I found this book at the local public library back in 2008.
Skiena writes about how he became a profitable jai alai bettor with some historical data. It's an interesting book if you don't know much about gambling or modeling.
So I haven't read it, but if it works, why is he sharing it?
I had a prof in grad school (http://www.umanitoba.ca/faculties/management/faculty_staff/a...) who worked out an arbitrage system for horse racing. Other than publishing a few obscure papers, he kept it to himself. Because it wouldn't actually work if he shared it to a mass market.
Just asking, before I go and read the page in depth.
At the time of writing, the market in question--jai alai--was so small that one could not bet it professionally. The action and the number of sites where it was played dried up.
That's a fair point. OTOH a lot of people who take the trouble to write a book would like to see it in print, and almost all publishers forbid you from releasing the manuscript in alternate forms once it has been published. Also, for somebody with his skill set, there are many shorter paths to riches than writing MAA monographs.
If you just read even the preface to this book ( http://books.google.com/books?id=UvWGgaE4ZA8C&lpg=PP1... ), it is obvious that his motivation for writing it has everything to do with a lifelong fascination with mathematical modeling and sincere desire to share that with others, and nothing to do with making money from gambling returns. The gambling was merely meant to put his model to the test in a way that penalized failure.
It's all right there in the book. But hey, why consult the text when we can all sit around grousing online about what his intentions might have been?
What do you mean by "work out" an arbitrage system? Arbitrage is arbitrage. Do you mean he wrote a system to find arb opportunities and place the bets automatically?
Arbitrage can mean buying a package of items and selling another equivalent package of items. So in horse racing (vague 20 year old memories here), it might mean betting on the a certain proportion of your money on the trifecta and other proportions on each of the horses not included in that trifecta. So you'd have do the math for each combination of 3 horses that could appear in the trifecta, and the corresponding payoffs. I doubt he had something that placed the bets automatically (because betting wouldn't have been on the internet) but he had some software to take the odds and identify arbitrage opportunities.
Whether parimutuel bets or your more traditional sports bookie, betting is basically like any other market. If you can predict it's outcome better than the rest of the market - you can beat it.
As someone else once told me, betting on sports is just like playing the stock market, except the market is smaller, with far fewer variables, and therefore easier to predict.
It's just like people who have stock trading systems in some market that are unbeatable - they very well may be true - but only for a limited time, as their trades themselves affect the market.
Arbitrage is another thing - on one hand, it may seem dishonest - it's looking for no-loose opportunities and taking advantage of them - on the other hand, an efficient arbitrage system keeps markets fair and moving. (In a perfect market, there would be no room for arbitrage.)
I read this book ages ago. It's not a "how-to" manual, as some people here seem to think. It's a mathematical adventure, how his childhood lucky bet set him on a particular obsession to see if such betting was amenable to mathematical analysis and prediction. I'll leave it up to you math experts to criticize the math in the book itself, but it was a really fun read even for someone like me, who can only do basic arithmetic. I guess that makes it sound like I'm easily impressed, but this is my way of saying it's also a good way to advocate the usefulness of math to people who don't see its real-world practical aspects.
http://www.amazon.co.uk/Calculated-Bets-Computers-Gambling-M...