Dark Horse
01-05-2012, 03:10 AM
Would anybody be interested in a project to determine the base speed figure (Beyer) of individual horses?
In my model, basically a detailed form cycle approach, I recognize five forms that apply to every horse in a race: excellent, good, neutral, poor, and terrible. As part of a project to add a more accurate speed conversion key, based on these five forms, I want to determine the base speed of each horse. I believe this could be relatively simple, and once the base figure is established I have reason to believe that horses divert from it, up or down depending on their form of day, with remarkable similarity.
Simple example. I've penciled in a difference of 10 Beyers between a horse's base speed (on a neutral day) and a horse's top speed (on an excellent day). This is not set in stone, but works well enough as starting point. Once I have the horse's base figure and top speed figure, his full performance range should fall into place. Another general guideline is that horses run their top speed as three year olds or older, so the learning curve from 2 to 3 year olds has to be analyzed separately. For these horses I may use a temporary base speed of the L3 or L2 averaged. (if such a horse goes 83-94-104, with all clean numbers, I toss out the 80's number, because the horse has made the jump up). For three year olds the base speed becomes more or less finalized. There are always exceptions, but the idea here is to establish a general framework; and add the exceptions later.
For this purpose I only use clean Beyers. That is, only fast and firm track. Figures where the horse gave up early aren't included either. I would be open to switching to BRIS figures, if they should prove as accurate.
A picture is worth a thousand words. Eventually I want to develop graphs for each horse's speed history, with the form of day attached to each date.
If you want to participate, great. Feel free to add the Beyer history, with race and date, for horses you're interested in. Make sure that the Beyers are clean, and that the horse didn't weaken (ok for speed horses to fall back in the stretch, but none of the giving up early stuff). A good place to start may be the Derby Trail. This type of stuff tends to fall under individual research, but it may be interesting to do this out in the open. So just testing the waters, hoping that something might come out of it. Thanks for reading.
Will add examples from last year's Kentucky Derby and Preakness for Animal Kingdom and Shackleford.
In my model, basically a detailed form cycle approach, I recognize five forms that apply to every horse in a race: excellent, good, neutral, poor, and terrible. As part of a project to add a more accurate speed conversion key, based on these five forms, I want to determine the base speed of each horse. I believe this could be relatively simple, and once the base figure is established I have reason to believe that horses divert from it, up or down depending on their form of day, with remarkable similarity.
Simple example. I've penciled in a difference of 10 Beyers between a horse's base speed (on a neutral day) and a horse's top speed (on an excellent day). This is not set in stone, but works well enough as starting point. Once I have the horse's base figure and top speed figure, his full performance range should fall into place. Another general guideline is that horses run their top speed as three year olds or older, so the learning curve from 2 to 3 year olds has to be analyzed separately. For these horses I may use a temporary base speed of the L3 or L2 averaged. (if such a horse goes 83-94-104, with all clean numbers, I toss out the 80's number, because the horse has made the jump up). For three year olds the base speed becomes more or less finalized. There are always exceptions, but the idea here is to establish a general framework; and add the exceptions later.
For this purpose I only use clean Beyers. That is, only fast and firm track. Figures where the horse gave up early aren't included either. I would be open to switching to BRIS figures, if they should prove as accurate.
A picture is worth a thousand words. Eventually I want to develop graphs for each horse's speed history, with the form of day attached to each date.
If you want to participate, great. Feel free to add the Beyer history, with race and date, for horses you're interested in. Make sure that the Beyers are clean, and that the horse didn't weaken (ok for speed horses to fall back in the stretch, but none of the giving up early stuff). A good place to start may be the Derby Trail. This type of stuff tends to fall under individual research, but it may be interesting to do this out in the open. So just testing the waters, hoping that something might come out of it. Thanks for reading.
Will add examples from last year's Kentucky Derby and Preakness for Animal Kingdom and Shackleford.