ATP data analysis and the myth of slow courts [Archive] - MensTennisForums.com

ATP data analysis and the myth of slow courts

tennis_analyst
03-17-2012, 11:34 PM
Hi all

This is my first port on this forum. Basically, I wanted to share some analysis I have done on the recently released statistics by ATP. Here (http://www.atpworldtour.com/Rankings/Top-Matchfacts.aspx) you can find some key stats on the top 200 players for every year since 1991 and this is the data I downloaded and used.

In this post, I try to shed light on the eternal debate about the surface speed and how it was slowed down lately. So I attach 3 charts (hopefully they will display well, still trying to figure out the system) that for each surface show these variables…

Percentage of first serve points won
Percentage of service games won
Percentage of points won returning first serve
Percentage of return games won

…averaged over top 50 players in each year. The pattern is fairly similar if I use all 200 players but some players ranked lower just don’t have enough matches so I cut at top 50.

These variables ***** for how easy it is to hold and break on a given surface in a given year, so we can get some nice insights from them. Here are my overall conclusions:

1) The slowing down of the Wimbledon grass is rather exaggerated. People say that the grass has been slower since 2002 (Federer stopped S&V in 2004) but since 2002 it actually became EASIER to hold serve and win points off the first serve! You can see an upward trend. At the same time, winning return games has become slightly more difficult, the line drops below 20%.

2) On hard courts the variations across years is small but the trends are striking: it has become easier to serve and more difficult to return if you compare early 1990s and late 2000s.

3) Clay has played fairly fast over the last 4 years.

This is obviously a very crude analysis but it goes contrary to the common belief that courts have been slowed down. So unless top 50 players suddenly became better servers and worse returners (unlikely, since the technology has benefited the defense game more), then we have to acknowledge that the courts are in fact playing quicker now.

This is rather surprising to me, other explanations for these results?

Later, I will post some analysis for players like Federer, Nadal, and Djokovic to see how their careers developed. Stay tuned!

PS. Any ideas how I can make those charts appear in the message and not as attachments? Thanks

scoobs
03-17-2012, 11:42 PM
post them to a photo hosting site and add them as images.

Li Ching Yuen
03-17-2012, 11:49 PM
You're probably the same guy that would deduce that clay is a super fast surface just because a lot of tie-breaks are being played on it and that players volley on it at times.

"Statistics are like mini-skirts..."

sexybeast
03-17-2012, 11:57 PM
Indeed you are right the whole surface debate has gone insane and surface change has been used to find an excuse for whatever people dont like about modern tennis. People dont want to bash raquets technology change because that is deterministic in its nature and cant be regulated.

Serve and volley has been killed because raquets can create incredible spin and has a big sweet spot so players can throw themselves at serves and just stretch themselves without even seeing how they hit the ball and get the ball back at players feet. It is like target practice for the one passing whoever is approaching the net.

Think of it like this, you are a goal keeper and let players shoot penalties with old heavy shoes and heavy balls from the 30s, now let them shoot the ball with modern balls and modern shoes, technology change is no good for the goalkeeper because he mostly uses his body and flexibility to cover the goal but for the one shooting penalties the change in equipment makes huge differense.

In the 90s almost every year some serve and volleyer could reach SF/F in RG(that is supposed to have been super slow in the 90s according to nostalgics), now players cant reach QF in Wimbledon serve and volleying.

romismak
03-18-2012, 12:00 AM
Yes grass is slower, clay is faster and HC - generally slower, but there are many types, and outdoor-indoor difference too. I am convinced that slowing down surfaces was neccesary for future of tennis, but not such big slowing down. Can we imagine 90s fast surfaces , HC, grass, carpet, indoor events with guys serving 130mph 1st serves like nothing and having such high FS%? Also that trend that players are not backing up their serves with volleys is reason why more service games are hold, before you just made good passing shot and you won the point, now after good serve you can control point from baseline and you donīt risk that you will immediately loose it, because of passing shot.

philosophicalarf
03-18-2012, 12:10 AM
This is rather surprising to me, other explanations for these results?


