No matter how good your model, it’s still You Play, You Pay. And Rowland’s disclaimer echoed a comment an influential racing veterinarian made to me: "Never invest in something that eats while you sleep."
Everyone should be allowed at least one vice, and mine is horses. I love them, spend as much time around them as feasible, and find that after years of this I’m still learning. Recently I’ve met a couple of people know a whole lot about horse racing. They don’t know a thing about the horse itself, but they have a very sophisticated understanding of the mathematics of predicting performance.
So that I could keep up my end of our conversations, I looked further into handicapping and discovered that horse races themselves are only a kind of graphical display to show the results of some massive efforts at statistical analysis, including some of the quantitative forecasting techniques used by financial analysts and whole lot of custom Excel programming. This should surprise no one–after all, what is betting on a horse if not decision making under uncertainty?–but what did surprise me is level of technical discussion about the math and how to work it through in Microsoft Excel statistics.
Take a look, for instance at a recent blog on "taking the price" from the U.K.’s Simon "The God of Odds" Rowland. Taking the price is locking in the odds when you bet. He discusses how to correlate a horse’s rating–the amount of weight the horse has been assigned to carry–with the actual odds on this competitor. He then gives the mathematical recipe for his custom Excel spreadsheet, which combines Monte Carlo simulation and the related Markov Chains technique. He wraps up his demonstration with a standard disclaimer: "It must be immediately apparent that this process is very susceptible to the GIGO (garbage in, garbage out) principle. No manner of mathematical manipulation will make up for essential shortcomings in the ratings and in the confidence attributed to those ratings."