As an on/off “student” of the game for many years, I’ve had the opportunity to learn from Dave, and to explore concepts and theory that was foreign to me prior to investing the time into his seminars. One of those concepts that have stuck with me - and still challenge me to this very day - is the concept of the 90% regression speed sort.
NewPace was my first introduction to the RegSpeed Sort, and it was quite an eye-opener for me. Being able to estimate the horse’s low and high end figure with a high level of confidence was in my opinion, one of Dave’s finest discoveries. Even when not using pencil/paper NewPace, RegSpeed Sort is such a simple concept, that I can have my pre-handicapping ritual for each race finished in mere minutes. The information provided makes a race come to life. It gives me an idea of a favorite’s theoretical strength, as well as guiding me in the direction of horses with longshot potential. Using the pacesetters to establish the pace of race further enhances this, opening the door for prices to pop.
What is it that challenges me? I wrestle with the Sartinian concept of using only comparable races across surface/distance structures, while ignoring wet dirt or soft/yielding turf, and whether I should ignore a line if my eyes determine that the horse was practically eased in the late stages, creating a speed rating/figure that’s really low because of a lack of effort. I however, question myself as to whether or not this is optimal, and if I could get a better read by say lumping all dirt races not on wet tracks together, and treating synthetic as an interchangeable entity between dirt and turf. I’m also not sure if I should ignore “wet fast” designations or if I should treat them like I would a dry fast track.
What are others experiences with RegSpeed Sort, whether pen/paper or HSH?
Hello Tim. I'm using DET, but was not an HSH user. My understanding is that DET (the AI) creates a "synthetic" paceline using several fractions and calls. The AI then provides the final output which is expressed as a Hit Rate. It's a very complex process. I no longer use a paceline selection, rather use the DET AI Best Horse and other reports and objects to make selections.
Back to your original question about wet dirt, soft/yielding turf, eased horses and synthetic (Tapeta) tracks. I think you would need to back-test several races in each category and check the results. There's a lot of opinions out there, but I'm not sure if any have been validated. Like you, I struggled with this-- checking the horse's previous starts, trainer stats, pedigree, etc. But, I don't think there's a true answer without digging deeper and documenting findings in each of the categories.
Here’s an excerpt from an email I sent Dave last night. I’ll leave it here for your consumption and feedback:
I’ve attached a race that I worked (yesterday). I didn’t bet. I’m trying to get into the habit of working/pre-handicapping races in this fashion so that I can readily identify future patterns and race scenarios. I’m not so much concerned about result as I am the process at this point in time. It’s likely going to take hundreds… make that thousands… of races of repetition to really get good at it.
On my worksheet, which is in Program# order, I have each runner’s LO and HI RegSpeed, plus the Best of the Last 3 (comparable) races. By comparable in this example, I used races on today’s surface, regardless of distance, excluding wet tracks and turf.
I use the TimeformUS Past Performances, so I use their main Speed Figure (which is adjusted for pace). They also present a Composite Early and Late rating with their interpretation of visual running style. I take the difference in the Early/Late ratings to create a composite “Energy” number, reminiscent of the later Sartin programs and RDSS.
The result? Probably a roundabout variation on the various iterations of NewPace.
In hindsight, a Win bet dutch of 7 and 8 looks rather desirable given the construction of and facts that we know about this particular race. To be honest, I didn’t have exactas or other exotics in mind, so I don’t have any insights to pass along.
The winner here was definitely within reach with the right tools, strategy, and mindset. At 17-1, it definitely didn’t “jump off the page” but by leveraging its marginal contender status with its price, it would have been a low risk reward.
I would not have had the #8 using my handicapping. The horse had a low projected speed rating, low early speed and low class. Your worksheet says-- "...... 1 point lower ...... qualifying as a contender due to probable margin of figure error." Not sure what this means. I'm not sure you can handicap and consistently pick winners assuming/guessing there are figure errors. To me this wouldn't make sense and I'm not sure how many winners you would find from "marginal contenders". Also, seems like quite a bit of work.
By low PSR, ES and Class, are you referring to your HDW data?
RE: the 8 being a single point below the projected POR (high RegSpeed LO of the frontrunner), I don’t find a single point to be a strong enough basis for “tossing” a runner.
The notes I was writing were for illustrative purposes. And yes, pencil/paper can be for lack of better words, a mundane/monotonous task. Unfortunately, DRF doesn’t offer data files for TimeformUS where figures and other data can be extrapolated into a homegrown program/database.
Yes, HDW, but I'm using several factors, not just early early speed. A horse may be close in early speed, but if it ranks low in the other factors, then it's most likely a toss. I don't look at each individual factor and look to qualify a horse as a contender because it just missed in one of them. Otherwise I may find a reason to qualify each horse in the race and for me that would not be a consistent handicapping process.
One thing to remember is that New Pace was developed with the non-HSH user in mind. Consequently, Dave had to develop a contender selection process that could be used with DRF or BRIS past performances, for example. I personally got better results using the PickMaster for contender selection. To select contenders, try experimenting with the NP contender selection rules but using the PM to select the factors, as opposed to using only the RegSpeed factors.