yet holds serve 92% of the time in his career. Incomprehensible. And I am convinced that his unreturned serve rate for his career would stomp the yard. Whether the margin between Karlovic and Sampras is "huge", or just "very clear", there's not much to pin on me here. At worst I was being hyperbolic, according to you. Remember that you agree with my central premise that Karlovic is a better server. The original post that you were responding to was a response to a poster that claimed that Sampras has a much better serve, and insinuated that the reason Karlovic doesn't notch more scalps over top players is because his serve pulls a disappearing act in big moments.

Which is further from the truth?

I believe that poster was referring to Pete's 2nd serve but yes, if he was indeed comparing their serves overall of course he was being ridiculous. In my book Ivo and Goran are in a league of their own and I don't think they'll ever be kicked off the top berth.

That said I'm all but certain that Ivo's unreturned-serve rate does not "stomp the yard." I don't know if you cared to follow the first link I provided in that GSOAT post of mine but there's actually not much to choose between Ivo and Goran (and Arthurs, for that matter) in ace frequency, and the likes of Raonic, Philippoussis and Krajicek aren't that far behind.

Obviously slice serve aces's method (as he himself acknowledges) isn't foolproof as it excludes all the second-serve aces that these guys hit throughout their careers, but having perused many %s of unreturned serves I can say that high rates of these freebies tend to correlate with high ace frequency. The notable exceptions are Sampras and Roddick because, well, we know it from the stats, but I reckon mainly because they both had a fantastic 2nd serve and also due to Pete's unusually high topspin component and Andy's sheer power (and probably a topspin rate not far below Pete's). And though Isner's ace frequency is surprisingly low (relatively speaking, of course) and I have only a handful of his unreturned-serve stats (which again are themselves surprisingly underwhelming), I suspect his actual stats are better since John IMO has the best 2nd serve and nastiest kick since Sampras (yes, even over Roddick).

Of course it's possible that Ivo is yet another exception, but given his relatively mediocre career % of 2nd-serve points won and I suspect his fairly average spin rate (which of course he makes up for with crazy angles due to his height) I rather doubt it. Since unreturned-serve stats are hard to come by we can only guess, and at the risk of sounding presumptuous I say my guess is more educated than most.

It isn't overlooked. It just doesn't offset everything else, especially when all the existing evidence indicates that Karlovic holds serve AGAINST THE TOP TEN at about as high/higher rate than Sampras did against the entire field, with a worse game to back up the serve.

Again you should be careful about any direct cross-era comparisons, because of all the factors I've mentioned. Remember I said the hold %s had probably gone up still since the '00s? Well, guess what, I did some stats Saturday night (well into it actually, which is why I missed most of the Wimby final live--the things I do for tennis...) and it turns out my guess was correct: the top 10 servers of this decade (2011-14) have won 87.9% of their service games so far, up from 87.1% in the '00s and 86.0% in the 90s. (Granted these averages aren't without flaws because as you know many of the ATP stats are flawed to begin with and sometimes players with an inadequate number of matches are omitted, but the top 10 %s of service/return games won should be more or less correct.) That's almost a full 2% improvement since the '90s.

So let's say Pete and a Raonic were to switch places, and assuming their service games would accordingly improve/suffer (and I see no reason to assume otherwise, give or take minor changes) we'd then be looking at 91% of career service games won for Pete and 89% for Milos rather than the other way around as is now. Changes the picture a little bit now, doesn't it? Now of course Karlovic has been at it much longer than Raonic and this swap effect in Pete and Ivo's case wouldn't be as significant, but since a mere 2% difference can mean a lot in these comparisons it's something to keep in mind. And remember, this is in addition to the competition factor. I'm not sure where you got Ivo's career hold % against the top 10 but even if we grant that the return quality of his opponents doesn't matter much for him you still can't take all these numbers at face value.

I understand all this is fairly obscure stuff and don't expect everyone to be aware of it. What I do find frustrating though is how it fails to register in people's heads no matter how many times I point this out. The ATP deserves its share of blame for not keeping better track of its records but the fault really lies with us fans who don't care to hear what they don't want to hear.

Thanks, the sentiment is appreciated. Yeah, I also got testy with you, as I was taken aback by your disparaging comment. Which is also more of a compliment than an insult, as that's clearly not your M.O.

Thanks from me, too. Alas I can be quite a crank on certain days. I've gotten better with age but still have much room for improvement.

I wasn't intending to cherry-pick, though. Their return statistics and % of tiebreaks won relative to how many they are expected to win, provide very practical evidence that Federer is a better tennis player. The tiebreaks are a small proof that the % of total games won is further in Federers favour than the return/serve games won alone would have you think. This wasn't meant to be some conclusive proof. Just another link in the chain.

I'd rather not get into yet another Fed-Pete comparison here but let me just say nobody denies that Fed is better day in and day out, but there's plenty of evidence that Pete was just as good if not even better against the top dogs on the most important stages. FWIW if pushed I'd rank Fed the best of the Open era myself (all time it's Laver for me, though sometimes I'm tempted to go with Gonzales and also maintain that Navratilova is the greatest of them all man or woman), but I really don't put much thought into the GOAT debate these days. The only GOAT ranking that commands my attention now is service-related, and I don't say that because I have the slightest personal stake in the matter.

I would dispute that. His slice is a passable shot. He doesn't carve it that well and it has a tendency to sit up. Nadal isn't known for having a good slice, either.

His rally forehand is not a good shot. He can slap it well when he gets a hold of one, often at triple digits. But it's not a tour-average forehand. His serve is his ONLY weapon of note. Without it he wouldn't crack the top 1000. Again, I suspect we are engaging in semantics here. In a literal sense, of course Karlovic is not just a serve. He needs to win points on his opponents serve SOME of time, hit SOME knifing volleys and put away SOME forehands. But he doesn't do any of that with any regularity. So I believe I am justified in my assessment.

But rallying isn't Ivo's metier and he knows it, which is why he takes those big cuts to begin with. Sure, he might miss more often than not, but is the strategy and mindset really all that different from how they (including a certain someone that did pretty well at Wimbledon) used to play on grass and sometimes on HCs back in those days? I think not.

