Predicting Horse Race Winners

 
Many horse racing systems are built around the fact that over 60% of winners come from the first two horses in the betting.  First2 is an attempt to improve the punters chances of winning, by selecting those races that are likely to have a winner coming from the first two in the betting, using a trained Neural Network.
 
Inputs:
 
Runners (number of runners)
 
Distance (in furlongs)
 
Handicap (race type)
 
Class (1 for A or listed, 2 for B, 3 for C..)
 
Stake>5k (Prize money for winner is greater than £5000)
 
Odds>2 (Odds of first horse in the betting is greater than 2/1)
 
Output:
 
Win (Is one of the first two in the betting likely to win?)
 
The data for this network are for flat races run in 1998. 
 
The network is produced with Grow hidden layer 1 checked. 
 
The controls have Optimize and Decay set for both Learning Rate and Momentum.  The Target error is set to 0.  The Correct after rounding validating stop is set and 100 examples have been selected at random for validating. 
 
Auto Save as been set the save every 100 cycles if the number of correct validating examples have increased.
 
The First.vi.tvq will be produced by AutoSave.  This file can be used to predict the winners.  In the 100 validating examples it correctly predicts if one of the first two horses in the betting will win in 69% of the races.  This is a 9% advantage over the known results.  Is that a big enough advantage to make a profit ?
 
See the First2.tvq sample. 
 
by Tom Huxley

Created with help of DrExplain