TimeformUS Pace Projector refined to include impact of blinkers, experience

It’s a snowy Tuesday here at TFUS HQ, the perfect sort of day to announce some refinements of our product—based on a considerable amount of research, of course. All of the work discussed here has now been implemented in TimeformUS PPs, and all of it was derived from the broad goal of making Pace Projector more accurate.

Additionally, as you may know, Pace Projector is partially informed by the Early Pace Ratings that you can see above all of our PPs. So the refinements triggered by the studies also resulted in sharper Early Ratings than before—and also sharper Late Ratings than before.

Screen Shot 2014-02-18 at 1.51.29 PM

All of the following may seem obvious and intuitive to you, and if so, pat yourself on the back. You’re the best. OK, on to the good stuff.

The headlines, based on a database study of over 250,000 starts, are as follows:

  • Second-time starters improve their early speed in blatantly quantifiable ways. Accordingly, we made adjustments to second-time starters on Pace Projector (and, of course, to their early ratings) to reflect these results, and Pace Projector and our Early Pace Ratings became more accurate because of our adjustments. Please note: The adjustments, in this case, impact ONLY our projections for a horse’s second start. In subsequent races, we draw exclusively on the past performances rather than any additional “Early Pace Maturity” projections. Indeed, our studies show that after the horses’ second-start, their improvement in early pace numbers levels off considerably.
  • Second-time starters and third-time starters improve their LATE speed in more subtle but clearly identifiable ways. And so we also made subtle adjustments to the late ratings displayed for second-time starters and third-time starters to better reflect what actually happens. We remove these “Late Pace Maturity” projections after the third start, and then rely strictly on each horse’s on-track past performances to inform his late ratings.
  • In what will come as a great relief to so many of our customers, we have safely concluded that horses adding blinkers improve their speed in blatantly quantifiable ways, and we could improve Pace Projector by accounting for these improvements. No, not all horses who add blinkers show better early speed. But enough do that it sharply affects their average early running positions, enough so that Pace Projector (and the attendant early ratings) becomes more accurate if we make an adjustment each time a horse puts on blinkers after not wearing them in his previous start. Once again, subsequent starts AFTER the first-time-blinkers race rely on the individual horse’s on-track results, rather than any continuing “bonus points” for continuing to wear blinkers. Our research indicated as much—the addition of blinkers was one thing, but keeping them on didn’t lead to further changes in running style, on average.
  • To be clear, we’re making adjustments only is when there is a CHANGE in equipment, because our research indicated this would improve the quality of our Early and Late Ratings and our Pace Projector.
  • And yes, you guessed it, there is a smaller, but measurable impact when horses take blinkers off. We were once again able to improve the accuracy to Pace Projector (and early ratings) by making an adjustment strictly for the projections for the race in which a horse is going to remove blinkers. And, as in the cases above, subsequent starts AFTER the blinkers-off race rely on the individual on-track results, rather than any continuing “Early Pace Chill Out” for the continuance of blinkers off.
  • Please Note: Because we’ve made adjustments to Pace Projector and Early/Late ratings to reflect these studies, YOU don’t need to bump a horse up or down in your own projections, unless you have every interest in “double-counting.”

Some notes on the data behind the study and subsequent conclusions/adjustments, for those of you who can’t live without it:

  • Please note our early and late pace ratings are all harmonized to the Timeform Global scale.
  • As mentioned, we conducted a database study of over 250,000 starts to draw the conclusions we drew.
  • We then ran the study against the same 10,000 race subsets (of 121,793 races total) we used in our August study of Pace Projector’s efficacy.
  • As you can see in the chart below, in a game of milliseconds, efficacy because of these new studies was improved slightly, and across the board. A random sample of 10,000 dirt races showed the least improvement of all, but when we drilled down to dirt routes and wet dirt tracks, we were able to pick up a couple % points of efficacy.
  • Our biggest win came in turf races, where we picked up more than 3 percentage points in both turf sprints and routes.
  • Please note we are judging efficacy here via Pace Projector’s strike rate for properly positioning a horse at the front, as described in the chart below.
  • All adjustments to the Ratings and Projector reflect the average changes we saw from the study. For example, if the removal of blinkers led to an average of X less early speed, we now, on a one-time basis, project that horse to have X less early speed in the race in which he is going to remove blinkers.
  • Yes, we think we can do better than just these enhancements, but we’re pleased to have picked up the efficacy for our customers, and look forward to further improvements to Pace Projector efficacy from time to time.
  • Please note we’re in competition here. So as to benefit our paying customers, we’re not going to broadly release precisely how much we adjusted anything. Please contact us privately at support@timeformus.com with any questions.

Screen Shot 2014-02-18 at 1.39.17 PM

Learn More About…

Early and Late Ratings.

Pace Projector.

This entry was posted in Data Studies, Product features. Bookmark the permalink.

8 Responses to TimeformUS Pace Projector refined to include impact of blinkers, experience

  1. BTW, the BRISnet group absolutely refuses to answer emails. You have to telephone anything you want to ask. My compliments on your considered responses throughout the site. And no typos. Anywhere.

    Like

    • David G. says:

      I would email Ed DeRosa and if he doesn’t give you a prompt response Gauche, then we’ll give him a nudge

      Like

      • LOL. The last email I sent was to Ed! No response. I suggested updating the web page and I meant it in the best way. Hard to put feelings in emails. Many thanks…

        Like

  2. Everything I am reading on your site is literate and articulate. It is also enjoyable to read. I think I’ll interrupt my BRISnet all-you-can-eat subscription for a month and try TFUS. The quality of the site and quality of writing is compelling.

    Like

  3. marc@tfus says:

    We tested the same subset with and without these factors, and will periodically run these tests against new/random subsets.

    Like

  4. Cal says:

    Quick question: was the 10,000 race database on which you tested the efficacy of the new calibrations a subset of the 250,000 or so starts database that you used to refine and “calibrate” your ratings and projections, or were these two sets completely independent of each other?

    Like

  5. marc@tfus says:

    Pace Projector had a good weekend. We looked at FTL, good idea; we felt sample size was too small to make a move.

    Like

  6. David G. says:

    Hi guys,

    Any studies or comments regarding early spread projections regarding those runners, other than first time starters, showing up with first time lasix?

    Thanks.

    And by the way, if you check the pace projector from this past Sunday’s 8th race from Santa Anita, it was dead on accurate, as well as positioning the eventual winner Rathbaun in a perfect spot sitting in 3rd position behind two eary pacesetters while well clear of the rest of the field. That’s exactly how the race unfolded. You can literally copy and past the image of pace projector diagram and pasted it on the TV display showing the runners and their saddle cloth numbers underneath.

    Rathbaun, 12/1 on the morning line, paid a whopping $61.20 to win

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s