 

[NonP]

I conceded a 2% edge to be diplomatic. There's no telling how much/little the numbers are fudged. What we DO know is, Karlovic DID win a higher % of total service points, a higher percentage of first serve points, and landed SIGNIFICANTLY more first serves in. Like, super-duper significantly more. So yes, in the context of this discussion, Karlovic winning a slightly lower percentage of second serve points (lets say 1.5-2% less), with everything else he has in his favour, is not enough to sway me.

We also know that, with the "correct" stats, from 2000-2002, Sampras won around 53% (I might be rounding up, mind you) of his second serve points. Yes, his game had deteriorated by that point, but....well...do we wanna go down that road? Pete's ground game TODAY would be better than Ivo's. So we can't use that.

A couple of things here. First off, I think you're underestimating just how much the 1st-serve % has improved since the '90s. Again it's no accident that the vast majority of the top 100 players in career 1st-serve % are from after the Sampras era. Just to drive the point home this year's Wimbledon SFs saw 76% on 1st serves from Fed and 74% from Murray, and 76% from Djoko and an admittedly subpar 51% from Gasquet. Take this from someone that has pored over gazillions of match stats: you'll be hard-pressed to find any Slam edition prior to the '00s (except maybe at RG where the clay-courters would mostly roll in their 1st serves) with such consistently high 1st-serve %s.

Now perhaps this doesn't affect Ivo much because the guy's a near 7-footer after all, but given Roddick's career average of 65% on 1st serves and Pete's own near 60% (I know the ATP's 59% for Pete is more or less correct because I've seen articles and broadcasts that say he's a career 60% server) let's say he'd serve around 63% in this era. That's about 3% lower than Ivo's 66% and 5% below Isner's 68%. Is this still an advantage to those guys? Sure (though it could be a result of Pete going for more on his serve, as evidenced by his higher # of DFs), but I'd say no longer significant, and certainly not as decisive as it might seem at first glance.

That rise in 1st-serve % could produce some remarkable results. For example if we had 2001 Wimbledon in this era we might be looking at an even higher tournament-record # of aces by Goran on his dream run. (Perhaps not quite as high per match as Ivo's ridiculous 41.25 aces this year, but let's not forget Goran played fewer sets in his earlier rounds, and of course went deeper.)

And second, let's not forget another swap effect, this time on 2nd serves. Alas the ATP's butchering of 2nd-serve stats (this one I know is significant, because slice serve ace's corrections show an clear uptick for Pete and everyone else from the '90s) makes direct comparisons even more useless, but let say on the conservative side that Pete's actual career % of 2nd-serve points won is 54.5% and that Ivo's own % in similar conditions would drop only by 1% from 53% to 52%. That's still a near 3% difference, and given how many matches are decided on 2nd-serve points I'd say that matters at least as much as Ivo's edge in 1st-serve % (after all even the best servers don't win 100% of their 1st-serve points).

Of course all that still doesn't address your main contention about these guys' ground games, which brings us to my third and last point. Like I said and as you yourself admit, a player's ground game has a much more significant impact on his return game than on his service game. Where we diverge is in the degree of this discrepancy and I say you're grossly underestimating it, at least as much as you probably think I'm overestimating it.

So let's take a closer look. I'm guessing you're already aware of these numbers and since one can easily look these up anyway I'm going to take the best and worst years of the following players and examine their %s of service/return games won. For fairness' sake I'm looking only at the period after the player's breakout year and excluding any years with extenuating circumstances, so Fed's pre-2003 years or Ivo's poor 2013 due to viral meningitis is out. You're welcome to do your own research if you think I'm cherry-picking.

Let's start with Sampras:

1994 - 88%/29.4%
2001 - 86%/17.3%

Federer:

2006 - 90%/31.6%
2013 - 87%/25.9%

Agassi:

1995 - 87%/36%
2004 - 88%/26%

Hewitt:

2001 - 83%/33%
2007 - 83%/27%

Roddick:

2003 - 91%/21%
2010 - 91%/18%

And now Karlovic:

2008 - 90%/12%
2011 - 90%/9%

(I know Ivo had a career-high 94% of service games won and also a career-high No. 22 year-end ranking in '07, but these numbers need to be put in context because he played 3 Challenger events that year as opposed to zero in '08 and managed to get past the 2nd round of any major/Masters only once at Madrid.)

As you can see players' % of service games won doesn't really change a whole lot throughout their career, and as I often point out it's really in their return game that one can see their steep decline. Now Roddick's and Karlovic's 3% drop might not sound like much, but that number is misleading because 1) these guys (especially Ivo) win such a low % of their return games compared to the best returners that there's not much room for a downfall and 2) even just a 3% drop in return games won is a much bigger concern than its equal counterpart in service games won (after all a drop from 90% to 87% in the latter is still only about a 3% decline, whereas a 3% decrease from 21% to 18% in return games is more like a free fall by 14%).

And notice how a weak/mediocre server like Hewitt won only 83% of his service games even in his absolute best year, as opposed to Agassi's 87-88% high. So could it be the case that guys like Agassi had an underrated serve and even a monster server like Karlovic has relied on the rest of his game more than people think? I say yes, and at the very least the stats show that the quality of a player's service game remains remarkably steady compared to that of his return game. Let me put it this way: the difference between top 100-200 and top 30-40 lies in the serve, while the difference between top 30-40 and top 10 lies in the return. Obviously this is a gross simplification, but an instructive one nonetheless for these serve comparisons.

I'm aware of the history of how service speeds were registered, and of the spike in '99-'00. Keep in mind, I meant that Pete's first serve speeds were that high if we use the updated method. We can assume this by tracking his speeds in the early 2000's; they WERE that high. He routinely averaged in the mid-120s on the first serve, with the same frame he used in his prime. More importantly, his serve was much pacier than Feds, which is what I was responding to.

Ah, but the radar readings even from the early '00s aren't quite comparable to the current ones. Let me demonstrate. Here are Fed's average serve speeds at this year's Wimbledon courtesy of its website (1st/2nd, plus the fastest in parentheses):

R128 - 115/99 mph (127 mph)
R64 - 120/102 mph (126 mph)
R32 - 118/100 mph (127 mph)
R16 - 117/100 mph (128 mph)
QF - 117/101 mph (128 mph)
SF - 118/100 mph (126 mph)
F - 117/102 mph (126 mph)

Pretty consistent. And here are Pete's from 2001 Wimbledon (clickable archived links provided):

R128 - 118/110 mph (129 mph)
R64 - 116/103 mph (132 mph)
R32 - 116/104 mph (130 mph)
R16 - 121/110 mph (136 mph)

Again quite consistent, though there's more variance in 2nd-serve and fastest speeds. And as you can see the last link has got Fed serving 114/97 mph topping out at 123 mph. So if these readings are to be believed and compared, we should be able to see Fed serving considerably faster than his younger self in '01 (as is commonly believed) and nearly as fast as Pete at the same event.

