Back in the 1960's, I was captivated by "Percentage Baseball" by Earnshaw Cook. Now long out of print and a collector's item, this book was a forerunner of the "science" of SABRmetrics (after the Society for American Baseball Research) that refers to the scientific (statistical) evaluation of the game. A couple of years ago, Michael Lewis' Money Ball, which chronicled the construction and success of the Oakland team based on these principles, was a best-seller; there are now thousands of amateur and professional SABRmetricians. I was hoping for more of that kind of thing from from "Mathematician at the Ballpark", but the author's focus is more on betting on baseball than its strategy. Nevertheless, this book could provide an impetus to a student to learn about statistics and probability. It is pretty heavy on theory, but the fact that it is applied to a game makes it much more palatable that your typical textbook (Ross has also written some of those). He describes a betting strategy that he devised, that takes into account the number of fans likely to be willing to wager on each side of a game, and he has the courage to disclose in the book that the approach was a money-loser in 2003, and he has posted the results of continuation of the experiment in 2004, when it lost even more. Nothing like experimental tests of hypotheses, is there?