Roddick vs Sampras: The Serve

[pc1]

Originally Posted by NonP

I assure you that wouldn't be true for the vast majority of players.



http://tt.tennis-warehouse.com/showt...33#post6850233

At least check out the last link if you're short on time.

As for your original question, Sampras. As good as he was, and apart from countless eye tests, Roddick is generally a notch below the likes of Pete, Goran, Ivo and Krajicek especially in % of unreturned serves. An interesting thing about this stat is that it usually corresponds with what one would expect from the mainstream consensus (for lack of a better term), which tells us that, wrong as it may be, conventional wisdom isn't entirely without merit.

Also there are other relative shorties whose serves have been nominated as among the best ever. McEnroe and Tanner come to mind, and Curren is another name that pops up in some (if admittedly cliquish) discussions.

And now's a good time to promote the greatest thread in TTW history:

http://tt.tennis-warehouse.com/showthread.php?t=306579

(Seriously, the whole thing is worth a read.)

In all seriousness it could be the best thread.

[mattennis]

I don't know what statistics say, but I wouldn't be surprised if service-related statistics favour Roddick.

Why?

Because players with absolutely great serves there have been dozens in every decade, and most of them never even reach a GS final.

And all these players with great serves (Forget, Wheaton, Stich, Krajicek, Ivanisevic, Rosset, Larsson, Arthurs, Johansson, Philippoussis, Isner, Karlovic...), they may have some (or many) service-related statistics with better numbers than those of Sampras or Roddick. I would not be surprise if that were the case.

Why? Because you don't win many GS tournaments and be nº1 in the world for years because of a great serve (as I said, most of the great servers of each era don't won GS tournaments, don't even reached GS finals most of them, even if they win almost as much service-games as Roddick or Sampras).

It is the whole package that is important, and it shows most in the return-games.

It is possible that Sampras won many more return-games than the rest of those great serve players (even though Sampras usually never tried hard for a second break of serve if he was already up a break).

[NonP]

Originally Posted by piece

NonP, what do you think the reason for the increase in hold percentages since the 90s is?

Well, I've written on this very topic at length in those posts I linked upthread and you can go back and read them for details, but to wit the top three reasons are technology, strategy and mentality.

Now I don't think the latter played as big a role as the first two, at least not as far as the '90s are concerned. For example in the past you could hear even a top-notch server like Gonzales dismissing flat serves as "for show-offs," which displays a mindset foreign to more recent players who, of course, try to win points outright on their serves.

Part of that mentality was due to the prevalence of S&V. That is, if you were coming to the net it was generally better to take a few mph off and get higher %s of 1st serves in, which resulted in fewer freebies. In fact the difference is stark when you compare the %s of free points on serve between the eras: before the '90s the average %s of unreturned serves rarely broke the 40% mark, even for big servers like Tanner and Becker, whereas now 40% is relatively common and even 50% isn't entirely unheard of.

And of course there's the matter of technology. Again it's s a myth that the modern racquets boost your maximum serve speeds significantly, but they do allow you to impart more spin, which helps increase your 1st-serve %s and thus your average serve speeds. And your speeds get an additional boost because the extra spin allows for bigger 2nd serves, too, and again since players are not following their serves to the net anymore and it's not as important to ease on the pace and get more 1st serves in. People usually ignore this serve part of the equation when they talk about how the new racquets have aided the return.

As you can see all these factors are related to each other, and it's hard to say which one contributed most to the steady increases in service stats. In any case people should've realized by now that all the chatter about surfaces slowing down or the supposedly improved return, spin and/or athleticism making S&V/all-court tennis obsolete is ultimately irrelevant, for the following reasons:

1) We know, contrary to popular belief, that players today are holding serve with more ease than perhaps ever.
2) We know, again contrary to popular belief, that the racquet/technology/spin/fill-in-the-blank revolution didn't drive S&Vers out of the game, since they were already becoming extinct in the '90s, well before the so-called revolution took root.
3) There are more than enough examples of players, some from the '90s, enjoying success thanks to S&V in the '00s and beyond, which belies the boilerplate talk that it is no longer a viable strategy today.

And really, people have no more excuse for remaining in the dark about all this, especially since they often happen to be the same ones who engage in all kinds of statistical contortions to prove their case and they could easily look up the above stats with just a couple clicks.

Originally Posted by slice serve ace

- so roddick leads in their total at AO, and is only slightly behind at USO and W

- roddick leads at all 3 slams in first 4 rounds, even at wimbledon, and is quite better at AO

- sampras leads huge at all 3 slams from QF to F

Yes, and I'm sure you already know that, at least at Wimbledon, Pete won a higher % of service games in the finals overall than in the earlier rounds. Which is pretty amazing if you think about it--even if you were to believe that Pete switched gears for the big matches you'd still expect the %s to be lower in Slam finals.