But does that pass the eye test? Let's see:

(2001 showdown between the two)

(this year's SF between Fed and Murray - couldn't find good clips of the final)

Now I don't know about you, but my eyes clearly tell me that Pete was serving quite faster than Fed this year and there's not a whole lot to separate the 2001 and 2015 Fed. And don't forget, in his 2nd exo against Fed, already well into his retirement, Pete had half of his 20 serves clocking in at 130 mph or more, 137 mph tops, while Fed's first serves were between 113-129 mph, just about what you expect.

So if we are to take these numbers at face value, we're expected to believe that a retired Sampras was serving faster than his younger self! Or maybe the radar guns are producing higher readings these days. Which explanation is more plausible? I'll let you make the call.

All that said I do think Fed has beefed up his serve a bit around '07-'08. Here are his aces per match since '04:

2004 - 7.0
2005 - 7.0
2006 - 6.8
2007 - 7.8
2008 - 8.6
2009 - 9.0
2010 - 8.4
2011 - 6.6
2012 - 8.3
2013 - 6.4
2014 - 8.0

(So far this year he's been averaging 8.2 aces per match, as opposed to 7.5 in '01.)

Quite a leap after '06-'07, peaking in '09 (BTW this is also the year with his highest % of service games ever won on clay), though the outliers in '11 and '13 strongly suggest that he's been trying get more free points on his serve because he has to, and not always with success. Still I think it fair to say that whatever additional pace he's gained since '07 is not as obvious as it's made out to be.

 

[jhhachamp]

Again you should be careful about any direct cross-era comparisons, because of all the factors I've mentioned. Remember I said the hold %s had probably gone up still since the '00s? Well, guess what, I did some stats Saturday night (well into it actually, which is why I missed most of the Wimby final live--the things I do for tennis...) and it turns out my guess was correct: the top 10 servers of this decade (2011-14) have won 87.9% of their service games so far, up from 87.1% in the '00s and 86.0% in the 90s. (Granted these averages aren't without flaws because as you know many of the ATP stats are flawed to begin with and sometimes players with an inadequate number of matches are omitted, but the top 10 %s of service/return games won should be more or less correct.) That's almost a full 2% improvement since the '90s.

So let's say Pete and a Raonic were to switch places, and assuming their service games would accordingly improve/suffer (and I see no reason to assume otherwise, give or take minor changes) we'd then be looking at 91% of career service games won for Pete and 89% for Milos rather than the other way around as is now. Changes the picture a little bit now, doesn't it? Now of course Karlovic has been at it much longer than Raonic and this swap effect in Pete and Ivo's case wouldn't be as significant, but since a mere 2% difference can mean a lot in these comparisons it's something to keep in mind. And remember, this is in addition to the competition factor. I'm not sure where you got Ivo's career hold % against the top 10 but even if we grant that the return quality of his opponents doesn't matter much for him you still can't take all these numbers at face value.

I think looking at only the top 10 servers in each decade is too small of a sample size. Your analysis makes the assumption that the average top 10 serve stays constant and doesn't consider the possibility the best servers now could simply be better than the best servers in the 90s. Having a major outlier (which I believe Karlovic is) is going to skew the data too much when looking only a sample of 10 players.

If you look at the top 100 players in each decade and compare the hold percentages, then I think you could draw better conclusions. I would not be surprised to see that it still shows that the hold rates have increased, but I don't think that it will be 2%.

 

[AngryBirds]

A couple of things here. First off, I think you're underestimating just how much the 1st-serve % has improved since the '90s. Again it's no accident that the vast majority of the top 100 players in career 1st-serve % are from after the Sampras era. Just to drive the point home this year's Wimbledon SFs saw 76% on 1st serves from Fed and 74% from Murray, and 76% from Djoko and an admittedly subpar 51% from Gasquet. Take this from someone that has pored over gazillions of match stats: you'll be hard-pressed to find any Slam edition prior to the '00s (except maybe at RG where the clay-courters would mostly roll in their 1st serves) with such consistently high 1st-serve %s.

Now perhaps this doesn't affect Ivo much because the guy's a near 7-footer after all, but given Roddick's career average of 65% on 1st serves and Pete's own near 60% (I know the ATP's 59% for Pete is more or less correct because I've seen articles and broadcasts that say he's a career 60% server) let's say he'd serve around 63% in this era. That's about 3% lower than Ivo's 66% and 5% below Isner's 68%. Is this still an advantage to those guys? Sure (though it could be a result of Pete going for more on his serve, as evidenced by his higher # of DFs), but I'd say no longer significant, and certainly not as decisive as it might seem at first glance.

That rise in 1st-serve % could produce some remarkable results. For example if we had 2001 Wimbledon in this era we might be looking at an even higher tournament-record # of aces by Goran on his dream run. (Perhaps not quite as high per match as Ivo's ridiculous 41.25 aces this year, but let's not forget Goran played fewer sets in his earlier rounds, and of course went deeper.)

And second, let's not forget another swap effect, this time on 2nd serves. Alas the ATP's butchering of 2nd-serve stats (this one I know is significant, because slice serve ace's corrections show an clear uptick for Pete and everyone else from the '90s) makes direct comparisons even more useless, but let say on the conservative side that Pete's actual career % of 2nd-serve points won is 54.5% and that Ivo's own % in similar conditions would drop only by 1% from 53% to 52%. That's still a near 3% difference, and given how many matches are decided on 2nd-serve points I'd say that matters at least as much as Ivo's edge in 1st-serve % (after all even the best servers don't win 100% of their 1st-serve points).

Of course all that still doesn't address your main contention about these guys' ground games, which brings us to my third and last point. Like I said and as you yourself admit, a player's ground game has a much more significant impact on his return game than on his service game. Where we diverge is in the degree of this discrepancy and I say you're grossly underestimating it, at least as much as you probably think I'm overestimating it.

