Quote:
Originally Posted by ubercapper
Did he really go from 27-30% to average (15-20%) or was the 27-30% win rate explainable by chance given the small numbers?
I understand statistics but don't come close to understanding when a sample size has statistical significance and leave it to experts to determine when that level is reached. From my reading of the work of Dr. Parkin and the Equine Injury Database, it takes some pretty big numbers to have confidence and when it comes to trainers we rarely get those sample sizes.
If that's the case, much of these discussions are speculation, or worse, spurious correlations, for example if trainer A is winning at 40% he must be cheating.
|
Wasn't looking at it from a purely statistical viewpoint, rarely are there enough samples for one trainer to make it statistically significant. Was looking at how his horses were performing (speed and pace at class levels) compared to before.