Is Returning Serve harder today than it used to be?

There has been a fair amount of discussion in relation to how much easier it is to break serve in today’s tennis, given the fact the courts have slowed down. I thought I would do a bit of a comparison against the 1990 as, frankly, that does not really accord with my recollection.

I have taken the percentage service games won for the top 20 servers for the last three years and compared against the equivalent figures for 1997-1999. The average percentages of service games won for the top 20 servers in each of those years was as follows:

1997 – 83.85%
1998 – 83.2%
1999 – 83.5%

2010 – 85.7%
2011 – 85.25%
2012 – 86.45%

So there is a fairly consistent pattern, the best servers today win, on average around 2-3% more service games than their equivalents in the late 1990’s.

There is no doubt that the courts have slowed down. However, the fact that servers are serving quicker than ever seems, from the statistics, to have negated the effect of that slow down. The statistics certainly seem to suggest that returning is no easier now, if anything it seems to be more difficult to break serve. Is there another explanation that I am missing?

Fantastic article Born_Slippy, although I do have to admit I was surprised at the stats!

Is it possible to find out the service games won percentage overall on tour, that would also reveal a lot imo.

Tremendous analysis, and really fascinating.

So many things have changed in the last 15 years it's impossible to say what's caused the difference, if indeed it is statistically significant.
There is a 1.4% difference between 1997 and 2011 and a 1.2% difference between 2011 and 2012
Also being harder to break serve doesn't necessarily correspond to being harder to return the serve.

I take it we can agree it is the most likely answer though. I'm struggling to think of any other one which is particularly plausible to be honest.

The difference may not be statistically significant, although I would have thought the pattern is sufficiently compelling to suggest it might be. However, what it seems to fairly clearly show is that, regardless of the court slow down, it certainly is not any easier to break serve.

What would explain the difference between 2011 and 2012 though?

Born Slippy wrote:I take it we can agree it is the most likely answer though. I'm struggling to think of any other one which is particularly plausible to be honest.

The difference may not be statistically significant, although I would have thought the pattern is sufficiently compelling to suggest it might be. However, what it seems to fairly clearly show is that, regardless of the court slow down, it certainly is not any easier to break serve.

I'd agree with your findings but would say you are confusing the issue. The art of great service return should be judged on the amount of service returns that end up not being returned or being winners. Service breaks can be a result of four points won by the reurner after lengthy rallies hence his rallying skills win him the service break and not necessarily his actual service return.

I don't think I'm confusing the issue Craig. I'm not seeking to judge the art of a great service return. I'm purely looking at whether the statistics indicate it is easier to break serve now.

Your statement is generally correct, although you also need to factor in that a player who is facing mainly serve volleyers will have more "winning" returns, so that wouldn't necessarily be a great way of judging matters. In addition, there is a significant skill in getting the return back against the big servers, even to give yourself some footing to try and out-rally them.

Where did you get the stats from, by the way? I can never find historical data like that.

Yes I see what you are saying and very interesting findings as well.

JuliusHMarx wrote:What would explain the difference between 2011 and 2012 though?

Don't know. Statistical quirk? The fact Raonic is up at a quite ludicrous 93% of service games won? The courts being slightly quicker this year than last?

JuliusHMarx wrote:Where did you get the stats from, by the way? I can never find historical data like that.

ATP matchfacts now go back to the 90s. You can look at all of their categories year by year, under Scores and Stats. Bit buggy though, at least for me, seems not to work quite regularly.

I do not know if ATP Match Facts will have average strokes per rally, but it would be good to see how it corresponds to the OP.

BPs converted versus BPs created would also be interesting.

PS: Excellent Born_Slippy.

Great find. Just the other day I was wondering whether this kind of data was available somewhere. Very interesting, not least because it goes against conventional wisdom, and - I have to admit - against what I expected to see.

Interesting. Would be good to see the data right across the period...and also across the early 2000s where we know conditions slowed.

Perhaps in mitigation to the apparently confounding data we have increasingly racquet tech (lighter racquets = increased racquet head speed) and string tensions have come down alot from average of 60-65 lbs to low-mid 40s (Raonic is strung at 40-44lbs with Luxilon). These factors have increased general service speeds markedly which may help service game % improvement.

lydian, would these also have helped control placement of serve and put more spin on serves, esp. second serves?

What a well timed article!! As I was considering this the other day. Certainly it was harder to return serve on the quicker courts of the 90's and it took quite a while for Agassi to be a consistent 'great' returner

However, despite the surfaces, the serve is quicker now and the returning of some even (arguably) better

Born Slippy wrote:

JuliusHMarx wrote:What would explain the difference between 2011 and 2012 though?

