GP 3/20/14 9th Race

[Ted Craven]

Originally Posted by Segwin

Continuing on:

Does a 170 CSR in a 5,000 claimer = a 170 CSR in a 10,000 claimer?

Does a 174 PoH rating in a 5,000 claimer = a 174 PoH rating in a 10,000 claimer?

Terry,

To start, here is a refresher as to how the Composite Speed Rating - CSR is calculated (from the RDSS documentation): http://paceandcap.com/forums/showpos...77&postcount=3.

Please note the Caveats expressed there as to why CSR may under-represent a horse today, often due to a poor - though possibly excusable - last line, or conversely, may over-represent a horse today by using Speed Ratings on surfaces or at distances to estimate performance a horse has never undertaken before (first time Turf, or sprint, or route ... etc):

The CSR was formerly only displayed on the NewPace screen, though technically not a part of NewPace. It is a weighting of the Adjusted Speed Ratings from the last 4 usable lines, regardless of distance, surface, race conditions or finish. It is not dependent on a single line, but rather a weighting of the last 4 usable lines.

The weighting is as follows: 100% of the last line SR + 62% of the average of the last 2 lines SR + 38% of the average of the last 3 lines SR + 24% of the average of the last 4 lines SR. Thus the last line SR has a major weighting in the CSR, with the weighting declining for older lines. Thus, if the horse ran well in the last line and faded in the stretch for whatever reason (wrong surface for the horse, wrong distance, inappropriate pace, trouble), since CSR uses ONLY final time Adjusted SR, that last line poor finish will hurt the horse's CSR. For any number of reasons, such a last line may be forgiven and the horse may indeed be a contender today, though the CSR may rank poorly. Then again, if the horse has run fairly in its recent history, the CSR will rank it fairly among others, for better or for worse.

The theory behind the rating is that the adjustments to final time Speed Ratings do all the heavy lifting of equalizing different distances and surfaces (and Daily- and Inter-Track variances). But you still can't accurately rate a horse on something it hasn't yet done by its past dissimilar performance (IMO). I don't the think the CSR rating helps much in these circumstances.

To your specific question: does the actual numerical CSR value (e.g. 170) represent an equivalence performance across 'Class levels' ?

I would answer that the CSR number itself - e.g. 170.0 - (distinct from its rank or differential from best in each race) is fairly meaningless, statistically because so much 'noise' can get incorporated into it, especially poor finishes last race which can be reasonably forgiven. That poor SR will play heavily in constructing the CSR number.

But if we consider our principles of paceline selection, and the concept of 'good races', 'good within bad' and 'bad' races, then there are any number of scenarios where we should simply toss that last 'other than good' race and rate the horse from a different, recent line (e.g. non-comparable surface or distance to today where the horse has either not demonstrated ability or worse has demonstrated dislike; pace was too fast for its early ability - today is different; first after a Layoff - horse is working into shape, etc; trouble; etc, etc).

So a horse's CSR number gets depressed by a bad last line, or 2nd line, etc - or one of the line components possibly over-inflates the total by coming from a surface other than today; or a horse raced in mediocre fashion against a fast pace of race; or whatever else. If the total CSR number = 170, that doesn't necessarily mean it is comparable to a 170 earned by a different horse (in the same or in a different race) where all the rated races and performances were much more fair, or comparable to today.

I would not recommend using the actual CSR number itself for anything, other than double-checking how close it is to adjacently ranked horses (e.g. within .50, you may choose to consider 2 horses tied in CSR rank). Instead, use the CSR rank, and the gap (% differential from best, shown on the Entries screen) to get a head-start on the top Contenders. (And that's also why you don't need the NewPace module to make use of the CSR ranks or gaps).

However, if you use CSR ranks, you MUST inspect each horse to see if the CSR is being either depressed or inflated as discussed above. (It would be nice to have a facility to permit the user to omit selected lines from a CSR re-calculation - and we might have that, soon).

The attraction of CSR is that it is an automated calculation not relying on the user to select a line - unlike most of the other ratings which are based on a selected line. At the same time - this is also its drawback. So far (IMO) our brains can see patterns which the CSR formulas do not.

Total Energy (TE) is a rating dependent on line selection and I would say that its numerical values are much more suited to equating horses who perform in different class levels (i.e. over a large group of races, the average TE of winners in CL 5 races will be lower than those in CL 20+ races, or Allowance, or OC or Stakes, etc).

As to this race in question, FWIW, my own guideline re first time Turf is to bet against them (unless they are held as a serious favourite, then I would tend to pass). I would be surprised at research which indicated that the payoffs of winners first time on the turf approached break-even. Dismissing that much money bet, puts you ahead of the game on all the rest of the horses entered.

The #7 winner was an OTE among cant-help-it Earlies/Fighters - AND owned BOTH the early Segments on the Segments screen (once you eliminated the non-turf horses). AND it was 3/2!

IMO, the trick then was finding the place horse at a sufficient price to warrant a single (or double?) on top of an exacta including the rest of the contenders. In retrospect, hard to know if the Place horse #4 (only horse with odds worth hoping for) could be counted on not to run (and die) with the Earlies, instead of stay out of trouble and close nicely (which it did).

Hope you find these opinions re CSR useful.

cheers,

Ted

[Segwin]

Excellent reply Ted! Thanks much for taking the time to detail it out.

I did play the 7/4/2 for the Exacta but alas it was a little too little.