Home court advantage in tennis
There’s some sort of home-court or home-field advantage in every sport, so why not tennis?
To test the hypothesis, I looked at all ATP-level matches from 2009. I excluded qualifiers and matches decided by a walkover or retirement. I also excluded matches where one of the players received a wild card into the tournament.
Limiting the field to matches where one player was on his home country’s turf and the other was not, that leaves us with 666 matches. (Maybe I should stop now.) All else equal, if there was no home-court advantage, we’d expect 333 wins from the home player. In the event, the home player won 358, or 53.8 percent of the matches.
Is all else equal?
Over the course of nearly 700 matches, we might expect everything to wash out. We would be wrong.
Using my simple algorithm for predicting the outcome of matches using ATP ranking points, we can calculate the likelihood that the home player would have won each match. For instance, when John Isner (565 points) played Christophe Rochus (1248 points) at Indian Wells last year, the ranking points gave Isner a 29.5 percent chance of winning. It turns out that Isner beat the odds and won the match.
As I said a moment ago, if all else was equal, the home player would have won 333 matches. But all else wasn’t equal. The average visiting player was considerably better than the average home player. Based on my calculations for each match, home players should have won only 288, or 43.2 percent. Remember, they won 358, or 53.8 percent.
I’m not sure what’s more surprising: the apparent magnitude of home-court advantage, or the relative weakness of the home players. (Remember, I excluded wild cards, so these are all guys who earned their way into the draw.)
Let’s spend a moment on the latter. One possible explanation is home-court advantage in qualifying. For instance, Jesse Witten (197 ranking points, about 250th in the world) fought his way through qualifying to the 3rd round of the US Open last year. There are even more extreme examples with the occasional player who is wildcarded into qualifying, and makes the main draw. An example is Milos Raonic, who with 40 ranking points (about 700th in the world) qualified for the main draw in Montreal.
Another possible explanation is that mid-level players (think 50th – 150th in the world) are more likely to play closer to home when given a choice. For many weeks of the year, there are multiple ATP tournaments, and in weeks where there is one (non-major) tourney, many players take a break. So an event in France may be more likely to attract French players to fill out the draw, even if the mix of seeded players is similar to that of non-French tournaments.
As we’ve seen, home players had a huge advantage in ATP matches last year, winning about 24 percent more often than we would otherwise expect them to.
One way to incorporate that in my algorithm is to multiply the ranking points of the home player by about 1.7. In the Isner-Rochus example above, that adjustment would have given Isner a 43 percent chance of winning.
To take another example, without factoring in home-court, Sam Querrey had a 53.3 percent chance of beating Stanislas Wawrinka yesterday. Considering home-court, that number is 67.2 percent. (Of course, Querrey lost.)
One big objection
A final thought. As I’ve mentioned in just about all of my posts on tennis, surface is not considered in any of my analysis. That’s a function of the ATP rankings, which also do not consider surface.
Home-court advantage may in part reflect something literal about the “home court”–that it is the player’s favored surface. Many more European players grow up on clay courts, and sure enough, that’s where most of the clay-court tournaments are. Querrey or Isner is much more likely to win a hard-court match than a clay-court match, whether it’s in Washington or Bangkok.
So when a player is competing in his home country, he is more likely to be playing on his favored surface. That doesn’t erase the validity of my results, it just suggests that they could be broken down further. Maybe a quarter of the home-court effect is really a surface effect, while the other three-quarters is due to the home crowd and home cooking. Eventually, we’ll find out.