So let's take a closer look. I'm guessing you're already aware of these numbers and since one can easily look these up anyway I'm going to take the best and worst years of the following players and examine their %s of service/return games won. For fairness' sake I'm looking only at the period after the player's breakout year and excluding any years with extenuating circumstances, so Fed's pre-2003 years or Ivo's poor 2013 due to viral meningitis is out. You're welcome to do your own research if you think I'm cherry-picking.

Let's start with Sampras:

1994 - 88%/29.4%
2001 - 86%/17.3%

Federer:

2006 - 90%/31.6%
2013 - 87%/25.9%

Agassi:

1995 - 87%/36%
2004 - 88%/26%

Hewitt:

2001 - 83%/33%
2007 - 83%/27%

Roddick:

2003 - 91%/21%
2010 - 91%/18%

And now Karlovic:

2008 - 90%/12%
2011 - 90%/9%

(I know Ivo had a career-high 94% of service games won and also a career-high No. 22 year-end ranking in '07, but these numbers need to be put in context because he played 3 Challenger events that year as opposed to zero in '08 and managed to get past the 2nd round of any major/Masters only once at Madrid.)

As you can see players' % of service games won doesn't really change a whole lot throughout their career, and as I often point out it's really in their return game that one can see their steep decline. Now Roddick's and Karlovic's 3% drop might not sound like much, but that number is misleading because 1) these guys (especially Ivo) win such a low % of their return games compared to the best returners that there's not much room for a downfall and 2) even just a 3% drop in return games won is a much bigger concern than its equal counterpart in service games won (after all a drop from 90% to 87% in the latter is still only about a 3% decline, whereas a 3% decrease from 21% to 18% in return games is more like a free fall by 14%).

And notice how a weak/mediocre server like Hewitt won only 83% of his service games even in his absolute best year, as opposed to Agassi's 87-88% high. So could it be the case that guys like Agassi had an underrated serve and even a monster server like Karlovic has relied on the rest of his game more than people think? I say yes, and at the very least the stats show that the quality of a player's service game remains remarkably steady compared to that of his return game. Let me put it this way: the difference between top 100-200 and top 30-40 lies in the serve, while the difference between top 30-40 and top 10 lies in the return. Obviously this is a gross simplification, but an instructive one nonetheless for these serve comparisons.



Ah, but the radar readings even from the early '00s aren't quite comparable to the current ones. Let me demonstrate. Here are Fed's average serve speeds at this year's Wimbledon courtesy of its website (1st/2nd, plus the fastest in parentheses):

R128 - 115/99 mph (127 mph)
R64 - 120/102 mph (126 mph)
R32 - 118/100 mph (127 mph)
R16 - 117/100 mph (128 mph)
QF - 117/101 mph (128 mph)
SF - 118/100 mph (126 mph)
F - 117/102 mph (126 mph)

Pretty consistent. And here are Pete's from 2001 Wimbledon (clickable archived links provided):

R128 - 118/110 mph (129 mph)
R64 - 116/103 mph (132 mph)
R32 - 116/104 mph (130 mph)
R16 - 121/110 mph (136 mph)

Again quite consistent, though there's more variance in 2nd-serve and fastest speeds. And as you can see the last link has got Fed serving 114/97 mph topping out at 123 mph. So if these readings are to be believed and compared, we should be able to see Fed serving considerably faster than his younger self in '01 (as is commonly believed) and nearly as fast as Pete at the same event.

But does that pass the eye test? Let's see:

(2001 showdown between the two)

(this year's SF between Fed and Murray - couldn't find good clips of the final)

Now I don't know about you, but my eyes clearly tell me that Pete was serving quite faster than Fed this year and there's not a whole lot to separate the 2001 and 2015 Fed. And don't forget, in his 2nd exo against Fed, already well into his retirement, Pete had half of his 20 serves clocking in at 130 mph or more, 137 mph tops, while Fed's first serves were between 113-129 mph, just about what you expect.

So if we are to take these numbers at face value, we're expected to believe that a retired Sampras was serving faster than his younger self! Or maybe the radar guns are producing higher readings these days. Which explanation is more plausible? I'll let you make the call.

All that said I do think Fed has beefed up his serve a bit around '07-'08. Here are his aces per match since '04:

2004 - 7.0
2005 - 7.0
2006 - 6.8
2007 - 7.8
2008 - 8.6
2009 - 9.0
2010 - 8.4
2011 - 6.6
2012 - 8.3
2013 - 6.4
2014 - 8.0

(So far this year he's been averaging 8.2 aces per match, as opposed to 7.5 in '01.)

Quite a leap after '06-'07, peaking in '09 (BTW this is also the year with his highest % of service games ever won on clay), though the outliers in '11 and '13 strongly suggest that he's been trying get more free points on his serve because he has to, and not always with success. Still I think it fair to say that whatever additional pace he's gained since '07 is not as obvious as it's made out to be.

Amazing stats. Let's not forget that Sampras used 1980s racquet to generate that pace. I'd like to see also the serve speed stats in their 2001 match when they both used the same racquet.

 

[NonP]

I think looking at only the top 10 servers in each decade is too small of a sample size. Your analysis makes the assumption that the average top 10 serve stays constant and doesn't consider the possibility the best servers now could simply be better than the best servers in the 90s. Having a major outlier (which I believe Karlovic is) is going to skew the data too much when looking only a sample of 10 players.

If you look at the top 100 players in each decade and compare the hold percentages, then I think you could draw better conclusions. I would not be surprised to see that it still shows that the hold rates have increased, but I don't think that it will be 2%.

I admit I haven't crunched the numbers for bigger samples yet, but I doubt it'd make much of a difference. Again the 1st-serve % is a big factor. There's no question players are getting more 1st serves in than in the '90s and most likely ever. I've taken a cursory look at outside the top 10 and didn't see a whole lot that didn't fit the trend, which is why I didn't bother to proceed further.

Also I don't think the fact that Karlovic is an outlier is very relevant here, because pretty much all these guys are outliers. We're talking about the best of the best.

And let me stress that the '10s stats are obviously incomplete. We're not even halfway in yet, so the average % could go up or down. My guess is that it might rise slightly still but probably not by more than a full 1% as we saw between the '90s and the '10s. The early '90s were an outlier compared to the rest of the time frames.