Don't know. Statistical quirk? The fact Raonic is up at a quite ludicrous 93% of service games won? The courts being slightly quicker this year than last?

I think the courts have been slightly quicker this year. Certainly Wimbledon and the US were. All to do with the climatic conditions at the time (a very warm spell before Wimbledon for instance) rather than anything else

Interestingly, if you look at the top 10 players with the highest % of first serve points won, they are all retired players (including Roddick). Does that indicate it's easier now to win return points?

But on 2nd serve % points won, there are 4 current players (plus Roddick) in there.

So it seems it's easier to win points against the first serve nowadays, but harder to win points against the second serve.

That is intriguing, although you've missed big Ivo sitting at number 2 all-time of the first serve points won. Counting Roddick as a current player i think means 8 of the top 20 on that stat are current.

From the late 90s I recall sets where you almost expected a tiebreak before the sets began..nowadays you have far more breaks it seems. But maybe that's because I mostly only watched Wimbledon in the 90s, I bet the statistics would be different for Wimbledon.

Instinctively, it looks like a just about statistically valid change to me, although hard to be sure.

There is also the discussion of whether its easier to hold serve on slower or faster surfaces. Intuition tells us faster...but is it?

On a slower surface you get more time to move to the return (or any other shot for that matter - hence we have more guys over 6'6 doing ok than before). On slower surfaces you dont have to take the high risk of volleying, you can be selective what you come in on. Also, if you have a good serve its still going to be hard to reply to...then the server can come in and make the midcourt mop-up winner perhaps easier than yesteryear when it come through quicker and lower.

I have to say if there was an award for thread of the year I would nominate this one. Unfortunately, I was in the princpal's office when it was first posted. Excellent research Born slippy and it isn't just that you did the research, but that the research you did is so dispositive of the discussions we have been having on court conditions and the quality of returning of Djoko and Andy. The fact of the matter is that I feel quite vindicated by the content of this research by Born Slippy.

This fits in to exactly in to exactly what I have been saying. We can not do more than incrementally and modestly speeding up a few tournaments. The proposals for banning technology and drastically speeding up conditions as I have been saying would have disastrous results as the numbers that we see posted in the OP would spike to near 90 percent or possibly over. Giving a big edge to servers on the first serve and first forehand will in fact decrease variety more than it will increase it. Is John Isner all of sudden going to become steffan Edberg or more likely will he just hit more aces and first ball winners off the forehand? As I have been saying for years the game needs slow conditions or the bigger male athletes on tour will easily hit through the court on the first serve and first forehand.

This also goes a long way in proving what I have been saying that Djokovic is the best returner of all time and that murray is right up there with him. Djokovic's 35 percent break percentage looks even more impressive in light of these numbers as dos Nadal and Murray's numbers in fact. This hurts those who constantly want to negate the accomplishments of todays stars and unfortunately there are some people who must constantly attempt to diminish their accomplishments and abilities. This is a massive setback for these types of critics.

lydian wrote:There is also the discussion of whether its easier to hold serve on slower or faster surfaces. Intuition tells us faster...but is it?

On a slower surface you get more time to move to the return (or any other shot for that matter - hence we have more guys over 6'6 doing ok than before). On slower surfaces you dont have to take the high risk of volleying, you can be selective what you come in on. Also, if you have a good serve its still going to be hard to reply to...then the server can come in and make the midcourt mop-up winner perhaps easier than yesteryear when it come through quicker and lower.

Come on lydian it sounds to me that you are grasping straws here a bit. You are making a lot of anectodal observations. Of course it is easier to hold serve on a fast surface that is why break numbers are always higher on clay and lower on grass and indoors there is no complicated scenario that makes holding serve easier on a slow surface now as opposed to the past.

I think you are a very knowledgable and respectful poster. However I think these numbers from Born Slippy really cuts against the idea that we need to give the attacker any advantage on the serve. And I am afraid what your suggetions about limiting lightweight racquets, face size, and string compositions will make these already historically high holding numbers shoot through the roof. As I have been saying the larger, stronger male tennis player needs slow conditions or he will just hit through the court way too easily.

