Introducing the TeamRankings.com Wimbledon Ratings and Wimbledon Predictions
June 24, 2011 - by Austin Link
Wimbledon started earlier this week, and we’ve decided to create a version of the power ratings to help predict the winner. It’s a little late getting up, but we are now unrolling on the site men’s Wimbledon ratings and Wimbledon women’s rankings, along with Wimbledon predictions and Wimbledon women’s predictions as of the start of the tournament. Here’s a quick explanation of both.
The Wimbledon rankings are predictive power ratings, and use results from all of the matches since January 1st, 2010. The ratings themselves don’t correspond to margin of victory as they do for other sports, but they allow an easy way to predict an individual’s chance of winning a match.
For the Wimbledon women’s rankings, if a player with rating A plays a player with rating B, their chance of winning the match is A/(A+B). For the men it’s similar, but the formula gives their chance of winning a set, making it easy to adjust for best of 3 or best of 5. Looking at how these ratings would’ve fared, they correctly picked almost 70% of game winners.
On the ladies’ side, the first thing that stands out is the positioning of the Williams sisters. Because they play so rarely, they sit much lower in the WTA tour rankings (25th for Serena Williams, 30th for Venus Williams) than they should based on ability. In fact, the power ratings show that Venus is the fourth best player in the world, and Serena is number one. Also interesting is that the results before her retirement place Justine Henin as the second best women’s tennis player alive. Talk about going out on top.
Looking at the men’s side, things are more conventional. In fact, the top 5 of our Wimbledon rankings matches up exactly with the ATP rankings. The ratings make it clear that there’s a distinct top 3 in the men’s tennis world, and then there’s another ten or so guys who have an outside chance to win any grand slam tournament. These rating differences are made even more striking when looking at each player’s chances of winning Wimbledon.
The Wimbledon women’s predictions from the start of the tournament show that the three most likely winners were Serena Williams (17.3%), Venus Williams (15.3%) and Caroline Wozniacki (11.2%). Beyond just having the two best winning chances, though, the Williams sisters (especially Venus) are also expected to outperform what the sports books would expect. Going into the third round Serena’s odds are up to 20.4%, and Venus’s are at 14.9%.
Our best value against the odds was Agnieszka Radwanska who has already dropped from the tournament, but we also expected Na Li to struggle more than others thought and she also bowed out early.
Going into the men’s tournament, there was a clear divide between the top 3 and the field. The Wimbledon predictions showed that Rafael Nadal was expected to win 28.3% of the time, Novak Djokovic 23.6%, and Roger Federer 20.9%.
It’s interesting to note that the field includes Andy Murray, somebody who was predicted by many to win Wimbledon. We think he still has a shot, but going into the tournament we just gave him 5.4% odds, and after two rounds things are the same (he won a couple of games, but is now facing harder than expected competition).
We did find one extreme value pick though: David Ferrer. Going into the tournament we gave him a 3.3% chance of winning the tournament even though he was getting up to 200-1 odds. He now has a 3.9% chance of winning, getting a boost by playing the worst remaining player (Karol Beck) in the third round.
Performance So Far
How did the predictions fare in the first two rounds? If anything, better than expected. The women’s predictions were correct on 46 of the 64 first round matches, and 22 of the 32 picks to make it to the third round did. 22 of 32 exactly mirrors how the seedings fared. For the Men it got 49 of the 64 first round matches, and again correctly selected 22 of the players who made it to the third round. Only 20 of the seeded players made it on the men’s side, so we actually outperformed the seeding. It’s all a little rough still, but the preliminary number look pretty good. We’ll keep it updated for the rest of Wimbledon, and possibly even going forwards.