Machine learning techniques have shown their usefulness in accurately predicting greyhound races. Many of the studies within this domain focus on two things; win-only wagers and using a very particular combination of race history. Our study investigates altering these properties and studying the results. In particular we found a race history combination that optimizes our S&C Racing system’s predictions on seven different wager types. From this, S&C Racing posted an impressive 50.44% accuracy in selecting winning wagers with a payout of $609.34 and a betting return of $10.06 per dollar wagered.
Schumaker, Robert P.
"Machine Learning the Harness Track: A Temporal Investigation of Race History on Prediction,"
Journal of International Technology and Information Management: Vol. 27:
2, Article 1.
Available at: https://scholarworks.lib.csusb.edu/jitim/vol27/iss2/1