Amazing stats. Let's not forget that Sampras used 1980s racquet to generate that pace. I'd like to see also the serve speed stats in their 2001 match when they both used the same racquet.

To be clear where the new sticks help is in average serve speeds, again due to getting more 1st serves in and heavier spin on 2nd. The top-end speeds shouldn't change much regardless of racquets. Maybe a gain of 2-3 mph, but not more than that. If anyone tells you these pros would gain something like 10 mph or more with any new racquet he just doesn't know what he's talking about.

 

[jhhachamp]

Also I don't think the fact that Karlovic is an outlier is very relevant here, because pretty much all these guys are outliers. We're talking about the best of the best.

It depends on how you define outlier. Most players have a near average serve and as you get further away from that, better or worse, there are fewer and few people. So, the other great servers are simply at the far end of the range, but they are not outliers. I see Karlovic as an outlier because at the far end of the spectrum, there is a significant gap between him and the second best server (my opinion). The gap from #1 to #2 is much greater than the gap from #2 to #3, #3 to #4, #4 to #5, and so forth.

 

[Chanwan]

I don't consider Roddick and Ivo to be better servers than Sampras. If your definition of "better" is hitting more aces, then I suppose you have a point.

I think people forget how shaky Goran's second serve could be. Sure, he was a sight to behold when he was firing off aces, but when you watched all of his matches during a season (which I did in HS and college when I had tons of free time to watch TV), you saw him serve a slew of DFs and in some cases some rather tentative 2nd serves.

To me, Roddick simply has a faster serve than Sampras. I would not say it's a better serve. There's more to the serve than the MPH. For placement, disguise, and the heaviness of the shot, I would take Sampras. There aren't too many guys whose serves I would take over Sampras.

I admittedly was more of a casual observer in the 90's than I am these years and I never saw anything close to a full season of Goran's matches. Also, my memory is simply not strong enough to rely on in this regard. So what I do is I base my analysis and opinions on stats and videos.

What the stats do tell me is that Ivo is in a league of his own and both Goran and Roddick are ahead of Sampras (but that's more debatable, Ivo being far ahead is not). What you need to remember is that Sampras was a vastly better player in everything else than the serve than those players.
Yet Roddick, Goran and especially Ivo posts equal or better stats on service games and all that. How's that if they don't have a better serve combo (2nd serve included)?

There aren't stats on serves + unreturnables, but I'm quite sure all of them would come out on top vs. Sampras in that. And that Ivo again would be in a league of his own.
As for DF's, Sampras hit a far amount as well.

There's an argument for Sampras in terms of clutch serving and succeeding vs. players like Agassi (best returner). But isn't that as much about him better a strong player in general, so he could win with his serve and hang with Agassi in the rallies, whereas most other big servers would get punished by Agassi once he got them in a rally?

 

[Laurie]

I admittedly was more of a casual observer in the 90's than I am these years and I never saw anything close to a full season of Goran's matches. Also, my memory is simply not strong enough to rely on in this regard. So what I do is I base my analysis and opinions on stats and videos.

What the stats do tell me is that Ivo is in a league of his own and both Goran and Roddick are ahead of Sampras (but that's more debatable, Ivo being far ahead is not). What you need to remember is that Sampras was a vastly better player in everything else than the serve than those players.
Yet Roddick, Goran and especially Ivo posts equal or better stats on service games and all that. How's that if they don't have a better serve combo (2nd serve included)?

There aren't stats on serves + unreturnables, but I'm quite sure all of them would come out on top vs. Sampras in that. And that Ivo again would be in a league of his own.
As for DF's, Sampras hit a far amount as well.

There's an argument for Sampras in terms of clutch serving and succeeding vs. players like Agassi (best returner). But isn't that as much about him better a strong player in general, so he could win with his serve and hang with Agassi in the rallies, whereas most other big servers would get punished by Agassi once he got them in a rally?

I pointed out a couple of years ago that Agassi has a winning record against almost every single attacking / serve volley player of his era:

v Becker: 9:5
v Rafter: 10:5
v Edberg: 6:3
v Krajicek 4:3
v Rusedski 9:2
v Phillippoussis 6:2
v Wayne Arthurs 2:0
v Stich 6:0
v Henman 2:2
v Todd Woodbridge 6:0
v Mark Woodforde 4:2
v Ivanisevic 4:3

That is an incredible record. There is only one attacking player he has a losing record against, guess who that is?

 

[jhhachamp]

I pointed out a couple of years ago that Agassi has a winning record against almost every single attacking / serve volley player of his era:

v Becker: 9:5
v Rafter: 10:5
v Edberg: 6:3
v Krajicek 4:3
v Rusedski 9:2
v Phillippoussis 6:2
v Wayne Arthurs 2:0
v Stich 6:0
v Henman 2:2
v Todd Woodbridge 6:0
v Mark Woodforde 4:2
v Ivanisevic 4:3

That is an incredible record. There is only one attacking player he has a losing record against, guess who that is?

Interesting data, but I'm not really sure how much it shows. Agassi was one of the greatest players of his era, so he likely had a winning record against almost every player, regardless of whether or not they were an attacking/serve and volley player or not.

Nice racket by the way, I also use Estusa!

 

[Laurie]

Interesting data, but I'm not really sure how much it shows. Agassi was one of the greatest players of his era, so he likely had a winning record against almost every player, regardless of whether or not they were an attacking/serve and volley player or not.

Nice racket by the way, I also use Estusa!

Quite the opposite, it shows a lot. Remember, the ideologues on this forum claim that surfaces were so fast then, counterpunchers had absolutely no chance against attacking players.

I would like to get ahold of another Estusa racquet but hard to come by these days. I have been advised to look at Yonex as a similar alternative. Those racquets are incredible to use but you got to have strength to use them.

 

[Bobby Jr]

...That rise in 1st-serve % could produce some remarkable results. For example if we had 2001 Wimbledon in this era we might be looking at an even higher tournament-record # of aces by Goran on his dream run. (Perhaps not quite as high per match as Ivo's ridiculous 41.25 aces this year, but let's not forget Goran played fewer sets in his earlier rounds, and of course went deeper.)

You touch on a very important point here - guys like Sampras and Federer played much, much tougher opponents on average, and therefore much better returners, than the likes of Karlovic and Isner because Fed/Sampras went deeper into eons more tournaments. I mean this year so far has been pretty average for Federer but it's still better than any year in Karlovic or Isner's career.

