This is the 2nd and final part of a 2 part series on our analysis of the court speeds. You can read the 1st part here.

The ace count is a poor way to measure court speed. Why do I say this? Well, there is a general reason, and a specific reason. In general, the serve is the only time in the rally at which a person hits the ball before the ball touches the court. Thus, the serve is less affected by changing court speed than is any other shot in the game. For any other shot, you have longer to set up the shot if the court is slower, and less time if it is quicker. When you serve, the court does not directly affect your set-up time. Thus if players are hitting the ball harder, then that harder hitting will be an intervening variable between number of aces and court speed to a much larger extent than will harder hitting on any other shot.

More specifically, with regard to Wimbledon, the argument that the court has slowed is somewhat complicated. In late 2001, the tournament decided to change the composition of grass and started using 100% ryegrass (formerly used mostly for race tracks) for more durability, and started packing the soil more densely. Starting in 2002, then, the court was different. Those who believe that the court has slowed down argue that the new grass made the ball sit up slightly more, which made it harder to attack the net, because net-rushers rely on the ball fading after the bounce to prevent the opponent from retrieving it and hitting a passing shot. The result of this difficulty in attacking the net is that the court gets chewed up in a different way than it used to.

In the past, the area around the T at the junction of the service boxes was very chewed up. Now, because the players don’t come to the net as much, the front half of the court is almost pristine, whereas the baseline is much more chewed up than before as the tournament goes on. As a result, the back half of the court becomes less and less slick so that, in the second week, it is not really a grass court at all because there is very little grass left. It would be more accurate to call the courts of the second week mud courts, and so very much like a clay court. (Clay courts are actually made of crushed brick). But those changes impact upon deep volleys and groundstrokes much more than they do upon serves, because the serve lands in the almost pristine front half of the court that is still slick and thus slides through the court. Thus, we should expect the impact of the new courts at Wimbledon in particular to be less marked on the serve than on any other shot. That means that it comes as no surprise to know that the ace count isn’t particularly changed because there is at best a very loose relationship between the ace count and the speed of the grass.

In general, a better way to measure the court speed would be to measure the average length of rally, and it is undeniable that rally length has increased markedly at Wimbledon, and the US and Australian Opens, over the years. In the 1995 Wimbledon final, the longest rally was five strokes. Now, it’s not rare for the average rally to be five strokes or more. It’s true that this doesn’t distinguish between court and ball/racket/strings, but as I said above, without a time machine there is no way to distinguish between these things.

In assessing the court speed, we do therefore have to allow a role for expert knowledge. Experts can be wrong, but we implicitly rely on them at many points in our everyday life, and the best evidence we have is often that of an expert. And when we ask the experts, we find pretty strong support for the claim that court speeds have been reduced, or more precisely that they have been homogenized, so that grass and hard courts play much more like clay than they used to. We couldn’t get much more of an expert than Serena Williams, who has won the US Open before the courts were changed in 2004 – with the explicit intent of slowing them down – (in 1999 and 2002), again afterwards (in 2008), and again after they were, in the views of many fans and players, slowed down again in 2011 (2012 and 2013). At the most recent US Open, Serena was asked how she prepared for the transition from clay to hard courts, and replied by saying that she did not really need to make a transition any more because, over the course of her career, the US Open courts have become so slow that they play like a fast clay court. Back in 1999, when the Arthur Ashe Stadium court was nearly new, the US Open courts were widely held to be among the fastest courts in the world. A plethora of other players have also commented on this change, while the commentators on Tennis Channel discuss it almost every week.

The appeal to expert testimony is bolstered by the fact that, if we look at the ace statistics more closely, we see that there is good reason to believe that what the players say reflects something real about the courts. If you look at the number of aces served at the Australian Open every year, you’ll see that there was a huge spike in the number of aces in 2000. Why might that be? Well, in the late 1990s, rebound ace, the surface used by the Australian Open between 1988 and 2008, was thought to be getting slower and slower. The top Australian players (Pat Rafter, in particular) didn’t like that and complained about it. So, for the 2000 tournament, the organizers gave the court a fresh coat of paint designed to make it faster. Several players commented in 2000 that the courts were playing as fast as any courts in the world. And that is the year in which the number of aces took a sharp turn upwards. Likewise, players commented that the US Open’s Arthur Ashe stadium, which opened in 1998, played much faster than did the old Louis Armstrong stadium, and sure enough the number of aces served at the US Open rose quite markedly in 1998, 1999, 2000, and 2001.

Additionally, the claim that the number of aces at each Slam isn’t converging seems to me to be based on reading 2012’s evidence alone, which raises all sorts of issues about sample size. In 2007, for example, there was clearly a convergence between Wimbledon, the US Open, and the Australian Open. Finally, although the ace count has risen slightly since 2000, it seems to me that the rate of increase since 2000 is very much less than it was between 1990 and 2000, which is consistent with the idea that the court speed is decreasing, as the slight increase in number of aces could be due to the use of larger rackets.

Statistical data cannot settle the question of court speed, and we have to combine use of it with appeal to the testimony of experts. In short, we have to appeal to judgment not science in measuring aggregate concepts that require interpretation. Doing that, I have said, provides good reason to believe that tennis courts are generally getting slower and less distinct from each other. Why should that matter, if the slower courts are producing a golden age in tennis? Well, the problem is that not everyone agrees that we’re in a golden age or that, if we are, it is likely to last. Tennis has fair claim to being the most truly global of individual sports and among the most global of any sport. That diversity is matched by a great diversity of playing styles.

As more and more courts play in a similar fashion, diversity of playing style declines, and the major finals increasingly become about grinding the opponent down. This is no bad thing in itself, but it lacks variety. Variety, as we all know, is the spice of life, in entertainment above all. If all the tennis tournaments are going to play in the same way as each other, it becomes hard to see why we should have four different major tournaments each year, rather than just a season-ending world championships. I doubt that many in the tennis world have the appetite for so radical a restructuring of the tennis circuit.

–Toby Reiner