Hanicapping Horses is NOT About Picking Winners

After finishing our inaugural season as an NFL handicapper at 62% against the spread, we are feeling fairly confident that we can be successful as oddsmakers. After reading "Picking Winners" (Beyer) and another winning horse racing book by Rove, our confidence level is further increased. We believe that the key to our success will be our probabilistic approach to handicapping. In each of our NFL predictions we determined our line, compared it to the Vegas line, and computed an Adjusted Cumulative Distribution Function that represented our chance of winning a bet against the Vegas Line. In a similar manner, successfully handicapping horse races will not depend on predicting the winner but rather it will depend on understanding each horse's probability of winning and identifying betting opportunities with positive expectation value.

The majority of the horse handicapping literature (if not all of it) focuses on identifying which horses are likely to win. This is intuitive. One naturally assumes that the ability to determine which horse is the likely winner would be the key to becoming a winning horseplayer. But this is an oversimplification. No race has a certain outcome, otherwise they would not bother running the races. Since each horse has a probability, no matter how large or small, of winning the race, the handicapper's job is to determine what the probability is for each horse and determine whether the probability multiplied by the payout is greater than the layout. If so, the bet is a good one.

Working against every horseplayer is the house take of 17% - 20%. For every $100 wagered, the house gets $17 - $20. Go with $100. Bet $20. Win. Bet $40. Lose. Bet $20. Lose. Bet $20. Win. Bet $50. Lose. Bet $50. Lose. Bet $20. Win. Bet $40. Lose. Bet $20. Win. With average play this player's expectation value is a loss of $56. Some days they will win and some days they will lose, but more often they will lose. If the same player, however, makes the same wagers but only wagers on the horse with the best expectation value they will win. They will win more by sitting out races without a positive expectation horse and betting more on races with a strong expectation horse. If, for example, they identify a horse with a 20% shot at winning that is paying 6 - 1, they gain $40 for every $100 that they wager.

Working against everyone that uses a handicapping service is the fact that many others are most likely using the same or similar service. By understanding the likely winners they lower the payouts for the likely winners, decreasing the likliehood of overcoming the house's 17% - 20% take. Using a data service or collecting your own data, on the other hand, and applying rudimentary handicapping skills to translate the data into probabilities, may identify positive expectation betting opportunities.