I think he addresses that issue specifically in his book. If you
go to his website he has a sample data set to download which has
several years worth of Milford data available so I'd imagine that
his data set was sufficiently large to get around that problem.
After I first bought his book I was disappointed at the lack of actual source code available so I e-mailed Prof. Skiena and he very nicely sent me a copy of his simulation code as well as made it available on his web site. The code was the source for the actual game simulations but did not include code for downloading and parsing results/entries and creating player % probabilities. I had trouble at first getting his simulation code to compile and run so I printed it all out and translated it from C to Java (language I know best) and got that to work. I then created from scratch my own program to do the entry/result parsing and player % probs. The results have been relatively mixed so far. I have been slowly but steadily building up my archive of entries/results so hopefully at some point I can make a stronger determination of what changes I think I should make to the system.
For now, I have taken some of Skiena's advice and tailored it towards methods of predicting greyhound races which, in the short term, so far seems to be doing pretty well. What is not doing well, it would seem, is the plainfield trifecta pools !! This past saturday I was at Mohegan race book and hit a tri-box at plainfield. Whoo-hoo!! Then the amounts posted and I won fricking 35 bucks.
Not to extend this post to far astray from the topic at hand but, in general, the pools need some rejuvenation at many venues. The exacta and tri pools at Miami always seem decent and the payouts usually make for a happy trip to the ticket window. Looking at the pools vs payouts at many of the tracks though and it seems obvious to me that there are a lot of people getting in the money. That is, a lot of people who know how to handicap a race vs the people just out to place a few wagers and have a couple of drinks. Damnit!! We need more suckers at the frontons and tracks!! Although at Mohegan race book this past saturday I recall two amusing situations. The first was a guy who appeared to be drunk as a skunk staggering towards the cashier and asking her, "What do I do ? Just pick numbers? Can you explain this to me?" Taken aback the cashier explained that he could bet on several different tracks and at those tracks there were several races available and he could make his choice based on the big screens all over the room, yada , yada, yada". With a few tracks closing in on post time I think about a dozen people wated to fricking kill him for tying up the line. The second thing was this guy who urgently ran up to the cashier and said,"Gimme the very next dog track with a trifecta and gimme the 4-5-6 tri-box." wtf?
Made me think, what we really need is a pari-mutuel "quick pick". That should fatten the pools and keep certain people happily ignorant of whatthey are actually betting on.