I am curious does this earth shaking post change your theories about the need to speed up conditions? If you still believe in faster conditions what do you think will happen at wimby when the big servers meet after the changes you suggest are made? And do Nadal, Djoko, and Murray go up in your estimation in terms of the all time returner list as a result of this research?

interesting article. I believe the quality of the serving has improved markedly since the 90s. Statistically the difference is marginal, but if we take it as given that returning has improved there is little room for doubt (certainly first serve speeds have increased a fair bit). As for the slightly bizarre difference between this year and last, it could indeed be down to the "freak" (that's a technical term) Raonic, as averaging over only twenty players won't necessarily iron this out. For me there are two main reasons for serve improvement:

- technology improvements. Better racquets mean players can hit the ball harder while maintaining a degree of control. Whoever made the point about being able to put more spin on the second serve and thus get more out of it is entirely correct also.

- increased athleticism. Players nowadays are fitter, simple as. This means they can keep on banging down huge serves for longer without tiring. I'm sure a comparison set by set would show that in the 90s there were more breaks towards the end of a match, not sure this is the case today.

what does this tell us? well for starters as socal suggests speeding up all courts to resemble 90s speeds would be a bad idea. Tennis has moved on, and I think we need to accept that. Of course we could also go back to the era of wooden racquets and increased court speeds, but I think this would be daft. All sports change. It also increases my admiration for Murray and Djokovic, who I believe are at the top of the all-time return lists (on HC maybe?).

Let's assume the following (albeit as generalizations across the tour, which might not be entirely true)
1. It's harder to break serve nowadays that 15 years ago
2. Courts are slower now than 15 years ago.

Clearly, if those assumptions are true (and personally I think more research is needed), then some other factor, not court speed, is helping the servers (unless we make the most obvious response to making it easier to break serve - return to the conditions of 15 years ago. That's probably too simplistic.)

The most obvious culprits are rackets/strings. If we could adjust the technology correctly, we could tip the balance back in favour of the returner. But then we would get endless rallies of 30, 40, 50+ strokes.
What we need is a change in conditions (rackets, strings, balls, rules??) that takes away slightly from the server, but favours slightly the attacking player once the rally is underway. Tricky.

My own view is that nothing ever changes until its forced to. People continue down destructive paths until they realise their folly, then change.

Formula 1 was moving towards disintegration as Ferrarri overwhelmed their rivals with their unique blend incredible team vision (Brawn), brilliant driver/developer (Schumacher) and mega budget. Standing back and leaving it be was heading towards Michael winning every race, and there were two years of procession races.

Only once this was obvious and on the verge of realisation did they change.

So my recommendation is to do nothing. Let the strings advance further (as they will), let the courts slow down more, and let the players ignore the rules so they can play ever longer rallies. Then eventually the problem will be obvious and the game will change. It'll actually happen when customers vote with their feet.

Parochial fans are unreliable but they change as events develop. For instance, I never heard any Nadal fans unhappy with the rigours of a slow tour until recently. It'll be the same when Djokovic gets hurt or outlasted regularly (as he was by Murray at New York).

Expecting more requires people to have long term vision, and frankly there aren't that any people around, and those that are do more interesting things than administer tennis.

Mad for Chelsea wrote:interesting article. I believe the quality of the serving has improved markedly since the 90s. Statistically the difference is marginal, but if we take it as given that returning has improved there is little room for doubt (certainly first serve speeds have increased a fair bit). As for the slightly bizarre difference between this year and last, it could indeed be down to the "freak" (that's a technical term) Raonic, as averaging over only twenty players won't necessarily iron this out. For me there are two main reasons for serve improvement:

- technology improvements. Better racquets mean players can hit the ball harder while maintaining a degree of control. Whoever made the point about being able to put more spin on the second serve and thus get more out of it is entirely correct also.

- increased athleticism. Players nowadays are fitter, simple as. This means they can keep on banging down huge serves for longer without tiring. I'm sure a comparison set by set would show that in the 90s there were more breaks towards the end of a match, not sure this is the case today.

what does this tell us? well for starters as socal suggests speeding up all courts to resemble 90s speeds would be a bad idea. Tennis has moved on, and I think we need to accept that. Of course we could also go back to the era of wooden racquets and increased court speeds, but I think this would be daft. All sports change. It also increases my admiration for Murray and Djokovic, who I believe are at the top of the all-time return lists (on HC maybe?).

That is exactly how I see it MFC, with these type of numbers on the serve there really is no logical way to do more than incremental and minor change vis a vis the conditions. I think the main reason is that players are taller and stronger than they have been in the past. Modern training and diet are transforming the modern athlete at an evoutionary breakkneck pace. The human animal is not the same as it was 100 or even 50 years ago. There was a reason why all the tournament directors slowed the tournaments 10 years ago, and this thread by Born Slippy illustrates the dilemma they faced with the super charged modern athlete.

