DanG
10th April 2007, 02:03.17 PM
I was killing time on a rainy Tuesday talking about the “K” ratings consistency concerning contenders.
Then Mike [MVM] drops this post into the mix and in my opinion it deserves its own thread. It is so well written and talks to the issues concerning artificial surface and the transition to / and from.
So…Mike, I hope you don’t mind if I take the liberty of hijacking your post, but it’s the “nuts” IMHO as they say in poker world.
---------------------------------------------------------------------------------------
MVM ~ I'm still trying to iron out some of the finer details in the way I deal with PolyTrack(s).
One of the things I was really wrestling with was the field finish (FinBL) compaction. At Hollywood and Turfway, the average First2Last finish beaten lengths margin was basically cut in half after their version(s) of Polytrack were installed. I'm not certain whether or not this is due primarily to attributes of the surface, or a difference in the way races are ridden.
Assuming that the difference in the way races are ridden primarily results in a slower pace, I need to look at the n% of the fastest paced races vs. n% of the slowest paced races and see if the resulting FinBL compression is as dramatic as it is on conventional dirt surfaces.
I've already (substantially) adjusted the value of a BL at the 2 tracks mentioned above to be much closer to the BL values I use in turf races than dirt races. It worked very well post-Hollypark and in general the figures, esp. inter-track make much more sense.
If anyone has noticed the divergence of PolyTrack figures in shippers vs. their dirt figures and the dirt figures of other non-poly entrants, take a second look. The figures for the Poly winners are much closer to their dirt figures than the (by way of comparison) polytrack figures of the also-rans, which are generally highly inflated. It would seem counterintuitive to believe that the horses who move up on PolyTrack (which some definitely do) are the losers, while the winners simply approximate their dirt form.
One additional "experiment" that I ran during the 1st 45 days of the current SA meet was to identify every pair of 4&Up horses exiting a tandem race at Hollypark who replicated their finish position relative to each other in a subsequent race at SA (A beat B in both events). The pct. of tandem horses replicating their relative finish positions was essentially identical to all intra-meet statistics I have compiled, however the lengths separating the 2 entrants (on average) nearly doubled. For me, this validated the BL adjustment I made, as the Poly2Dirt result was almost exactly the inverse of the Dirt2Poly adjustment.
Hollywood is part of my primary circuit, so in spite of the short meet/small sample size I did that study first. When I followed up with TP, Kee and WO, it became readily apparent that each "flavor" of PolyStyroRubberWaxFishOilMichelinTrack is quite different.
TP appears quite similar to Hol as it relates to FinBL compression, in that the post-poly difference is substantial. But there are so many shippers, many of whom come from circuits far less competitive than SoCal. These shipping tracks have comparatively huge FinBL margins and resulting StDevs that make a very large sample size a necessity. Fortunately, TP has a plethora of racing dates and provided this.
Kee, just as in seasons prior to Poly is it's own animal, and the only one of the 4 artificial surfaces that seemed to have a pronounced "some handle it, some don't" aspect to it. The FinBL reflected this, and there is some indication that there was an optimum pace at Kee (the time-honored 12, 12, 12, 12) and that substantial deviation from this cadence was heavily penalized, both for Closers and Frontrunners.
I just completed a cursory study on WO (last on my list because I don't play it). There is a marginal degree of FinBL compaction, but it is accompanied by a fairly pronounced decline in Field Movement (Field Movement is a rating I use to determine how much position change there was within a race from 1st call to Finish. Sum total of the difference (absolute value) between Call1 position and Finish position of all entrants/field size). In other words, far less passing (in all positions) occurs at Woodbine than it did prior to the installation of PolyBackBaconTrack. I don't watch the races at Woodbine too often, so I don't have any subjective opinions to include, but it would appear (at first glance) that perhaps the jockey colony in the Toronto area has adjusted their riding styles, and the races are run a bit more conservatively early.
Editorial: I have seen protracted closer/frontrunner biases often attributed to surface characteristics when (upon close watching of the actual races) they are often times self- fulfilling prophecies which are attributable more to the general makeup and current mindset of the local rider colony. This phenomenon happened in the extreme at EmD this past summer, especially at the 6.5 furlong distance, where the races where generally ridden as if they were carded from Seattle to Olympia rather then backstretch chute to finish.
