Pace and Cap - Sartin Methodology & The Match Up - View Single Post

Jeebs · 09-13-2017, 09:49 AM

I've been forging ahead with my own homegrown spreadsheet (which I showed Ted, Barb and a few others at Saratoga), using Brisnet PP pace figures correlating to time charts specific to today's track and have added a few extra wrinkles to replicate the RDSS experience as much as possible. One thing I recently added was a Pace of Race screen, as the Brisnet PP Generator provides the E1 and E2 for the lead horse and LP pace for the overall race. However, I find that because the pace figs include a built-in variant that is not publicly available, I don't have the ability to "tone down" the variant (RDSS defaults to 50% of the DTV). This was apparent in a race I sampled on opening day with a line, as the PoR projected to a 1:07.2 off a 1:12 and change line for a horse. I knew when I saw a 176.8 Total Energy PoR that it was not "realistic", and likely the result of the variant. I affirmed this by looking at the adjusted times over at TimeformUS, and their final time pace was 1:08.8. Not quite as extreme, but still indictative.

A "correction" I am tinkering with is to "normalize" the pace figures by taking the difference between the PoR figure at each point, subtracting it from 80 and splitting the difference. Example, an E1 PoR of 100 would adjust down to 90 (100-80 = 20 x 50% = 10 points) and then applying that "variant" to the horse's line. Such a correction toned that 176.8 PoR I mentioned earlier down to a 173.2, which is still fast, but much more in line.

I am sure that there are better means of correction out there (this obviously punishes races that run "fast" on race pace figs, yet rewards "slow" times (race pace figs below 80, but pretty much moves things more to the center) that I am overlooking. One idea was to take the raw sectional/final times, apply gaps (like what is mentioned by Ted in an old thread and what Hat has done) so that the times are "comparable" to today's distance/surface, compare those to my par charts and extrapolate a variant that way.

Perhaps some can chime in here with some ideas. It would be appreciated!

09-13-2017, 09:49 AM	#1
Jeebs Grade 1 Join Date: Dec 2010 Posts: 909	Track Variants I've been forging ahead with my own homegrown spreadsheet (which I showed Ted, Barb and a few others at Saratoga), using Brisnet PP pace figures correlating to time charts specific to today's track and have added a few extra wrinkles to replicate the RDSS experience as much as possible. One thing I recently added was a Pace of Race screen, as the Brisnet PP Generator provides the E1 and E2 for the lead horse and LP pace for the overall race. However, I find that because the pace figs include a built-in variant that is not publicly available, I don't have the ability to "tone down" the variant (RDSS defaults to 50% of the DTV). This was apparent in a race I sampled on opening day with a line, as the PoR projected to a 1:07.2 off a 1:12 and change line for a horse. I knew when I saw a 176.8 Total Energy PoR that it was not "realistic", and likely the result of the variant. I affirmed this by looking at the adjusted times over at TimeformUS, and their final time pace was 1:08.8. Not quite as extreme, but still indictative. A "correction" I am tinkering with is to "normalize" the pace figures by taking the difference between the PoR figure at each point, subtracting it from 80 and splitting the difference. Example, an E1 PoR of 100 would adjust down to 90 (100-80 = 20 x 50% = 10 points) and then applying that "variant" to the horse's line. Such a correction toned that 176.8 PoR I mentioned earlier down to a 173.2, which is still fast, but much more in line. I am sure that there are better means of correction out there (this obviously punishes races that run "fast" on race pace figs, yet rewards "slow" times (race pace figs below 80, but pretty much moves things more to the center) that I am overlooking. One idea was to take the raw sectional/final times, apply gaps (like what is mentioned by Ted in an old thread and what Hat has done) so that the times are "comparable" to today's distance/surface, compare those to my par charts and extrapolate a variant that way. Perhaps some can chime in here with some ideas. It would be appreciated!