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

Dorianmode · 08-14-2017, 10:09 PM

Hi,
As Ted knows, I've developed a set of Excel programs which started out a few years ago as 'profiles', to which I added my own version of something like (but different ... and which I worked out myself) NewPace. There were a lot of things wrong with the early versions, and I recently discovered a MUCH easier way to do the same thing. I now use them together. Without RDSS, (on which they are totally based), they would not work. I download the runners to the profiles, for comparisons. I have been very pleased with the results. Ted and I have talked about all this, and will again soon, I understand, about offering my version of NewPace, (which I need to rename, as a statistical calculation, as it's based on the TrackMaster speed figures, which we already have in RDSS), and my computations are different. In the last couple years I added some regression analysis, (and automated it, as there is no way to mess with that while preparing a race), to tell me what was the best way to weight the readouts which come from comparing today's runners to my profiles. Then also I added a multiple regression analysis. The two regression numbers are now my most powerful predictors of the winner. Of course one must still do basic handicapping, and make basic judgments about contenders, class droppers, and which speed figures to omit or average, etc, etc. It's not a black box. I don't have enough organized data to present it to these fora yet, but will when I have enough. Most recently I worked out also a calculation (also automated in Excel) based on a few things, which rate the advantage of my top two contenders, over the next three. That is showing some great promise, but also needs more work, and surprisingly it's pointed me to look a three-way dutch, if the numbers for two are not optimum, which I had not thought about, until it happened. If all this sounds impossibly complicated, ... it isn't. I can prepare a race in about 5-7 minutes, as all the steps in Excel are macro driven (automated). I just push he command buttons, (which I programmed), and boom, boom, boom, the numbers pop out.

I like Mike's sheet ... mine are a lot more "busy" and have a lot of calculations.

One of the most fun successes with this all was on Saturday ... I'm still chuckling about it. DelMar race 6, 8/12/17. It had the winner, the place horse, the show horse and the 4th place runner... all in precise order. LOL. It's good, but it's not THAT good.

I do have separate profiles for the place, show and perfecta horse, that come out of the numbers at the same time the win numbers are generated. It's probably mathematical heresy, but I recently cleaned out my profiles. I was thinking that many races and many numbers just "blur" the picture, and hides the significant among the noise. I now use only seven races (which are frequently updated) in the profiles. If the race turns out as well as the one above, I replace one in the profile with that one.

08-14-2017, 10:09 PM	#8
Dorianmode always learning Join Date: Jul 2011 Location: Minneapolis / Rancho Santa Fe Posts: 277	Hi, As Ted knows, I've developed a set of Excel programs which started out a few years ago as 'profiles', to which I added my own version of something like (but different ... and which I worked out myself) NewPace. There were a lot of things wrong with the early versions, and I recently discovered a MUCH easier way to do the same thing. I now use them together. Without RDSS, (on which they are totally based), they would not work. I download the runners to the profiles, for comparisons. I have been very pleased with the results. Ted and I have talked about all this, and will again soon, I understand, about offering my version of NewPace, (which I need to rename, as a statistical calculation, as it's based on the TrackMaster speed figures, which we already have in RDSS), and my computations are different. In the last couple years I added some regression analysis, (and automated it, as there is no way to mess with that while preparing a race), to tell me what was the best way to weight the readouts which come from comparing today's runners to my profiles. Then also I added a multiple regression analysis. The two regression numbers are now my most powerful predictors of the winner. Of course one must still do basic handicapping, and make basic judgments about contenders, class droppers, and which speed figures to omit or average, etc, etc. It's not a black box. I don't have enough organized data to present it to these fora yet, but will when I have enough. Most recently I worked out also a calculation (also automated in Excel) based on a few things, which rate the advantage of my top two contenders, over the next three. That is showing some great promise, but also needs more work, and surprisingly it's pointed me to look a three-way dutch, if the numbers for two are not optimum, which I had not thought about, until it happened. If all this sounds impossibly complicated, ... it isn't. I can prepare a race in about 5-7 minutes, as all the steps in Excel are macro driven (automated). I just push he command buttons, (which I programmed), and boom, boom, boom, the numbers pop out. I like Mike's sheet ... mine are a lot more "busy" and have a lot of calculations. One of the most fun successes with this all was on Saturday ... I'm still chuckling about it. DelMar race 6, 8/12/17. It had the winner, the place horse, the show horse and the 4th place runner... all in precise order. LOL. It's good, but it's not THAT good. I do have separate profiles for the place, show and perfecta horse, that come out of the numbers at the same time the win numbers are generated. It's probably mathematical heresy, but I recently cleaned out my profiles. I was thinking that many races and many numbers just "blur" the picture, and hides the significant among the noise. I now use only seven races (which are frequently updated) in the profiles. If the race turns out as well as the one above, I replace one in the profile with that one. Attached Images Last edited by Dorianmode; 08-14-2017 at 10:20 PM.