Ask Andy Beyer

[mountainman]

Originally Posted by classhandicapper

There's a new version being developed. The early BETA is available at the homepage and has that feature .

Tech idiot here. Please explain, sir. For instance, will the new version incorporate that feature???

[mountainman]

Incidentally. Beyer should put me on his xmas list for the 20x a night I invoke his figs. In fact, I almost changed my name here once to "Antibrisman," but that just doesn't roll smoothly off the tongue.

Simple guy here, so a bottle of Old Spice will do, Andy. Chix dig it.

[mountainman]

Much appreciated Beyer's tasteful homage to the incomparable Glenn Gallivan, by the way. I loved the guy.

[steveb]

Originally Posted by cj

I don't think Beyer is using pars very often any longer to make figures. It is a good place when you are first starting out, but most then move on to projections based on the quality of the field. If I were starting over today, i wouldn't even begin with pars. I'd start with someone else's figures that I thought were of good quality.

we would disagree there.
i don't think i am just starting out, and there is no way i would use a projection method.
i guess though it depends on your definition of 'projection' as i know how fast any particular race should be run before the event(but it has NOTHING to do with how fast the runners in this race may have run previously), so maybe that would be a 'projection' although i don't see it as such.


pars(or in my case standards, because they are not averages as such) are fine.
if you are projecting, then that is basically saying that you expect this race to be run in that time, or something similar.
not so different to me, but mine will never be changed after the fact.

[classhandicapper]

Originally Posted by mountainman

Tech idiot here. Please explain, sir. For instance, will the new version incorporate that feature???

Yes. There's a notes section related to each horse for today's race that you can use for whatever you want.

Go to the home page. On the left, set it to PPs (instead of wagering) and select a race card. That will take you into the new Classic PPs. Then there's a tab for going into Formulator Beta.

Once you get into the new Formulator beta PPs, there's a box on the right near the earning box for notes.

[mountainman]

Originally Posted by classhandicapper

Yes. There's a notes section related to each horse for today's race that you can use for whatever you want.

Go to the home page. On the left, set it to PPs (instead of wagering) and select a race card. That will take you into the new Classic PPs. Then there's a tab for going into Formulator Beta.

Once you get into the new Formulator beta PPs, there's a box on the right near the earning box for notes.

I get that, Wayne. And tx. But I'm asking if the new Formulator will allow me to print out a card with PRE-RACE annotations BEFORE the races have been run.

[classhandicapper]

Originally Posted by mountainman

I get that, Wayne. And tx. But I'm asking if the new Formulator will allow me to print out a card with PRE-RACE annotations BEFORE the races have been run.

I don't think that's available yet. I don't know if it's in the long term plans. You can print the regular Card, Trip, and Race notes in the PPs. I let someone know what you are asking for.

[bobphilo]

Originally Posted by cj

You just try to gauge what you expect the horses to run based on past performance, not some vague par which is really a problem these days with so many different conditions even within the same class. It is tough but over time you can get really good at it. I start with the last four races of the top eight finishers. I toss out races that aren't similar enough to today based on surface, distance, finish position (both today and the races I'm comparing) and how long ago it happened. You can get a big sample sizes this way compared to the small samples you get using pars.

If there are 5 sprints run on a card, I have a chance at 32 data points per race for a total of 160. Of course that rarely happens given today's field sizes and how little horses actually race, but I might have a good 80. I'll narrow it down by tossing out the the possible outliers, so maybe the 15 highest and 15 lowest, and base the track variant on the remaining 50.

The trick with projections is you have to be careful not underrate improving horses. The more horses you use the less likely that is to happen as you'll be using horses that regress as well and again tossing out the outliers before making a variant.

Excellent explanation of the projection method. Takes more time than pars but yields much more accurate figures. Beyer said that he used pars when he was just starting out and had nothing else to work with but switched to projections when he had generated some numbers.

[bobphilo]

Originally Posted by cj

You just try to gauge what you expect the horses to run based on past performance, not some vague par which is really a problem these days with so many different conditions even within the same class. It is tough but over time you can get really good at it. I start with the last four races of the top eight finishers. I toss out races that aren't similar enough to today based on surface, distance, finish position (both today and the races I'm comparing) and how long ago it happened. You can get a big sample sizes this way compared to the small samples you get using pars.

If there are 5 sprints run on a card, I have a chance at 32 data points per race for a total of 160. Of course that rarely happens given today's field sizes and how little horses actually race, but I might have a good 80. I'll narrow it down by tossing out the the possible outliers, so maybe the 15 highest and 15 lowest, and base the track variant on the remaining 50.

The trick with projections is you have to be careful not underrate improving horses. The more horses you use the less likely that is to happen as you'll be using horses that regress as well and again tossing out the outliers before making a variant.

I used to use the same method except I rarely went beyond the top 4 or 5 finishers unless they were close. Horses finishing far back are usually not preservered with and are probably not running representative races. I know this gave me a smaller sample but a more representative one.

[classhandicapper]

Originally Posted by mountainman

I get that, Wayne. And tx. But I'm asking if the new Formulator will allow me to print out a card with PRE-RACE annotations BEFORE the races have been run.

Update.

That feature is already in place.

[dilanesp]

Originally Posted by bobphilo

I used to use the same method except I rarely went beyond the top 4 or 5 finishers unless they were close. Horses finishing far back are usually not preservered with and are probably not running representative races. I know this gave me a smaller sample but a more representative one.

Yes, and that speaks to a problem I have always had with figure gurus, including Beyer. They assume- and advocate- that a 77 run in a performance where a horse is far back means the same thing as a 77 run in a winning performance.

It actually could either be too low OR too high. Too low because the horse wasn't perservered with. But also too high for a reason any track and field fan can tell you- that runners often run their best times in races where they finish off the board, because the better runners "pull them along" and pace them to a faster time.

The more a horse is beaten, the less you can trust the number as a handicapping tool.

[classhandicapper]

Originally Posted by dilanesp

It actually could either be too low OR too high. Too low because the horse wasn't perservered with. But also too high for a reason any track and field fan can tell you- that runners often run their best times in races where they finish off the board, because the better runners "pull them along" and pace them to a faster time.

The more a horse is beaten, the less you can trust the number as a handicapping tool.

I agree.

The horses that don't finish in the top few are sometimes horses that were used early trying to win but got put away by superior horses. So they rarely run the best final time figure they are capable of. That goes double if they were already well out of the "purse" group and get eased late by the rider.

The latter is WAY more apparent in harness racing where horses that were in the back and made no real move into contention sucked up late with a faster final time than they could possibly earn if they moved earlier and actually tried to win the race.

[Tom]

That is why they let us bet on the races.
You have to handicap the race to know how to look at the figure.

[mountainman]

Originally Posted by classhandicapper

Update.

That feature is already in place.

Didn't work for me, sir. How might it be done?? (starting to sound like Dr Evil here: "All I want is sharks that shoot frikken laser beams out of their heads!!")

[AndyC]

Originally Posted by dilanesp

Yes, and that speaks to a problem I have always had with figure gurus, including Beyer. They assume- and advocate- that a 77 run in a performance where a horse is far back means the same thing as a 77 run in a winning performance.

It actually could either be too low OR too high. Too low because the horse wasn't perservered with. But also too high for a reason any track and field fan can tell you- that runners often run their best times in races where they finish off the board, because the better runners "pull them along" and pace them to a faster time.

The more a horse is beaten, the less you can trust the number as a handicapping tool.

Beyer makes no such assumption. He is not making performance numbers, just speed figures. Factors used that may enhance or detract from the numbers is up to each individual handicapper.