Serve statistics are almost entirely unaffected by bounce (not pulling that out of my ass, I've got 18 months of hawkeye bounce height data behind it). There are some exceptions, eg people called Isner, or those with good kick serves. As an average it's true though.

In the rally though, higher bounce means more time to retrieve behind the baseline. The ITF's court pace rating incorporates separate friction and bounce measurements for this reason - thus a "medium" rated court can be fast/high, medium/medium, or slow/low.


This is also the same explanation offered by the head groundsman of Wimbledon. He said the speed through the court hasn't changed at all since the 90s, but they roll it harder now to make it last the two weeks (which results in a higher bounce).

LisaKoh
03-18-2012, 12:15 AM
Problems with this analysis/data set:

1. This all rests on the assumption that percentage of service games won/held are an indicator of the surface speed. This variable is a factor, but it does not tell the whole story.
2. Perhaps one overlooked indicator of surface speed is average length of rallies/points. (i.e. 3 versus 10 shots, etc.)
3. Instead of focusing on correlates, why not go with the BBC data that demonstrated at least 9 mph slowing from 2003-2008?

JanKowalski
03-18-2012, 12:18 AM
Those statistics prove nothing :shrug:

Looner
03-18-2012, 12:20 AM
Lol, so courts are not slower because of the way it's played on them. I guess if you can walk on water, it's not water. Great thinking here.

Li Ching Yuen
03-18-2012, 12:21 AM
Lol, so courts are not slower because of the way it's played on them. I guess if you can walk on water, it's not water. Great thinking here.

It's all in the mind, bro.

Today, we are actually witnessing serve and volley. FROM THE BASELINE.

Think about it.;)

MaxPower
03-18-2012, 12:23 AM
The fact that average serve speed and average 2nd serve speed has gone up since 1991 is conveniently ignored in your analysis. Sampras served like a beast for his time. Today his serve speeds would be nothing special.

Isner/Karlovic and alike would bomb so many aces playing on the 1990 courts that it would look like a parody. The only reason the past 20 years has remain pretty constant IS THE GRADUAL SLOWING DOWN OF THE COURTS. How do you think it would have looked if they didn't slow it down? Players have become stronger and taller. Thats a fact. Check the player stats have changed over the the past 20 years too. Don't you think if the courts hadn't been slowed down like you try to argue that the numbers would be different?

but that's the fallacy of statistics anyway. You can make any case. In the end all you need for this is your own freaking eyes as a tennis fan. Go back watch vids from early 90s, late 90s, early 00s, late 00s and todays. Then you tell me about your myth. I tell you where to get glasses

abraxas21
03-18-2012, 12:30 AM
Hi all

This is my first port on this forum. Basically, I wanted to share some analysis I have done on the recently released statistics by ATP. Here (http://www.atpworldtour.com/Rankings/Top-Matchfacts.aspx) you can find some key stats on the top 200 players for every year since 1991 and this is the data I downloaded and used.

In this post, I try to shed light on the eternal debate about the surface speed and how it was slowed down lately. So I attach 3 charts (hopefully they will display well, still trying to figure out the system) that for each surface show these variables…

Percentage of first serve points won
Percentage of service games won
Percentage of points won returning first serve
Percentage of return games won

…averaged over top 50 players in each year. The pattern is fairly similar if I use all 200 players but some players ranked lower just don’t have enough matches so I cut at top 50.

These variables ***** for how easy it is to hold and break on a given surface in a given year, so we can get some nice insights from them. Here are my overall conclusions:

1) The slowing down of the Wimbledon grass is rather exaggerated. People say that the grass has been slower since 2002 (Federer stopped S&V in 2004) but since 2002 it actually became EASIER to hold serve and win points off the first serve! You can see an upward trend. At the same time, winning return games has become slightly more difficult, the line drops below 20%.

2) On hard courts the variations across years is small but the trends are striking: it has become easier to serve and more difficult to return if you compare early 1990s and late 2000s.

3) Clay has played fairly fast over the last 4 years.

