1. Court speed – Act One: XC marks the spot

A widely held opinion about professional tennis in the last few years is that court speeds have been sufficiently harmonised so that the pace of general playing conditions has converged. Examples of this opinion can be found here and here.

How true is this on the ATP Tour? And how can we measure it?

Measuring first.

Let’s assume that on quicker surfaces it is easier to hold serve and that the score lines of sets are accordingly tighter: more tie breaks, 7-5, 6-4 set score lines. We can as a result measure the pace of courts according to a calculation known as cross-courts, a betting market from Sporting Index and others. Under the cross-court markets (hereafter known as XC), the two game scores in each set are multiplied together. So a set that finishes 7-6 scores 42 XCs; 6-4 will score 24 XCs. A match that finishes 6-4 3-6 7-6 will score 84. The higher the XCs, the faster the court: not always, but typically.

Measuring XC is a more reliable indicator of court speed than the total games of matches (see Annex 1). There are 23 games each in matches that finish 7-6 6-4 and 6-1 2-6 6-2. But experience tells us which one we expect from a grass court and which from a clay court. In the former score line XC will score 66; the latter 30.

Ideally, analysis would draw on data from the 1990s until 2013, but absent many a long night in front of the computer, this analysis looks at data between 2008 and 2013. Notes on the data:

  • The data comprise all ATP tour level matches from 2008 to 2013 to date, with the exception of Grand Slams, Davis Cup and Olympics matches (which are 5 set or extended 3 set matches).
  • All the matches have to have been completed (ie no retirements, no withdrawals).
  • All matches are “vanilla” 3 set matches that culminate if necessary in a third set tie break.
  • A total of over 12,000 matches.

The results

Average XC by year by surface

Surface

2008

2009

2010

2011

2012

2013*

Indoor

55

55

55

55

55

57

Hard

52

54

53

52

52

54

Clay

52

52

50

49

50

50

Grass

58

56

55

54

57

55

* To date in 2013

  • XCs for indoor and grass courts are consistently higher than for hard courts and clay courts.
  • Clay courts have recorded the lowest XCs each year since 2009.
  • In other words, and despite other non-dependent factors that may contribute to the differences above, it is likely that the different surfaces do have different speeds, sometimes up to 10 per cent.
  • Although the differences are not large and the surfaces are not immune to changing position in the hierarchy, they are consistent for the six year period under analysis.

So the next time someone casually tells you that the court speeds of different surfaces are the same you may wish to point out that there is no evidence in the last 5 years to support that.

If the analysis above, supporting the heterogeneity of court surfaces, could be termed “macro-empirical”, there is interesting stuff to be found here that comes at the issue at a micro level.

This was Act One.

Act Two, and next on Cleaning the Lines:

  • Which tournaments offer the slickest courts?
  • And how to control for the likes of Isner and Raonic who live in a cross court town called Anomaly, or Memphis maybe.

Annex 1

This was how the data turned out if instead of using XC as our data point, we had used the average number of games for a match. Across a large data set, there is not much discernible difference between the surfaces (more can be seen at one decimal place); which underscores the importance of looking at XC in this instance.

Average games for a match by year by surface

Surface

2008

2009

2010

2011

2012

2013*

Indoor

23

24

23

23

23

24

Hard

23

23

23

23

23

23

Clay

23

23

22

22

22

23

Grass

24

23

23

23

24

23

* To date in 2013

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s