What happened to Wired's professional bettors?
August 20, 2008 8:45 AM Subscribe
Remember this Wired article about "computer teams" analyzing horse races in Asia? What became of that? Were the theories sound? What exactly were they doing? Are there any other articles, academic or otherwise, about modeling horse races?
The article sparked my interest, and I'm interested in the mathematics and theory behind this, yet I'm unable to find any references to statistical betting beyond the Wired article. In fact the names in the article don't even Google well.
I have a hard time believing that there are betting quants earning 24% returns in Hong Kong somewhere. Did these "funds" eventually go bust? Did they exist at all?
The article sparked my interest, and I'm interested in the mathematics and theory behind this, yet I'm unable to find any references to statistical betting beyond the Wired article. In fact the names in the article don't even Google well.
I have a hard time believing that there are betting quants earning 24% returns in Hong Kong somewhere. Did these "funds" eventually go bust? Did they exist at all?
Best answer: I haven't seen the Wired article, but there is a German paper that I've seen that might be of interest -- Risk Love and the Favorite-Longshot Bias: Evidence from German Harness Horse Racing.
"Abstract:
Empirical studies of horse race betting in the U.S., the UK, and Australia have established the so called favorite-longshot bias. Studies find that on average, bets on longshots lose much more than do bets on favorites. This means that longshots are overbet and favorites are underbet. By using a large data set of pari-mutuel harness horse races, we show that the favorite-longshot bias exists in Germany as well. We provide evidence that there is bias not only for simple win bets, but also for other types of bets as well, and that the bias is a time- and track-invariant phenomenon. The bias is consistent with the assumption of the (local) love of risk of the betting audience. "
Full citation: Winter, S., Kukuk, M., 'Risk Love and the Favorite-Longshot Bias: Evidence from German Harness Horse Racing.' Schmalenbach Business Review, Vol. 58, October 2006
Not really my area of interest (I seem to lose my money spread betting on financials), but I've known a few guys working in The City who are pretty cagey about their "hobby". And there was a Russian guy I worked with at Dow Jones, who left to pursue his strategy full time. Apparently it involved a carefully selected combination of favourites and long shots, and he felt that his full time job was cutting into his track time.
I'll try to grab a URL for the paper (I've only got a .pdf that was emailed to me).
posted by Mutant at 10:03 AM on August 20, 2008
"Abstract:
Empirical studies of horse race betting in the U.S., the UK, and Australia have established the so called favorite-longshot bias. Studies find that on average, bets on longshots lose much more than do bets on favorites. This means that longshots are overbet and favorites are underbet. By using a large data set of pari-mutuel harness horse races, we show that the favorite-longshot bias exists in Germany as well. We provide evidence that there is bias not only for simple win bets, but also for other types of bets as well, and that the bias is a time- and track-invariant phenomenon. The bias is consistent with the assumption of the (local) love of risk of the betting audience. "
Full citation: Winter, S., Kukuk, M., 'Risk Love and the Favorite-Longshot Bias: Evidence from German Harness Horse Racing.' Schmalenbach Business Review, Vol. 58, October 2006
Not really my area of interest (I seem to lose my money spread betting on financials), but I've known a few guys working in The City who are pretty cagey about their "hobby". And there was a Russian guy I worked with at Dow Jones, who left to pursue his strategy full time. Apparently it involved a carefully selected combination of favourites and long shots, and he felt that his full time job was cutting into his track time.
I'll try to grab a URL for the paper (I've only got a .pdf that was emailed to me).
posted by Mutant at 10:03 AM on August 20, 2008
Best answer: Selecting The Best of K Multinomial Parameter Estimation Using SPRT. (.pdf)
Efficiency & Probability In Exotic Bets.
posted by Floydd at 11:11 AM on August 20, 2008
Efficiency & Probability In Exotic Bets.
posted by Floydd at 11:11 AM on August 20, 2008
Found a free source and upon review, the references have lots and lots of papers citing market efficiency at the horse races. Interesting topic / question ...
posted by Mutant at 11:35 AM on August 20, 2008
posted by Mutant at 11:35 AM on August 20, 2008
Response by poster: Sorry I didn't get back to this question sooner, I had spent a lot of time thinking about this for whatever reason, and wanted to read the papers before getting back here. I don't think anyone was waiting with baited breath ...
I spent the last week in Vegas and can't stand the damned continuous shuffle machines they use for blackjack now, how is that fair? Blackjack isn't infinite decks it is 3-5 decks. Anyway, I would get tired of walking around, and as my friends hit the tables, I spent countless hours at the sports gambling room of the Bellagio. Anyway I began to think about parimutuel betting and horses as securities, then whether horses could be valued, etc. etc. Anyway after reading the papers, I'm no longer so sure that sports betting is such a gamble, at least with regards to horse racing. I would be very surprised if sports betting had worse chance of ruin than the stock market (or more formally, a more wild form of randomness than stocks exhibit). Of course the scalability is quite limited.
Anyway, from my cursory research, it would appear the very in-the-money and very out-of-the-money bets appear the most profitable over the long run, or the long-shot/favorite bias that mutant talks about. I think that's also a good option strategy, philosophically, but that's an aside. Anyway I'll see if I can get some empirical data to verify this, I'm having a hard time gathering data from horse races, nowhere near as accessible as stock data. So sad, if I find anything I'll post it here.
posted by geoff. at 10:19 PM on September 4, 2008
I spent the last week in Vegas and can't stand the damned continuous shuffle machines they use for blackjack now, how is that fair? Blackjack isn't infinite decks it is 3-5 decks. Anyway, I would get tired of walking around, and as my friends hit the tables, I spent countless hours at the sports gambling room of the Bellagio. Anyway I began to think about parimutuel betting and horses as securities, then whether horses could be valued, etc. etc. Anyway after reading the papers, I'm no longer so sure that sports betting is such a gamble, at least with regards to horse racing. I would be very surprised if sports betting had worse chance of ruin than the stock market (or more formally, a more wild form of randomness than stocks exhibit). Of course the scalability is quite limited.
Anyway, from my cursory research, it would appear the very in-the-money and very out-of-the-money bets appear the most profitable over the long run, or the long-shot/favorite bias that mutant talks about. I think that's also a good option strategy, philosophically, but that's an aside. Anyway I'll see if I can get some empirical data to verify this, I'm having a hard time gathering data from horse races, nowhere near as accessible as stock data. So sad, if I find anything I'll post it here.
posted by geoff. at 10:19 PM on September 4, 2008
This thread is closed to new comments.
Here are some Gambling Times columns by William Ziemba.
posted by Floydd at 10:00 AM on August 20, 2008