BTW do you have the 1st- and 2nd-serve stats (including unreturned serves) for all of Pete's Wimby finals? I've been meaning to compare and perhaps start a thread on Pete's & Fed's Wimby title runs (along with another one on unreturned serves in general--lots of work involved there), and so far I have every one of Pete's except those for the 1997 and '98 finals. I do have roughly accurate numbers for the '97 because I once watched the whole thing and wrote down whatever match stats I could think of (I'll definitely do a thread on this match once I verify everything), for the '98 final I have nothing but the 1st-serve stats.

I know you have lots of other obscure stats at hand that I'd be interested in, so let me know if you want to exchange private messages instead of posting everything here and possibly hijacking the thread. As my contributions of sorts I have Fed winning 75.6% of his 1st-serve points (68/90) and 48.8% on 2nd (20/41) in last year's Wimby final, and winning 19/21 service games in the final and 92.44% (110/119) for the whole tournament (4/20 and 33.33% or 38/114 of return games respectively). Those numbers should not be easy to come by, if you don't have them already.

Originally Posted by mattennis

I am interested in NonP opinion about the service-related statistics constantly improving since 1991.

Are you implying that ATP-statistics are mainly wrong?

I don't believe for a second that Goran Ivanisevic didn't hit more than 10 aces in a tennis match prior to 1992. (I know for a fact that he had matches with more than 20 aces even in 1989).

The '90s ATP stats can be unreliable, especially the %s of 2nd-serve points won which can be off by as high as 12% (at least among those I've seen) as the double faults are often counted twice, but both the # of aces and the %s of service games won should be mostly accurate. Trust me, I actually crunched a few numbers to assure myself that Goran's total ace count in '91 wasn't just a simple data entry error.

See above for a detailed explanation of how the improved service stats came about.

Originally Posted by pc1

In all seriousness it could be the best thread.

Sure it is.

[slice serve ace]

Originally Posted by NonP

BTW do you have the 1st- and 2nd-serve stats (including unreturned serves) for all of Pete's Wimby finals? I've been meaning to compare and perhaps start a thread on Pete's & Fed's Wimby title runs (along with another one on unreturned serves in general--lots of work involved there), and so far I have every one of Pete's except those for the 1997 and '98 finals. I do have roughly accurate numbers for the '97 because I once watched the whole thing and wrote down whatever match stats I could think of (I'll definitely do a thread on this match once I verify everything), for the '98 final I have nothing but the 1st-serve stats.

I know you have lots of other obscure stats at hand that I'd be interested in, so let me know if you want to exchange private messages instead of posting everything here and possibly hijacking the thread. As my contributions of sorts I have Fed winning 75.6% of his 1st-serve points (68/90) and 48.8% on 2nd (20/41) in last year's Wimby final, and winning 19/21 service games in the final and 92.44% (110/119) for the whole tournament (4/20 and 33.33% or 38/114 of return games respectively). Those numbers should not be easy to come by, if you don't have them already.

hijacking the thread? with the number of trolls in this forum, nobody will notice


so, service stats for pete, wimby 97 and 98? here


1997

1st serve 45/76, 59%
1st serve pts won 39/45, 87%
2nd serve pts won 21/31, 68%
unreturned serves 45/76, 59%


1998

1st serve 85/155, 55%
1st serve pts won 69/85, 81%
2nd serve pts won 44/70, 63%
unreturned serves 71/155, 46%

[Moose Malloy]

I don't believe for a second that Goran Ivanisevic didn't hit more than 10 aces in a tennis match prior to 1992

The ATP has him with several matches with 20+ aces in '91. They have no stats on any matches prior to '91.

And they don't track Davis Cup, which has been the site of some rather high ace counts(Guga's 47 ace match etc) or Grand Slam Cup(some 40+ ace matches for Goran there)

[droliver]

Originally Posted by slice serve ace

i compared roddick and sampras service games won at slams and in these categories

-total service games won for each slam
-round 1 to round 4 for each slam
-quarterfinal to final for each slam

since roddick never advanced past 4R at roland garros, we can leave that out (enough to say sampras has better stat at rg, 82.9% against 81.7% for roddick)

i got some really interesting results, here they are;


australian open


total

sampras 89.3%
roddick 90.5%


R1 to R4

sampras 90.1%
roddick 93.1%


QF to F

sampras 87.6%
roddick 82.3%


wimbledon


total

sampras 94.1%
roddick 93.4%


R1 to R4

sampras 94.2%
roddick 94.6%


QF to F

sampras 94%
roddick 90.2%


us open


total

sampras 91.6%
roddick 91%


R1 to R4

sampras 92.3%
roddick 93%


QF to F

sampras 90.2%
roddick 84.8%


- so roddick leads in their total at AO, and is only slightly behind at USO and W

- roddick leads at all 3 slams in first 4 rounds, even at wimbledon, and is quite better at AO

- sampras leads huge at all 3 slams from QF to F


i'll let you draw the conclusions, can't write anymore

The numbers you cite are indirect measures of the serve and are confounded by the other factors that affect service holds (ie, strength of ground game). They suggest Sampras is a better player, but do not really prove anything about the serve.