 

[jhhachamp]

Quite the opposite, it shows a lot. Remember, the ideologues on this forum claim that surfaces were so fast then, counterpunchers had absolutely no chance against attacking players.

I would like to get ahold of another Estusa racquet but hard to come by these days. I have been advised to look at Yonex as a similar alternative. Those racquets are incredible to use but you got to have strength to use them.

I think the best players are the best players for a reason and they are generally going to have winning head to head records against nearly all other players of their era, regardless of style of play. Sure, those players might be difficult for Agassi on faster surfaces, but he would also have chances to play them on slower surfaces where he would have a larger advantage.

I have 2 old Power Beam Pros that I use mostly as my 3rd and 4th rackets in case my main 2 rackets have broken strings. I hardly ever use them. I usually use my 2 Power Beam Pro Braided rackets. I actually bought 4 of the Power Beam Pro Braided rackets a long time ago from Europe right before the company went out of business. I had to wire about $800 to Europe, it was a bit sketchy. I have 2 of these rackets sitting at my parents home brand new with the packaging still on them! With how little I actually play tennis these days, all those rackets should last me a while!

 

[jhhachamp]

You touch on a very important point here - guys like Sampras and Federer played much, much tougher opponents on average, and therefore much better returners, than the likes of Karlovic and Isner because Fed/Sampras went deeper into eons more tournaments. I mean this year so far has been pretty average for Federer but it's still better than any year in Karlovic or Isner's career.

I don't think the difference in average opponent ranking is as great as you seem to think. Yes, Sampras and Federer advance further in most tournaments, but they also get easier opponents in the early rounds due to their high seed. Unless someone has some data on average opponent ranking, then we are both just guessing at this anyway. Also, I think you would agree that Sampras and Federer played a higher percentage of their matches in best of 5 format than Karlovic and Isner because they have had so much more success in the grand slams (and 5 set matches were more common in the 90s as well I believe). Despite this, Karlovic served nearly twice as many aces per match as Sampras and Federer.

 

[krosero]

Remember I said the hold %s had probably gone up still since the '00s? Well, guess what, I did some stats Saturday night (well into it actually, which is why I missed most of the Wimby final live--the things I do for tennis...) and it turns out my guess was correct: the top 10 servers of this decade (2011-14) have won 87.9% of their service games so far, up from 87.1% in the '00s and 86.0% in the 90s. (Granted these averages aren't without flaws because as you know many of the ATP stats are flawed to begin with and sometimes players with an inadequate number of matches are omitted, but the top 10 %s of service/return games won should be more or less correct.) That's almost a full 2% improvement since the '90s.

Thanks for calculating these. You've mentioned before in general terms this upward trend in service stats. I think it's quite possible that the trend began earlier and may extend back decades. I may never have enough stats to prove such a thing, but when I do come across stats from the early Open Era and before, I'm often surprised at the low hold rates.

Very good to have your stats on this, please keep them coming.

I much prefer the old-fashioned keyboard.

Ah, if my grandparents could hear you say that...

 

[NonP]

It depends on how you define outlier. Most players have a near average serve and as you get further away from that, better or worse, there are fewer and few people. So, the other great servers are simply at the far end of the range, but they are not outliers. I see Karlovic as an outlier because at the far end of the spectrum, there is a significant gap between him and the second best server (my opinion). The gap from #1 to #2 is much greater than the gap from #2 to #3, #3 to #4, #4 to #5, and so forth.

Even if we took Ivo out (and a corresponding server from the '90s) it still wouldn't make much of a difference. Remember, we're talking about service games won. It's not like Ivo was winning 98% of his service games while the other top servers were eking out just 80-85% year after year. And the %s of 1st-serve and 2nd-serve points won wouldn't tell a whole other story, either: Ivo does stand at the top with regard to the former stat, but just 1% above Raonic's 82% career average (it's hard to know exactly what Goran's % is due to the ATP's butchered '90s stats), and as I've pointed out Ivo's career 2nd-serve % is actually a good but not stellar 53%.

A single stat might not show just how much impact a player's ground game has on his return game as opposed to his service game, but as I've shown above that becomes clearer once one puts all the stats together and uses a little common sense. I really can't emphasize this point enough: the difference between top 100-200 and top 30-40 lies in the serve, while the difference between top 30-40 and top 10 lies in the return. If all those blocks of text and compiled stats I just provided don't convince you there's really not much else I can do at this point.

There aren't stats on serves + unreturnables, but I'm quite sure all of them would come out on top vs. Sampras in that. And that Ivo again would be in a league of his own.

There's no way to know for sure because %s of unreturned serves are hard to come by, but like I said the ones I do have don't really show such a clear gap between Ivo and the other monster servers. And of course this isn't a fail-safe comparison in and of itself, because an unreturned serve has as much to do with the returner as the server.

Let me give you a striking example. Here are the two giants going at each other, on grass:

2013 Newport QF - Isner def Karlovic 7-6, 7-6
53% of Isner's serves were unreturned
45% of Karlovic's serves were unreturned

Looks pretty unremarkable, as both of these guys ain't the best returner by a long shot and routinely serve over 40%, even 50% of freebies especially on the fastest surface.

But then Moose once observed this eye-popping factoid:

They had a pretty amazing stat after the first set:

34% of all Isner's service points were unreturned serves
71% of all Lllodra's service points were unreturned!

those Isner's numbers are some of the worst returning I've ever heard of(ATP only tracks aces, but total number of unreturned serves are a more important stat imo)

and people here really think he(or Karlovic) could have been Wimbledon champions in the 90s? without ever breaking serve? Guys like Sampras & Goran had a way higher break % on that grass than those guys have on any surface. Big serve doesn't automatically mean great fast court player(Rusedski was also someone who struggled breaking serve on grass, which partly explains why his Wimbledon record was so much worse than say Goran's)

Isner & Karlovic have more in common with Wayne Arthurs than Sampras, Goran, Krajicek, etc.

