clock menu more-arrow no yes

Filed under:

2014 Wimbledon championships: Breaking down the men's and women's seeds

How do the seeds at this year's Wimbledon compare to their predictive rankings?


Wimbledon is the only Grand Slam to make explicit surface-based adjustments to its seedings, albeit for the men only. The formula is a little complicated, but it does give extra weight to grass results in an attempt to provide a little more balance to the draw.

I think it's great that they make an effort to make surface-specific adjustments, but how well does their specific formula work? We can test that pretty well by comparing the tournament seeds to the AB grass ranks of each player. Here's how they stack up against one another:

This graph exposes two major flaws in Wimbledon's formula.

Jerzy Janowicz has an ATP rank of 23 and a tournament seed of 15; his boost comes from a lot of grass bonus points earned during his run to the 2013 Wimbledon semifinals. That run, however, was greatly assisted by early exits from Federer and Nadal, both of whom Janowicz was slated to face in succession.

Wimbledon's grass bonus points don't look at how hard players actually had to work to earn their points. The net result is that the draw and upsets luck play an even bigger role than normal in influencing the seeds.

Complete Wimbledon draw

To be fair, the grass bonus points still do some good, tightening up the bracket in select spots (the rest of the top 16 looks pretty aligned). But do the side effects outweight the benefits? How many "undeserved" seeding slots would you allow to adjust the rest of the field? Adjusting based off of predictive rankings instead of bonus points solves this problem pretty nicely.

The other flaw is highlighted by Marcel Granollers nabbing a seed despite being worse on grass than half the field. He's seeded because his base points (i.e. non-grass) are high enough that they overcome his practically zero grass points. The problem with bonus points, especially ones that are based on a player's best results, is that losing a lot on grass is treated the same as not playing on grass.

Forget fancy stats for a second: Granollers is 1-7 on grass in the last four years. But the surface-adjusted ranks still award him a seed. Can you say with a straight face that your surface adjustments are working well when this happens? Predictive analysis says Granollers is much worse on grass, but the surface adjustments don't penalize him at all.

The women's seeds, on the other hand, don't have any surface adjustments. I have no idea why. Here is the women's graph.

Eyeballing this graph next to the men's shows a little more scatter around the black line, which shows what happens when you have no surface adjustments at all.

There will be a lot of dangerous floaters (Sabine Lisicki, Ekaterina Makarova, Venus Williams) that the top seeds might have to face in the third round, which means the luck of the draw will once again be huge. There's no one in the seeded group that passes the "worse than 50 percent of the field" test (i.e. the Granollers Test), but the alignment definitely isn't as tight as on the men's side.

We'll see how the draw shakes out Friday; it will go a long way in deciding everyone's chances.

Surface adjustments are always a good thing, but the way Wimbledon does it is a much smaller step in the right direction than the All-England Club probably realizes. Their process could be improved painlessly by adjusting with predictive rankings instead of the current ATP point system.