https://www.paulickreport.com/horseplayers-category/figure-makers-companies-with-a-graphic-demonstration/
It\'s good except for the AP grass race thing. I didn\'t split them because of \"class\", I split them because the pace in the Secretariat was way faster than in the other two. When the pace is slow there isn\'t enough time to make up the lost time, especially on grass, which when dry is generally much faster than dirt, with the pace generally slower.
These are the Secretariat, Beverly D and Million, run consecutively. From Jakes comments about using the time as is, they would either have to have the Secretariat MUCH faster, or the other two MUCH slower. (Talking about maybe 5 points). See if Robert Bruce and the others make any sense if you do that.
I invite Jake to post their sheets for the same 3 races. It\'s a good way for both of us to show differences in methodology.
Thanks for posting these. I agree that the article was good, though there was something I also noticed about how you use class that just didn\'t ring true. I\'m glad you cleared that up.
In all methods studied, the projection method to determine the expected time of a race is key to the process of calculating a variant. The systems that interpret the horses\' times to project the expected times or figures (Beyers, TimeformUS, T-Graph) are taking into account which are the most representative of a horse\'s past races to get this and take into account the various factors involved in weighing the importance of these races. The totally computer generated systems (Brisnet, Equibase) are simply averaging the horses past races without regard to human interpretation of which races are most representative of a horses\' ability to give more weight to.
While the methods that are completely computer generated without human intervention are less prone to human error, IMO, the systems that take into account which past races are most representative and weigh them accordingly including, T-Graph, produce more accurate projections and ultimately, more accurate final figures.