If the game loses fans and their is a genuine broadbased call for faster conditions the tournament directors will make the change again. If there is no desire in the marketplace and as I have been saying the fans by in large like the modern game then we will not see any serious changes. And frankly that is probably the way it should be. Tennis like all modern sports will be impacted by the continued to advancement of modern medicine, training, and nutrition. Speed and fitness play an ever increasing role in all modern sports and so be it with tennis as well.

Very good research on stats BS , really it is surprising to know breaking of serve is more tougher these days inspite of conditions slowing down massively, but I don't think so thats anything to do with the quality of the return. For instance Murray might have excellent quality of return yet Federer on his day can go un broken even though Murray might have won several rebound aces in the game.

I see more tall guys in the game meant they hold service easily, I guess we need to look into stats what was the average height of players in 90's to today's, I guess in 90's average height might have been 5'8-5'9 and now it could 6'1-6'2, and its already leading to 6-3' I guess.

Anyways a good though provoking article.

Born_Slippy, can you explain a bit more about how you got the figures.
I selected the top 200 players for 1997 from http://www.atpworldtour.com/Rankings/Top-Matchfacts.aspx?y=1997&s=0
copied the data into a spreadsheet, sequenced by % service games won, and averaged the top 20 (i.e the top 20 servers).
I got 85.65%, not 83.85%

There is a matchfacts leaders tab or something like that on which i just chose 1997 and took the top 20. I think players have to play sufficient matches to make the list, which probably explains the difference.

The serves are much quicker in the modern game, though in the 90's the string technology was not a evolved as it is now which left returners very little to work with in terms of control in the 90's.

Agassi and Murray for me are the 2 greatest returners of serve I have seen.

I would wonder how much more quicker serves would be if they quickened the surfaces up

legendkillarV2 wrote:I would wonder how much more quicker serves would be if they quickened the surfaces up

They wouldn't be measured quicker as the measurement is the speed off the racket. So if serves are faster nowadays it's either because of better tehcnique, physical improvement or technology changes.

JuliusHMarx wrote:

legendkillarV2 wrote:I would wonder how much more quicker serves would be if they quickened the surfaces up

They wouldn't be measured quicker as the measurement is the speed off the racket. So if serves are faster nowadays it's either because of better tehcnique, physical improvement or technology changes.

I would go with a mix of all three (improved technique is one I hadn't thought of...)

JuliusHMarx wrote:

legendkillarV2 wrote:I would wonder how much more quicker serves would be if they quickened the surfaces up

They wouldn't be measured quicker as the measurement is the speed off the racket. So if serves are faster nowadays it's either because of better tehcnique, physical improvement or technology changes.

True though that is, the speed of the impact off the surface would be different.

With that statement I shall not nominate myself to research such speeds!

Mad for Chelsea wrote:

JuliusHMarx wrote:

legendkillarV2 wrote:I would wonder how much more quicker serves would be if they quickened the surfaces up

They wouldn't be measured quicker as the measurement is the speed off the racket. So if serves are faster nowadays it's either because of better tehcnique, physical improvement or technology changes.

I would go with a mix of all three (improved technique is one I hadn't thought of...)

I'm not sure many (or any) could have a better technique to produce faster serves than, say, Tanner, Philipoussis or Ivanesivic, but maybe more players have mastered (or got close to it) the technique, making the average serve speed faster. That's just me speculating.

Coaching techniques also advance Julius as shots are invented to use the new technologies better. Things like the modern forehand, topspin, and serve have been developed over generations and new twists and techniques popularized. This was one of my points to lydian that the skill set expands and does not diminish for the modern champion

JuliusHMarx wrote:Born_Slippy, can you explain a bit more about how you got the figures.
I selected the top 200 players for 1997 from http://www.atpworldtour.com/Rankings/Top-Matchfacts.aspx?y=1997&s=0
copied the data into a spreadsheet, sequenced by % service games won, and averaged the top 20 (i.e the top 20 servers).
I got 85.65%, not 83.85%

Have had a chance to properly look at this now. Your stats would have the likes of Safin (played one match in 1997) appearing and will therefore clearly be higher. I just used the following link and then did a search directly from there on 1997 service games won:

http://www.atpworldtour.com/Matchfacts/Matchfacts-Landing.aspx

It only includes those who played a substantial number of matches - looks like 40 or so is the cut-off.

Which makes sense if you don't play a full season your statistics can be skewed up or down based on an inadequate sample size.

Slippy, if this data is correct I wonder whether this connects to my observation about slower courts encouraging taller players towards the top of the game?

Slower courts help taller players thrive because now they can rally, who then make it harder to return by being there to chuck bombs down.

