Quote:
Originally Posted by cj
I remember seeing one from years ago but I don't remember the results. The toughest thing I think is defining the criteria. What do you use? Last race only, last race on surface, last race on surface at similar distance? Do you average figures?
The other problem is having access to the figures. Who is going to pay to buy PPs for all the data providers just to study the figures? You couldn't just buy a few weeks and have a real sample. I think you'd need at least a year, maybe more.
|
This was done in the mid 1990-s by Sports Stats ( Jim Bayle) and it was going to be repeated around 2000 but it wasn't as (I believe) Jim was too busy gambling on sports full time.
There are references to it here on PA from years ago. It used simple metrics of best last figure, best average last three figures, best average last five figures and perhaps a few others.
At the time the win percentage for Beyer, Bris and TrackMaster (now the Equibase figure) was about the same. Again, if I recall correctly it was in the 28% range.
The difference was the TrackMaster figure had a slightly higher R.O.I. on all three metrics. This was fairly easy to explain as it wasn't in as much use at the Beyer figure or Bris figure.