Then Mike [MVM] drops this post into the mix and in my opinion it deserves its own thread. It is so well written and talks to the issues concerning artificial surface and the transition to / and from.
So…Mike, I hope you don’t mind if I take the liberty of hijacking your post, but it’s the “nuts” IMHO as they say in poker world.
---------------------------------------------------------------------------------------
MVM ~ I'm still trying to iron out some of the finer details in the way I deal with PolyTrack(s).
One of the things I was really wrestling with was the field finish (FinBL) compaction. At Hollywood and Turfway, the average First2Last finish beaten lengths margin was basically cut in half after their version(s) of Polytrack were installed. I'm not certain whether or not this is due primarily to attributes of the surface, or a difference in the way races are ridden.
Assuming that the difference in the way races are ridden primarily results in a slower pace, I need to look at the n% of the fastest paced races vs. n% of the slowest paced races and see if the resulting FinBL compression is as dramatic as it is on conventional dirt surfaces.
I've already (substantially) adjusted the value of a BL at the 2 tracks mentioned above to be much closer to the BL values I use in turf races than dirt races. It worked very well post-Hollypark and in general the figures, esp. inter-track make much more sense.
If anyone has noticed the divergence of PolyTrack figures in shippers vs. their dirt figures and the dirt figures of other non-poly entrants, take a second look. The figures for the Poly winners are much closer to their dirt figures than the (by way of comparison) polytrack figures of the also-rans, which are generally highly inflated. It would seem counterintuitive to believe that the horses who move up on PolyTrack (which some definitely do) are the losers, while the winners simply approximate their dirt form.
One additional "experiment" that I ran during the 1st 45 days of the current SA meet was to identify every pair of 4&Up horses exiting a tandem race at Hollypark who replicated their finish position relative to each other in a subsequent race at SA (A beat B in both events). The pct. of tandem horses replicating their relative finish positions was essentially identical to all intra-meet statistics I have compiled, however the lengths separating the 2 entrants (on average) nearly doubled. For me, this validated the BL adjustment I made, as the Poly2Dirt result was almost exactly the inverse of the Dirt2Poly adjustment.
Hollywood is part of my primary circuit, so in spite of the short meet/small sample size I did that study first. When I followed up with TP, Kee and WO, it became readily apparent that each "flavor" of PolyStyroRubberWaxFishOilMichelinTrack is quite different.
TP appears quite similar to Hol as it relates to FinBL compression, in that the post-poly difference is substantial. But there are so many shippers, many of whom come from circuits far less competitive than SoCal. These shipping tracks have comparatively huge FinBL margins and resulting StDevs that make a very large sample size a necessity. Fortunately, TP has a plethora of racing dates and provided this.
Kee, just as in seasons prior to Poly is it's own animal, and the only one of the 4 artificial surfaces that seemed to have a pronounced "some handle it, some don't" aspect to it. The FinBL reflected this, and there is some indication that there was an optimum pace at Kee (the time-honored 12, 12, 12, 12) and that substantial deviation from this cadence was heavily penalized, both for Closers and Frontrunners.
I just completed a cursory study on WO (last on my list because I don't play it). There is a marginal degree of FinBL compaction, but it is accompanied by a fairly pronounced decline in Field Movement (Field Movement is a rating I use to determine how much position change there was within a race from 1st call to Finish. Sum total of the difference (absolute value) between Call1 position and Finish position of all entrants/field size). In other words, far less passing (in all positions) occurs at Woodbine than it did prior to the installation of PolyBackBaconTrack. I don't watch the races at Woodbine too often, so I don't have any subjective opinions to include, but it would appear (at first glance) that perhaps the jockey colony in the Toronto area has adjusted their riding styles, and the races are run a bit more conservatively early.
Editorial: I have seen protracted closer/frontrunner biases often attributed to surface characteristics when (upon close watching of the actual races) they are often times self- fulfilling prophecies which are attributable more to the general makeup and current mindset of the local rider colony. This phenomenon happened in the extreme at EmD this past summer, especially at the 6.5 furlong distance, where the races where generally ridden as if they were carded from Seattle to Olympia rather then backstretch chute to finish.