This is obviously a very crude analysis but it goes contrary to the common belief that courts have been slowed down. So unless top 50 players suddenly became better servers and worse returners (unlikely, since the technology has benefited the defense game more), then we have to acknowledge that the courts are in fact playing quicker now.

This is rather surprising to me, other explanations for these results?

Later, I will post some analysis for players like Federer, Nadal, and Djokovic to see how their careers developed. Stay tuned!

PS. Any ideas how I can make those charts appear in the message and not as attachments? Thanks

hard and grass courts have definitely been slowed down. clay definitely plays faster though.

i can think of 2 explanations for the situation you describe:

1- racquet technology allows for faster serves and faster strokes. it doesn't just make improvements in the defensive department.

2- players are taller these days. even 8 years ago you'd think of a player like andy roddick as one of the tallest in the top 20. nowadays his height is pretty standard for a tennis player.

romismak
03-18-2012, 12:32 AM
The fact that average serve speed and average 2nd serve speed has gone up since 1991 is conveniently ignored in your analysis. Sampras served like a beast for his time. Today his serve speeds would be nothing special.

Isner/Karlovic and alike would bomb so many aces playing on the 1990 courts that it would look like a parody. The only reason the past 20 years has remain pretty constant IS THE GRADUAL SLOWING DOWN OF THE COURTS. How do you think it would have looked if they didn't slow it down? Players have become stronger and taller. Thats a fact. Check the player stats have changed over the the past 20 years too. Don't you think if the courts hadn't been slowed down like you try to argue that the numbers would be different?

but that's the fallacy of statistics anyway. You can make any case. In the end all you need for this is your own freaking eyes as a tennis fan. Go back watch vids from early 90s, late 90s, early 00s, late 00s and todays. Then you tell me about your myth. I tell you where to get glasses

Agree with you

Thatīs why i wrote slowind down was necessary for future of the game-of course i meant some slowing, not This slowing. Guys like Becker and Sampras were considered best servers of their eras. But serve speed, or ace per match, ace per game ratio wasnīt in the league of Isner, Raonic, Karlovic-2007. Now if we imagine this serves, with isner having normally FS% in 70s at old grass, old indoor carpets or outdoor HC all quick like Cinci than it would be real ace fest. Also big servers would be more vulnarable on return too, so serve statistics- service games hold, FS points won, big servers would be untouchable and decent servers would be in league of Isner, Raonic nowadays. Basically every non-clay match would have tie-break at least

MuzzahLovah
03-18-2012, 12:40 AM
Well interesting, but serving might just be a trend in the men's game and not be indicative of court speed. I mean the women play on the same courts and it certainly doesn't seem that they are becoming more serve oriented/dominant.

Nathaliia
03-18-2012, 12:57 AM
i agree with max power
however, following lately challengers on extremely fast surfaces (which you dont have on main tour anymore, like carpet), i have noticed the breaks are relatively easy to make when one of the players focuses to send an ace and sends 3 double faults instead.
Also, when someone flukes a good returning game, it's easy to send return winner. poor serving guy won't even land after the motion.

In Wolfsburg on carpet you could see instead of 7-6 6-7 7-6 many results like 6-2 6-3 because of the things i mentioned. These guys were pretty much one trick ponys going for one shot (because they can't keep a long rally anyway, on any surface). On a surface that is slightly slower a decently solid serve gives you advantage for 2-4 next shots when you fight for your point construction.

Li Ching Yuen
03-18-2012, 01:19 AM
Everyone's got this deluded idea that back in the day the courts were so fast that you couldn't return serve at all. Sure, there were players with great serves like Sampras and Ivanisevic that were holding serve pretty easily most of the time, but that doesn't mean that everyone else was doing the same. A fast court can't hide a shit serve.

One aspect that has changed over years in my eyes is clay. Today people hold serve much easier on clay than before, and this supports the point I was making in a similar thread: the average quality of serving has improved greatly over the years.

fast_clay
03-18-2012, 01:20 AM
when stats are able to quantify style call me

sexybeast
03-18-2012, 01:25 AM
Everyone's got this deluded idea that back in the day the courts were so fast that you couldn't return serve at all. Sure, there were players with great serves like Sampras and Ivanisevic that were holding serve pretty easily most of the time, but that doesn't mean that everyone else was doing the same. A fast court can't hide a shit serve.

