Handicapping Horses

     I had written a Master's Thesis in the Business School at the University of Chicago on the use of a linear discriminant function in handicapping horses. I had become interested in multivariate analysis and linear discriminant functions in particular, wanted to test them with a lot of data that were cheap and also timely so that I would know how well the function discriminated in a short period of time. Lots of past performance data on horse races were available in the Daily Racing Form and they were cheap. (I seem to remember the Form cost something like 25 cents.) Other data, like numerical ratings of horses and jockeys, were to be found in a monthly Turf and Sport Digest. These data were used to derive a linear discriminant function to separate winners from losers. Although race results are known in a matter of minutes after a race has ended, it is more practical to gather them at the end of each day or the next morning,.

     I had given up handicapping horses until the electronic computer lured me back. Here was a fast way to calculate all the sums, sums of squares, and sums of cross products for several variables that were fundamental to deriving the function. I actually convinced Professor Ralph Cassady Jr., Director of the Bureau of Business and Economic Research, to fund a six month subscription to the Daily Racing Form as part of this "research." So, over the meeting at Hollywood Park in the summer of 1953, Karr and I took turns going to the track to bet our selections until my wife put a stop to it. She did not think it seemly that I should be encouraging students to gamble. Our joint handicapping effort ended, but not our continued study and use of SWAC.