I have long been a fan of the folks at www.tradesports.com and www.intrade.com. They brought a whole new approach to sports gambling. Instead of acting as a bookie and charging a vig, tradesports and intrade are clearinghouses. They provide a platform for bettors to come together and make deals amongst themselves. Since tradesports and intrade aren’t taking any position on the bet, they don’t face any risk or need a lot of capital. Consequently, they can charge commissions that are much lower than a bookie. Instead of the typical vig (which is 5% per bet), the fees at these clearinghouses are more like 1-2% per bet. It is puzzling to me why more of the betting money doesn’t flow their way. I suspect that they represent the way of the future. (And just like PartyPoker and eBay in my last post, these gambling clearinghouses have network externalities, so the bigger they get, the more attractive they will be to bettors.)
Another thing I love about the guys who run tradesports and intrade is that they have been very generous in providing access to their data for academics to study. There are now a handful of papers using their data. I am both proud and embarrassed to say that I was the first academic to get their data. Proud, because it is a great data source. Embarrassed, because it has taken me (and my co-author Ricard Gil of U-C Santa Cruz) four years to finally produce a paper using the data! In the meantime, a whole academic literature has emerged in this area, spearheaded by Justin Wolfers.
The timing is not bad, though, since the data we are working with are from the (2002) World Cup. We argue in the paper that soccer betting at intrade and tradesports turns out to be a great natural laboratory for testing theories of market efficiency. Unlike stock markets, where you never know who has what information when, in soccer almost all the information comes in the form of goals. No one knows a goal is coming very far in advance and all the traders see the goal happen at the same time. Thus, there is a very direct test of how effective the market is at factoring in new information. We also look at arbitrage opportunities (very few and far between) and find some bettor bias towards favorites (just as in other sports such as the NFL).