Test Method



At 11am each day the application is run, in turn it:

  • Acquires all the Horse Races for the day in the UK
  • Acquires supporting data needed by the algorithm
  • Generates a set of predictions
  • Recommends a set of bets.
  • Emails the recommendations to 3 independent witnesses

At the end of the day, the App:

  • Acquires all the days racing results and calculates the % gain
  • The results are then emailed to the 3 independent witnesses.

The results are shown on the linked pages below

  • Batch 1. 10/06/2013 to 28/06/2013. A bug was found and corrected at the end of this run.
  • Batch 2. From 29/06/2013 onwards. These will run with updated code from Batch 1.
  • Batch 3: This run includes nightly retraining of the Genome against fresh historical data, over a sliding window.
  • Batch 4: This run is using a revised fitness function for training the GA.
  • Batch 5: This run compares two GA training strategy results.
  • Batch 6: Runs two betting strategy regimes.
  • Batch 7: Runs a low e/w Betting Strategy option
  • Batch 8: Uses a Horse Score Filtering option in the Betting Strategy
  • Batch 9: Uses distance behind winner rather than position.
  • Batch 10: Uses an alternate betting strategy to reduce risk.