Anyone for tennis?

In the lull between the European Championship semi-finals and final, I looked around for something statistical-ish to do and noticed that some small, London tennis tournament was on.

Tennis seems like a perfect thing for stats. Like baseball, it is easily delineated into short actions at the end of which the game moves a small, small step along. Shots tend to be hit from the baseline, deep, or close to the net; central, from the left, or the right; forehand or backhand; points scored by winners, or errors.

I sat down after watching the Women’s final (I wanted to bask in Serena’s tennis glory, and Kerber’s play wasn’t bad either) to have a go at coding the how points were won. To keep things (relatively) simple, I was just looking at the last shot of the point – the winner, or the error (forced or not), or an ace. I won’t go through everything, but here’s my stats scoreboard.

Tennis stats scoreboard

I should note straight away: my stats don’t match up with the official ones. After the first set, I had winners at 16-6 and the BBC broadcast showed 24-6; unforced errors for me 16-6 again (coincidentally), but 14-5 by the BBC. Part of this may be that I spent the first couple of games classing unreturned serves as unforced errors before I realised that I’d included “unreturned serves” in the key I’d knocked up before the match. I’m not sure where the discrepency on winners comes from – maybe they were including aces.

But it’s interesting to be able to break things down beyond what the stats you get on the broadcast. Williams’ winners were pretty evenly split across her forehand and backhand, but Kerber’s were pretty clearly forehand based. Williams also made a lot more unforced errors than her opponent.

I’ve also toyed with the idea of looking at “winners [minus] unforced errors” as a point of interest. Using this, Williams has a net +5 points on her forehand but -7 on her backhand; for Kerber this is +5 and -5. You’d have to look at a much, much larger sample of stats to see if this might be actually useful in any way, but it seems interesting.

Both players (according to my coding) forced the same amount of errors from their opponents, and it’s interesting to see that most of the errors Williams forced were to Kerber’s forehand, and the other way round for the errors Kerber forced from Williams.

Where Williams was making a lot of unforced errors, she was making it up with winners and ‘serve points’ (aces and other unreturned serves; I think the BBC broadcast had these down as ‘Unreturned serves’). 26 serve points. 26! That’s over a third of all points she won, directly from her serves. But it also should be noted that while Kerber didn’t have an ace, she did have 10 unreturned serves, which seems a decent haul.

Anyway, I could break things down further – and might in future – regarding shots won going cross-court or down the line or something like that, but won’t right now. Presumably, there’s some data companies out there that have access to stats similar to these (and a lot more extra – position and aim of serves is an obvious place to start), but I haven’t seen things like this in the public sphere before.

I imagine it would be hard for someone on their own to do something meaningful with tennis stats like this due to the majority of matches not being available to watch (live or via sites on the internet), but there’s a chance (probably slim) that I might return to something like this during another Grand Slam.

Hope you enjoyed a taste of something different. There are probably tennis stats purists out there grumbling about me coming along during Wimbledon fortnight and hogging their court. Sorry, I’ll be gone soon, see you next year.


Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )

Connecting to %s