Now I've yet to come across a single match where a player had over 70% of his serves unreturned throughout (the closest one I know of is Ivo serving at least 66.7% of freebies against Bracciali in the 1st round of '05 Wimbledon, though he went on to lose the match)--even 60% is pretty rare for the very best servers--so I very much doubt that Llodra was able to keep up his insane % for the rest of the match. Still this match (or the Bracciali match, for that matter) shows just how a Karlovic's or Isner's stats sometimes need to be put in context. TheFifthSet has insisted on Ivo's service stats being largely unaffected by his opponents, and while that's largely true even he's not immune to quality returning himself as his record against a Murray or Davydenko demonstrates. And there's no question that Pete, Roddick and Fed faced stronger competition more often or had deeper runs at the best events. I do think the competition factor is often overstated but still it's something to keep in mind when we compare the service stats of a Karlovic vs. a Sampras.

As for DF's, Sampras hit a far amount as well.

He sure did, but that's not necessarily a bad thing as is commonly assumed because it shows that he was very aggressive on 2nd serves (a la Laver, under the tutelage of Pete's childhood coach Fischer). In fact I've long maintained that today's players play it too safe on their 2nd serves and that guys like Karlovic or Murray should if anything be double-faulting more.

There's an argument for Sampras in terms of clutch serving and succeeding vs. players like Agassi (best returner). But isn't that as much about him better a strong player in general, so he could win with his serve and hang with Agassi in the rallies, whereas most other big servers would get punished by Agassi once he got them in a rally?

There was once a long discussion on the GSOAT thread that compared Agassi's and Chang's records against the entire field vs. big servers/S&Vers, as these two were pretty much neck and neck in % of return games won, and the conclusion was that Agassi comes out comfortably ahead. Not to say their overall games didn't play a role, but if you break serve with roughly equal frequency but do better against a certain group/cohort that's quite telling, no?

In short Agassi was basically kryptonite against the attacking guys, because he could put you under so much pressure especially on your 2nd serves. You're talking about a guy that often had his opponents struggling to win 60%, 70%s, sometimes even 80% (as Rusedski would tell you) of their 2nd-serve points. Heck, even Pete said after the '95 AO final that he had to go for so much more on his 2nd serve 'cause Dre was returning too good. And I think zagor once put it well when he said Agassi was making his opponents look like WTA players at '95 Wimbledon.

I wouldn't go so far as to say Agassi is hands down the best returner ever, as Connors is before my time and Djoko and Murray ain't half bad themselves. But I can't imagine anyone returning better than Agassi did in much of '95. Some of his highlights from that year will make you literally chuckle, he really was that crazy good.

You touch on a very important point here - guys like Sampras and Federer played much, much tougher opponents on average, and therefore much better returners, than the likes of Karlovic and Isner because Fed/Sampras went deeper into eons more tournaments. I mean this year so far has been pretty average for Federer but it's still better than any year in Karlovic or Isner's career.

Yes, and those stats provided by slice serve ace on Pete vs. Roddick are quite telling. Again I do think this competition factor can be exaggerated because Karlovic/fill-in-the-blank is Karlovic after all, but not to the extent where it becomes completely irrelevant.

Thanks for calculating these. You've mentioned before in general terms this upward trend in service stats. I think it's quite possible that the trend began earlier and may extend back decades. I may never have enough stats to prove such a thing, but when I do come across stats from the early Open Era and before, I'm often surprised at the low hold rates.

Very good to have your stats on this, please keep them coming.

Thanks, krosero. Unfortunately tennis' poor recordkeeping makes it virtually impossible to tell, but we're valiantly filling in the gaps!

Ah, if my grandparents could hear you say that...

Hook me up with them and I'll gladly tell them myself. Seriously, I think people today lose sight of the importance of genuine (preferably face-to-face) interaction. I generally avoid texting and always try to use email where possible, and if we can talk in person, even better.

 

[Bobby Jr]

I don't think the difference in average opponent ranking is as great as you seem to think. Yes, Sampras and Federer advance further in most tournaments, but they also get easier opponents in the early rounds due to their high seed. Unless someone has some data on average opponent ranking, then we are both just guessing at this anyway.

Most of the time Isner and Karlovic would be seed also meaning the pool of non-seeded players and the chance for drawing them early on is exactly the same as it is for the top seed. It's the 3rd or 4th round where you could argue a higher seed gets a lower rated opponent, not the first round.

Average opponent (ranking) data is one thing but the fact the top guys get further in tournaments indicates their opponents are more likely to be peaking that week regardless of their rank - and therefore more likely to be returning well which hinders a server's stats somewhat.

 

[jhhachamp]

Even if we took Ivo out (and a corresponding server from the '90s) it still wouldn't make much of a difference. Remember, we're talking about service games won. It's not like Ivo was winning 98% of his service games while the other top servers were eking out just 80-85% year after year. And the %s of 1st-serve and 2nd-serve points won wouldn't tell a whole other story, either: Ivo does stand at the top with regard to the former stat, but just 1% above Raonic's 82% career average (it's hard to know exactly what Goran's % is due to the ATP's butchered '90s stats), and as I've pointed out Ivo's career 2nd-serve % is actually a good but not stellar 53%.

Karlovic is an outlier in terms of the serve alone, not in terms of the service game in total. If Karlovic had a ground game as good as Federer or Sampras, I believe he would hold close to 98% of the time.

 

[jhhachamp]

Most of the time Isner and Karlovic would be seed also meaning the pool of non-seeded players and the chance for drawing them early on is exactly the same as it is for the top seed. It's the 3rd or 4th round where you could argue a higher seed gets a lower rated opponent, not the first round.

Yes, you are correct that a top seed has the advantage over a lower seed in the 3rd/4th rounds in grand slams. That applies to Isner. Karlovic has spent a significant portion of his career outside the top 32, and often significantly lower. I took a quick glance through his ranking history, and I would estimate that he has been ranked outside the top 32 at least 2/3 of his career. The degree of difficulty of early round matches at smaller tournaments would also be more difficult as well.

Average opponent (ranking) data is one thing but the fact the top guys get further in tournaments indicates their opponents are more likely to be peaking that week regardless of their rank - and therefore more likely to be returning well which hinders a server's stats somewhat.

This may also be true, but I think the effect would be relatively minor. Do you have any comment on the thought that Sampras played 5 set matches more often than Karlovic?

 

[coloskier]

I love Federer, but having seen both Fed and Sampras live in their primes, Fed serve isn't even close to Sampras. Sampras had so much spin on the ball along with his pace that he would consistently cause the returners to shank the ball, if they could actually get a racket on it. Even on the 2nd serve. 2nd serve aces definitely goes to Sampras. And Sampras's kick serve was unreadable and kicked at least 6" higher than Fed's.