Not saying its proven fact, but its not inconsistent.

It is possible but I don't think the stats help conclude that either way.

Incidentally, just going back to a point Julius made, it is second serve points won which I think have taken a big jump - perhaps because of the extra spin now available with the tech.

First serve points, relying more on pure pace, have dipped slightly from the peak in 1996 - 1998 (78.3 - 78.4%) - albeit they have remained remarkably consistent from 2001 to now (ten of the last 12 years between 76.1% and 76.8%).

In that case, it's probably technology that is helping the servers hold serve more.

However, we'd need a statistician to help us decide if lowering the sample size (by selectively excluding players in the top 200) would lead to more, or less, accurate figures.

Julius what about the taller and physically stronger players, isn't that as well allowing players to hold more easily? I personally think the cotinued rapid evolution of the modern athlete goes a long way towards explaining these numbers.

Evolution is the process of selection for fitness for purpose; in other words, the creature changes to suit the environment because a different environment demands a different solution. It works that way round; environment first, followed by adaptation.

That taller players thrive shows not that players have grown - tall men are nothing new - but that changes in conditions and equipment now suit them more than they did, say, 20 years go.

It stands to reason that the tennis environment is better suited to great big men than it was a while back.

socal1976 wrote:Julius what about the taller and physically stronger players, isn't that as well allowing players to hold more easily? I personally think the cotinued rapid evolution of the modern athlete goes a long way towards explaining these numbers.

Except that Born Slippy's figures indicate that the improvement lies with the second serve, not the first. To me that indicates technology more than other factors. Howver, I still think more statistical research is required before drawing number-based conclusions.

bogbrush wrote:Evolution is the process of selection for fitness for purpose; in other words, the creature changes to suit the environment because a different environment demands a different solution. It works that way round; environment first, followed by adaptation.

That taller players thrive shows not that players have grown - tall men are nothing new - but that changes in conditions and equipment now suit them more than they did, say, 20 years go.

It stands to reason that the tennis environment is better suited to great big men than it was a while back.

Actually the large number of tall people who are also as agile as they are is something relatively new in human anthropology. In Iran you can really see it much more so than in America. The average persian male was probably 5 foot five for a couple of thousand years at least if ancient greek historians are accurate or the british who visited in the 16th and later 17th century. In the last 100 years with industrialism, oil dollars, the society has changed 180 degrees. Till a century ago 95 percent of the population had no electricity and starvation was routine form of death. Nutrition was poor and for most of the population was mainly bread. In the last half century really the persians are getting relatively statesque. You go out in tehran and the young generation is all these tall girls and boys relatively speaking. You look at their grandparents and you would be stunned at how this pygmy gave birth to this statuesque goddess. Until you understand that even up to a century ago these grandparents were the result of widespread poverty and malnutrition. Anthropological historian estimate that the average Russian for example in 1917 was living onf 900 calories a day and probably a lot of that was vodka. Industrialization, technology, modern science, and just pure wealth and the ability to feed children milk and meat has shot up the size of the species. And in asia particularly this is pronounced. Iran is not alone there are dozens of asian countries and third world countries that experienced the same trend. In the west the standards of living rose and industrialization caused this spike much earlier. But for most of the world these malnutrition and intense poverty issues weren't solved till just half a generation ago, in some they are still not solved. The human animal is growing and getting stronger and faster at an unprecedent evolutionary pace that is unique to the last half century.

If you watched the NBA you would understand this. Take some footage of the recent US olympic team and watch the size, speed, and leaping ability of today's players and then watch some NBA highlights from the early 1960s or even the early 70s. Then look at the sizes of the players and their builds in comparison. I mean it isn't even funny the difference in size, speed, and strength within the last half lifetime. The reason the NBA is so good of an indicator because height and speed combination is so valuable to the NBA. Look at volleyball for god's sake these guys look like they are spiking the ball from the back line.

Certainly people are much taller generally, but that would just mean more people unsuited to top level tennis unless something had happened to the tennis environment to benefit them. Many of the outcries you cite haven't yet produced tennis players, the game is still dominated by countries which have been wealthy for a long time.

If height had always been a big advantage how come we'd not ever seen the sport dominated by a massive player, who being rarer would have this advantage all to himself?

If it is all about the second serve I'd say it sounds like the strings again.

For the sake of the debate can you define what you mean when you say 'taller players' . Are we talking over 6ft, 6' 3'', 6' 6'' ?. The current top 4 are all between 6' 1" and 6' 3". That's tall to me at 5' 6" but may not be to others.

I'm assuming we're talking about 6' 6" and above, but I'm not defining the point.