The stats that are most important for assessing the serve as a stroke are
- Aces
- Winners
- 1st serve %
- Double faults

[slice serve ace]

indirect, yes. the total % of unreturned serves would be the closest to direct measurment of the stroke itself (aces are only subsection of that stat)

my intention with stats that i posted is to show that pete's service game held much better when the going gets tough. sure, everybody knows that his game outside of the serve is much better than roddick's, but i also think that unpredictabillity and better placement of pete's serve is more usefull against better opposition than predictable roddick's 140 mph bomb down the T

[Bhagi Katbamna]

Sampras had this ability to pull out huge first and second serves at critical times. Roddick would win service games easily during the first few games of the set but towards the end when he really needed some first serves and easy games, he couldn't do it(esp. against top players).

[brettsticker86]

roddick for sure. could hit it 10-20 mph harder, clutch city and had a huuuuuge kick serve. no question.

[fed_rulz]

Originally Posted by slice serve ace

indirect, yes. the total % of unreturned serves would be the closest to direct measurment of the stroke itself (aces are only subsection of that stat)

my intention with stats that i posted is to show that pete's service game held much better when the going gets tough. sure, everybody knows that his game outside of the serve is much better than roddick's, but i also think that unpredictabillity and better placement of pete's serve is more usefull against better opposition than predictable roddick's 140 mph bomb down the T

if by "tough" you mean later rounds, you might want to reconsider that. One oft-mentioned point in this forum is the notion that the 16-seeding system in slams resulted in huge upsets more often... which translates to easier rounds later in slams more often for the top seeds that made it. also, the concept of "surface specialists" --- a term invented to refer to folks who sucked on surfaces, other than the ones they're "specialists" in. These "specialists", while great on their surface, lacked big match clutchness (naturally.. grass specialists get one tournament per year to fine-tune their mental strength)... which could also explain the higher hold stats.

either that, or we could just analyze stats without adding caveats to it when it doesn't suit our view points (e.g. explaining away consistency of current top seeds to a 32-seed system, or homogeneity in surfaces or bemoaning lack of surface "specialists")?

[slice serve ace]

Originally Posted by fed_rulz

if by "tough" you mean later rounds, you might want to reconsider that. One oft-mentioned point in this forum is the notion that the 16-seeding system in slams resulted in huge upsets more often... which translates to easier rounds later in slams more often for the top seeds that made it. also, the concept of "surface specialists" --- a term invented to refer to folks who sucked on surfaces, other than the ones they're "specialists" in. These "specialists", while great on their surface, lacked big match clutchness (naturally.. grass specialists get one tournament per year to fine-tune their mental strength)... which could also explain the higher hold stats.

either that, or we could just analyze stats without adding caveats to it when it doesn't suit our view points (e.g. explaining away consistency of current top seeds to a 32-seed system, or homogeneity in surfaces or bemoaning lack of surface "specialists")?

ok let's reconsider

if bolded stands, sampras should play lower ranked oponents more often at QF to F stage than roddick

since they come from the different seeding sistem, let's compare them in 2 categories; how much they played players ranked 1-16 and 1-32


australian open

roddick has 10 matches from qf-f
sampras has 14 matches from qf-f

vs players ranked 1-16

roddick played 5 of 10 matches, 50%
sampras played 10 of 14 matches, 71%

vs players ranked 1-32

roddick played 7 of 10 matches, 70%
sampras played 12 of 14 matches, 83%



wimbledon

roddick has 12 matches from qf-f
sampras has 24 matches from qf-f

vs players ranked 1-16

roddick played 6 of 12 matches, 50%
sampras played 14 of 24 matches, 58%

vs players ranked 1-32

roddick played 9 of 12 matches, 75%
sampras played 19 of 24 matches, 79%



us open

roddick has 12 matches from qf-f
sampras has 27 matches from qf-f

vs players ranked 1-16

roddick played 8 of 12 matches, 67%
sampras played 21 of 27matches, 78%

vs players ranked 1-32

roddick played 11 of 12 matches, 92%
sampras played 26 of 27 matches, 96%



so sampras leads in every single category at every slam. looks like bolded doesn't stand in this comparation. sampras didn't benefit from lower ranked oponents - in fact he played tougher oposition more often, if we go by ranking, no matter if it's 1-16 or 1-32. so i stand by what i said before; sampras service game is in different class from roddick's from QF to F. is it because of sampras' superior backup game or the serve itself, that is arguable.