 

[NonP]

Average opponent (ranking) data is one thing but the fact the top guys get further in tournaments indicates their opponents are more likely to be peaking that week regardless of their rank - and therefore more likely to be returning well which hinders a server's stats somewhat.

This is most likely true, as there's no question Pete, Fed and other top players try to peak for the later rounds. One extreme example is Pete himself who amazingly won a HIGHER % of service games won in his 7 Wimbledon finals (97.0%) compared to all the rounds combined (95.5%). (Fed himself won 90.4% in his own Wimby victories, as opposed to 94.1% overall.) And more predictably both Pete and Fed's %s of return games won suffered in the finals: 23.6% overall vs. 18.3% in the finals for Pete and 28.6% vs. 16.1% for Fed.

And then there's Karlovic himself. Now I think it safe to assume that Ivo has rarely if ever served better at a major than he did at this year's Wimbledon, and I don't say that just because of his ace counts. Here are his serve speeds from the 1st-4th rounds (same format as above--1st/2nd, fastest in parentheses but this time with fastest 2nd):

R128 - 128/116 mph (137/125 mph)
R64 - 128/117 mph (137/127 mph)
R32 - 126/112 mph (135/124 mph)
R16 - 126/119 mph (136/126 mph)

Now I've perused many Karlovic stats over the years and 119 mph is the highest number I can recall from him on 2nd serves. (Even with the most state-of-the-art racquets and radar I doubt Pete himself would crack 120 mph in 2nd-serve average except as a one-off.) And I know for a fact that his 2nd serves used to average around 110 mph just a couple of years ago, so this is no doubt a good showing by him. (I should add that Tsonga was battling an abdominal injury not long before Wimbledon--his serve speeds in the 3rd round against Ivo were much lower than usual, which probably explains why Ivo played it fairly safe on his 2nd serves.) The 1st-serve averages are more or less in line with his usual readings.

But of course speed isn't everything, so let's look at the usual service stats (aces/DFs, 1st serves in, 1st-serve points won, 2nd-serve points won):

R128 - 42/8, 61% (76/124), 92% (70/76), 54% (26/48)
R64 - 53/4, 72% (122/169), 87% (106/122), 57% (27/47)
R32 - 41/5, 76% (99/131), 90% (89/99), 66% (21/32)
R16 - 29/4, 69% (97/140), 76% (74/97), 49% (21/43)

As you can see Ivo except in the 1st round had very high %s of his 1st serves in, probably because he was still getting into a groove. And notice the remarkably low #s of DFs (again except in R128) despite his aggressive 2nd serves, not to mention the ridiculous ace totals.

So Ivo was clearly serving as well as perhaps ever, but then there's one match that sticks out, and not surprisingly it's the one he lost to Murray... which also happens to be the one with his fastest 2nd-serve average! And despite his near 70% serving he scored a merely decent (for him) 29 aces and 76% in 1st-serve points and under 50% on 2nds.

And I invite everyone to examine Ivo's other match stats against Murray, Davydenko, Ferrer, Fed and other elite returners and compare them against the rest. Again Ivo is such a beast there will be days when his serve would prove too much even for these guys, but then there are other days when their return would do what it's supposed to do. And top players like Pete and Fed would face their fellow top dogs more often, and again in the later rounds when they would be peaking more often than not. If you really think this would have a negligible effect even on flamethrowers like Ivo and Isner you really aren't thinking and looking hard enough.

Do you have any comment on the thought that Sampras played 5 set matches more often than Karlovic?

Are we talking about best-of-5 or 5-setters? FYI about 4.9% (48/984) of Pete's career singles matches were 5-setters, 3.6% (19/532) for Ivo, so not a big difference between them.

Besides I don't see how any of this matters. The fact that Pete played 5-setters at a higher rate can be simply attributed to his stronger opposition. And the stats I've posted so far have to do with individual games/points, not individual matches. I know how these numbers and averages can be distorted by the length of a given match so I try to avoid per-match stats where possible.

 

[zep]

With this poll result, TMF must be struggling to sleep at night.

 

[HoyaPride]

I admittedly was more of a casual observer in the 90's than I am these years and I never saw anything close to a full season of Goran's matches. Also, my memory is simply not strong enough to rely on in this regard. So what I do is I base my analysis and opinions on stats and videos.

Right.

I'm not even sure how Goran would be a better serve according to the stats. Unless, that is, you place more weight on aces.

Ivanisevic

Aces - 10,183
DFs - 3,572
1st % - 55%
1st won - 82%
2nd won - 50%
BPs faced - 3,900
BPs saved - 66%
Service games won - 86%
Service points won - 68%

Sampras

Aces - 8,858 Aces
DFs - 3,009
1st % - 59%
1st won - 81%
2nd won - 53%
BPs faced - 3,751
BPs saved - 68%
Service games won - 89%
Service points won - 69%

It could be argued that Ivanisevic was the better server because he averaged more aces per match, but he also averaged more DFs per match. He also got fewer of his 1st serves in. Goran's serve was sort of like James Blake's forehand--massive and impressive...when it was on. People only seem to remember the ace count and forget all of the matches where his greatest strength was to a large extent his undoing. There aren't many matches I can remember where Pete's serve truly let him down but there are several matches where Goran double faulted his way to oblivion.

 

[HoyaPride]

I love Federer, but having seen both Fed and Sampras live in their primes, Fed serve isn't even close to Sampras. Sampras had so much spin on the ball along with his pace that he would consistently cause the returners to shank the ball, if they could actually get a racket on it. Even on the 2nd serve. 2nd serve aces definitely goes to Sampras. And Sampras's kick serve was unreadable and kicked at least 6" higher than Fed's.

When people climb out of a 0-40 hole by blasting 5 aces/unreturnable serves, commentators often say "That was Sampras-like." There's a reason for that.

 

[AngryBirds]

When people climb out of a 0-40 hole by blasting 5 aces/unreturnable serves, commentators often say "That was Sampras-like." There's a reason for that.

I read somewhere that Sampras had a very unique shoulder muscle, which allows him to effortlessly generate more pace on his serve than just about anybody. Need confirmation on this.