Biases At The Tracks

I've been meaning to learn more about horse racing and handicapping for awhile, but haven't really found the time to bone up on it.  Over at the Wages of Wins Journal, econ prof Steve Walters talks about some interesting inefficiencies that economists have discovered at the races.  Apparently, for awhile, there was assumed to be a longshot bias.  In other words, thrill junkies and dolts alike will place extra bets on the longshot horse, just cause.  So, because of the pari-mutual system, whereby odds are set by the bettors themselves, this creates value in betting on favorites.  But, you knew it wasn't so simple:

But not so fast. In a 2003 paper in Applied Economics, Russell Sobel (of West Virginia Univ.) and Travis Raines (of UNC-Chapel Hill) point out some, er, anomalies in this anomaly. It turns out that a few studies have turned up an opposite bias—over-betting on favorites and lower average returns there—that appears at odds with the “risk-lover” theory. And they think they have an explanation.

We don’t need to get into all the gory mathematical details, but Sobel and Raines (who actually studied betting on greyhound racing) think that the pattern of betting returns has a great deal to do with the level of sophistication of the bettors in attendance and their propensity to use information (or not use it). And they present some persuasive evidence on that score. For example, “longshot bias” tends to be a weekend phenomenon. Do risk-lovers only show up on weekends? Or is the weekend betting pool made up of more casual, less-informed bettors, who spread their bets more evenly over the field—perhaps because they’re using an “Aunt Louise’s age” betting theory?

Sobel and Raines present a lot more evidence in favor of their “information theory” of betting results, but one key takeaway is this: If you plan to actually make money betting the ponies, you have to pay attention to your fellow bettors as well as the ponies themselves. On days dominated by more serious, better-informed bettors, there appears to be a “favorite bias” on the simplest type of bets, especially in high-grade races. When the suckers are out in force, however, “longshot bias” will increase relative returns on the favorites.

Speaking of biases, I'm slowly getting addicted to this blog.

Comments

[wcw]

Now, I am not much of one for the ponies with the fat track take. However, I have bet on horses any number of times for fun, and my observations absolutely concur with the conclusion, "you have to pay attention to your fellow bettors as well as the ponies themselves."

Just as in poker or in the markets, watching other players bet tells you a lot. The board at the track follows the betting before each race and lets you watch as money flows into various bets. See if you can spot the smart money and the emotional bettors moving the odds. It's a lot more fun than just laying out your $2 to win if you have a nice narrative for what constituencies are thinking.

I say this knowing I pick badly, of course. I go to the track with my wife, who inherited her grandfather's ability to look at a horse's ass (never the face, never the jockey) right before the race and pick winners. As a rule, her wins cover my losses. She once picked eight dachshunds in a row (UC Davis Picnic Day; there are no dachshund tracks, more's the pity). They ran heats of four. There was usually only one obvious loser in each bunch (and once she picked that loser, "Noodle" and of course he won). What's 3 to the 8th?