One aspect that has changed over years in my eyes is clay. Today people hold serve much easier on clay than before, and this supports the point I was making in a similar thread: the average quality of serving has improved greatly over the years.

Players are taller in this generation, I remember Sampras seemed to be so tall in the 90s and he was the same height as Federer/Nadal who seems just avarage height and smaller than Djokovic, Murray, Del Potro, Berdych, Soderling, Tsonga, Roddick and other top players.

Li Ching Yuen
03-18-2012, 01:28 AM
Plus the slowness of a court affects most of all the rallies, not necessarily serve. It's ok if it's not 2 aces per game, but it's downright awful when people are retrieving Inside Out forehands like it's part of a drill especially with the kind of groundstrokes players can produce these days.

And because of the grittyness of the court, an approach shot is very much like a marriage: a very likely possibility to lose half of your shit.

Li Ching Yuen
03-18-2012, 01:29 AM
Players are taller in this generation, I remember Sampras seemed to be so tall in the 90s and he was the same height as Federer/Nadal who seems just avarage height and smaller than Djokovic, Murray, Del Potro, Berdych, Soderling, Tsonga, Roddick and other top players.

I don't have the link to that blog, that was posted here, but what you are saying is also a myth. Maybe some of the posters that saved it can help with a link.

Roddickominator
03-18-2012, 01:40 AM
The racquets(specifically string technology) and the balls are just as big a factor as court speed/bounce.

Isner and Karlovic would not be serving 50 aces a match with faster/lower bouncing courts and old strings. They would have to take pace off the ball to serve at a decent percentage, and they'd have to develop their net game further because they wouldn't be able to hang from the baseline. The serves would still be incredibly effective, but not tennis armageddon like some seem to think.

romismak
03-18-2012, 02:09 AM
The racquets(specifically string technology) and the balls are just as big a factor as court speed/bounce.

Isner and Karlovic would not be serving 50 aces a match with faster/lower bouncing courts and old strings. They would have to take pace off the ball to serve at a decent percentage, and they'd have to develop their net game further because they wouldn't be able to hang from the baseline. The serves would still be incredibly effective, but not tennis armageddon like some seem to think.


No old strings. Modern racquets on old surfaces- if they didnīt slowed down surfaces. I think you meant my coment one of those.

With faster surfaces, like old grass, indoor carpet and faster outdoor HC, Isner, Raonic or Karlovic-2007 would have more aces- donīt know how many more, but more, but other things will be more too - like service winners, unreturnable serves and easy points like 1 groundstroke, volley to win point after big serve. Because of lower bounce or faster court, there would be less reaction time on return. I donīt know how about rallies, simply there wonīt be this super long rallies lke Rafa-Nole, but 1st serves and service games hold i am sure that decent servers would have big serverīs stats and big servers would have better stats like nowadays.

fast_clay
03-18-2012, 02:11 AM
balls and string advancement, and then the ball reaction to hard court surfaces are the factor - largely the increase in surface area of the ball after a few games...

racquet technology peaked around 92-93

Benny_Maths
03-18-2012, 03:37 AM
Today, we are actually witnessing serve and volley. FROM THE BASELINE.



Such as this:

h6A_Tn4rGkQ

tennis_analyst
03-18-2012, 02:23 PM
So, what most likely happened is that there are better servers now than before. The courts have been slowed down but still not enough to prevent a rise in serve statistics.

It is still interesting though because I had a feeling that in the 90ies the game was serve-dominated and now it is return-dominated. Of course, there is a bias here since we see mostly top players. So watching Sampras and Becker in 1994-1996 and watching Nadal and Djokovic now would give you the feeling that it was easier to hold serve in 1990ies. But overall for the men's game now the serve wins you more points and the return wins you less. That was surprising to me, but as some of you wrote this has little to do with the court speed probably but just with players being taller and more powerful.