Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.



Pitfalls with simple probability
Posted:
Aug 26, 2012 12:30 AM


All I have historical data on horse trainers and their race results going back several years.I calculate the simple probability if Trainer A uses a gear change then in the last 3 years his success rate was Number winners/Total Runners to give me a strike rate or probability.
Questions:
1.How do I test for the optimal historical period to use for my probabilities.ie 3 years or 10 years of data .I could go back a decade for a larger N but then then things change over time Or is it best to just run my own trials
2. What are the pitfalls for using this simple approach to predict future outcomes



