Woke up this morning to an interesting article at ESPN.com calling into question the true randomness of the draw process at the US Open over the previous ten years. Basically, they were tipped off by people that had been looking at the draws and felt that the #1 and #2 seeds on the men's side were drawing far easier first-round opponents than they should, on average. ESPN crunched some numbers (quite a bit, it appears) and came to the conclusion that not only were the tipsters correct, but that it was actually even worse on the women's side of the draw, almost to the point of impossibility.
The article and its research were done by ESPN's Outside the Lines group, generally known for breaking well-researched stuff on their Sunday morning show and not a place you want to be showing up if you're in the sports world. They're basically accusing the USTA of tinkering with and potentially fixing their draws so that the highest seeds get an early walkover, a pretty serious charge, and getting some response from some of the victims, mostly USTA wildcards like Scoville Jenkins, Coco Vandeweghe, and Devin Britton. The responses suggest that the players might actually believe they got screwed and definitely calls into question the integrity of the event staff.
I stewed over this possibility most of the morning while I had my Internet switched to a new provider and tried to figure out what the motivation would be if there truly was a conspiracy. ESPN doesn't really get into this too deeply (so I will), but the implication is that the 1 and 2 seeds get an early freebie to help them advance. So, let's assume the conspiracy is real and ask why it happened?
1. Money! USTA wants more money. It gets more money if the higher seeds advance. Plausible, I guess, but the seats are usually filled at the Open start to finish, particularly at night. I guess an argument could be made that TV wants higher seeds advancing, but that would probably play better on the men's side with its marquee names. Often, the top two seeds on the women's side were second banana to the famous names elsewhere in the bracket like the Williams sisters, Sharapova, etc. and there was no indication the conspiracy was attached to names, just seeds.
2. I don't know? I really can't think of another reason.
Neither of those are compelling, so let's also look at some other issues:
1. It's the USTA, so why didn't this disproportionately affect Americans? Well, it did, just the opposite way you would expect. They served up tasty triple-digit-ranked American wildcards to the top seeds. If anything, you would think a conspiracy would be afoot to hide some of these wildcards elsewhere in the draw, sneaking them through a round or two before throwing them to the wolves. They essentially guaranteed a beating on national television for some of their most promising young players. Sounds like a plan.
2. Aren't the draws public? I always felt like the majors had some sort of press conference where famous people drew names out of a hat. The ESPN article even shows a picture of Kim Clijsters doing just that and I definitely remember the snickering when John Isner and Nicholas Mahut drew a rematch at Wimbledon in front of the press corps. The article suggests that a computer, with the same program as the Australian Open, pulls the draw at a push of a button, but even so, there are witnesses and nobody has previously questioned the integrity of this.
3. It was only two players, half the years, and the advantage only appeared in the first round. Why didn't this extend to #3 and beyond? Why was it "super obvious" some years, then disappear in others? Why wouldn't they also rig other parts of the draw to weight the easiness through to the later rounds. I looked at later rounds and there was no appearance of funny business in how the seeds were placed or really even into the second round. It was pretty much confined to what the article suggested.
It all sounds like an awful lot of work with very little that can go right and a lot that could go wrong. So, with that, I tried to see what else I could find, see if there was a pattern to the math that couldn't be explained away by randomness. One thing I've found from my poker playing days is that when a cheating scandal happens, once you find it, you realize how incompetent the people hiding it really were. Faking data in research or rigging a draw randomly is not easy to do and once you know there's something fishy, stuff starts to jump out all over the place.
I wanted to start with the lowest hanging fruit, so I broke up the suspicious pairings into groups. The victims were either field, qualifier, or wild card, since seeds can't meet until round three. The US Open has eight wild cards, 16 qualifiers, 32 seeds and the rest of the field. In any given US Open draw, a seed (1's and 2's included) would be expected to face a wild card one in twelve tournaments and a qualifier one in six. The remaining 75% of the time, they should face a generic middle of the road pro in the top 100. The qualifiers would generally come from the next 100, and the wild cards could be anyone, but most often were young American players that would not have made the main draw otherwise. Drawing a wild card in the first round would be a lucky boost to any player, let alone high seeds.
On the men's side, of the 20 opponents the 1 and 2 seeds faced over the last decade, they drew qualifiers three times and wild cards five times, against expectations of 3.3 and 1.7, respectively. On the women's side, they drew qualifiers six times and wild cards an astonishing seven times, 300% greater than expected. This seemed like a good place to start. Was the USTA manipulating wild cards to place them in harm's way?
As kind of a reverse check, I decided to see who the wild cards were drawing on the men's and women's side to see if there was a pattern. There were eighty men's and eighty women's wild cards over the ten years. It would be expected that of the eighty, 20.2 would draw seeds, 4.4 would draw other wild cards, 10.1 would draw qualifiers, and the remaining 45.3 would draw the field. This where it gets a bit interesting and I thought I might be onto something.
On the men's side, 44 drew the field, 29 drew seeds, 7 drew qualifiers and 0 drew another wild card. On the women's side, 47 drew the field, 24 drew seeds, 7 drew qualifiers, and 2 wild cards drew each other. In both cases, wild cards were generally more likely to face a seed than they should and less likely to face each other or a qualifier, more so on the men's side, bolstering ESPN's argument.
Now that the numbers appeared to be pointing toward something, I figured the final straw would be catching some chicanery in the draws. The funny business seemed to start on the women's side from 2001-03 (with a men's cameo in 2001), shift to the men from 2004-07, then shift back to the women in the last three years.
When draws are announced, they don't know the qualifiers and they tend to be a more homogenous group, so there's not a whole lot of manipulation possible there. The wild cards, however, are mixed in nicely with the rest of the draw and were definitely showing up more often than they should. My hypothesis was that somehow, they were trading wild cards in for other players who created bad first round matchups for the top two seeds, either a dangerous floater in the field or maybe a wild card that would be expected to advance.
To test this, I looked at the women's fields from 2001-2003 and 2008-2010, where out of 12 opportunities, the top two seeds somehow drew seven wild cards (remember, one in 12 chance), five times to the top seed. This is where the conspiracy starts to fall apart. From 2001-2003, only one wild card was any threat to the field, and Martina Hingis was also seeded 9th, far away from any 1 or 2. In 2009, Kim Clijsters won the tournament as an unseeded wild card, but she faced Venus Williams in a quarterfinal that certainly would have been preferable in a later round. Along this route, I thought maybe there was a dangerous floater that got switched out, but really only Ana Ivanovic in 2010 qualifies and she was matched against Clijsters in a quarterfinal as well.
Once I didn't see anything on the women's side, I kind of doubted there was anything on the men's. Just a fluke that the 2 seed drew a wild card four years in a row. The types of things that look really bad on the surface, but no real advantage ever appeared to be gained. Finally, I just gave in to Occam's Razor.
Overall, I applaud ESPN for looking into this and they do appear to have stumbled onto a wacky statistical fluke. Maybe there even is a bug in the program that steers wild cards and qualifiers to 1 and 2 seeds. I don't know, but there doesn't appear to be malice. On the plus side, all parties will be vigilant when the next set of draws comes out in a couple of weeks and maybe the USTA will be more open with the process going forward.
No comments:
Post a Comment