Turf figures or not?

[Gapfire]

What to do with Turf figures on other surfaces?

Currently, we have our system set up to include all races when evaluating a race. What ends up happening, is runners with turf lines end up on top of our rankings. We calculate our turf figs the same way we do other surfaces, but when the runners switch surfaces, their turf figures always seem higher.

Quite often these runners end up winning at a price. However, most times they do not. What do the sharp handicappers on this board do? Toss the turf efforts, and look for the next available race on today's surface, or evaluate the turf race, and downgrade the figure?

Thanks in advance,

Kevin
www.gapfire.com

[GaryG]

When you use turf form to evaluate a horse running on dirt you are asking for trouble. There are a few that have equal ability on different surfaces, but they are very few. A horse going from turf to dirt, or vice versa, is essentially a shipper, even though he is only shipping a few feet. Also, the requirements of the surfaces can be polar opposites.

[Gapfire]

When you use turf form to evaluate a horse running on dirt you are asking for trouble. There are a few that have equal ability on different surfaces, but they are very few. A horse going from turf to dirt, or vice versa, is essentially a shipper, even though he is only shipping a few feet. Also, the requirements of the surfaces can be polar opposites.

So you would just toss the turf effort? I ask in the context of database handicapping and automatic calculations of a race. In that we calculate all races at all tracks.

I don't have time to go through each race by hand, so I would have to set up the database to go to evaluate the horse on his next available dirt effort.

Every time I think about doing this, it seems like one of these guys ends up winning at a huge price. So, I haven't made the change yet.

[GaryG]

The turf race can be used as an indicator of current form, unless the horse simply hates the surface. In that case his turf races can be tossed.

[Gapfire]

The turf race can be used as an indicator of current form, unless the horse simply hates the surface. In that case his turf races can be tossed.

Thanks for the sound advice Gary.

[thaskalos]

So you would just toss the turf effort? I ask in the context of database handicapping and automatic calculations of a race. In that we calculate all races at all tracks.

I don't have time to go through each race by hand, so I would have to set up the database to go to evaluate the horse on his next available dirt effort.

Every time I think about doing this, it seems like one of these guys ends up winning at a huge price. So, I haven't made the change yet.
My experience has been exactly the opposite.

A horse shows repeated high recent speed figures on the turf, but finds itself on the DIRT today. A survey of its less recent dirt figures indicates that the horse is much more adept on the TURF than on today's dirt surface.

More often than not, the horse will be a much SHORTER price than his "dirt" form will warrant (as a result of those repeated high recent turf figures)...and it will very likely disappoint its backers.

Throw the turf races out, I say...and compare apples to apples.

You will be proven wrong once in a while...but you will be rewarded in the long run.

[Greyfox]

Throw the turf races out, I say...and compare apples to apples.

.

thaskalos - Would you recommend throwing turf out if the switch is to poly?

[gm10]

I would suggest, think a bit broader than just surface, think 'circumstances', which are influenced by a number of factors - such as pace, going, surface, draw, etc. Sure, certain surface/going combinations show a high correlation with certain circumstances, but that doesn't mean 'surface' equals 'circumstances'.

[Gapfire]

I would suggest, think a bit broader than just surface, think 'circumstances', which are influenced by a number of factors - such as pace, going, surface, draw, etc. Sure, certain surface/going combinations show a high correlation with certain circumstances, but that doesn't mean 'surface' equals 'circumstances'.

Yes, I agree. However, I'm looking for a way to deal with these surface changes automatically. I couldn't possibly go through each race, and make these determinations manually. Good thought though.

[thaskalos]

thaskalos - Would you recommend throwing turf out if the switch is to poly?
To be honest, Greyfox...I haven't come to a firm conclusion about this one.

It SEEMS that the turf is much more closely related to poly than to regular dirt...but in my handicapping, I still rely strictly on "grass form" when playing turf races.

If I find that the turf races feature accomplished poly horses who have never - or even rarely - raced on turf...then I will PASS these races. It just seems safer to me...

[gm10]

Yes, I agree. However, I'm looking for a way to deal with these surface changes automatically. I couldn't possibly go through each race, and make these determinations manually. Good thought though.

Not sure if this is what you're after but Equibase have an XML feed for 'scratches'/changes'. I wrote a small program that grabs it and updates all the relevant data, it works fine, never had any problems with it.

If you mean historical data, no, you'll have to come up with some 'measures' that capture the concept, you can't determine that manually for every horse/race.

[Gapfire]

Not sure if this is what you're after but Equibase have an XML feed for 'scratches'/changes'. I wrote a small program that grabs it and updates all the relevant data, it works fine, never had any problems with it.

If you mean historical data, no, you'll have to come up with some 'measures' that capture the concept, you can't determine that manually for every horse/race.

Thanks,
I just finished writing a similar script for our site. And, yes, I meant historical data. Concept capturing is tricky, but not impossible. It's just a matter of whether or not the time invested in writing the scripts is worth the return. That's why I like to have a very clear idea before I set out programming anything.

[gm10]

Thanks,
I just finished writing a similar script for our site. And, yes, I meant historical data. Concept capturing is tricky, but not impossible. It's just a matter of whether or not the time invested in writing the scripts is worth the return. That's why I like to have a very clear idea before I set out programming anything.

Hmmm I doubt you can have a clear idea at this point. It'll be more of an ongoing project imo. One thing you can invest in without any risk is infrastructure that allows to specify, store and swiftly analyze any new variables/measures that you wish to investigate. You need a reliable car for you travels ;) .

[raybo]

Not having any knowledge with what you are doing, I assume you are creating figures, etc.. In that case, I also assume you are using pars to enable the figs. First, you should be using dirt pars for dirt figs, turf pars for turf figs and poly pars for poly figs.
Now, in order to, for example, use turf figs for evaluating a dirt race, or vice versa, it seems logical that you would also have the ability to compare the pars for the different surfaces and apply the difference between these, to the figs, to get your rankings on the new surface.

In other words, downgrade or upgrade the figs, to better equalize the figs for a surface change.

Horses that don't like a certain surface would have to be dealt with differently, however.

[Gapfire]

Not having any knowledge with what you are doing, I assume you are creating figures, etc.. In that case, I also assume you are using pars to enable the figs. First, you should be using dirt pars for dirt figs, turf pars for turf figs and poly pars for poly figs.
Now, in order to, for example, use turf figs for evaluating a dirt race, or vice versa, it seems logical that you would also have the ability to compare the pars for the different surfaces and apply the difference between these, to the figs, to get your rankings on the new surface.

In other words, downgrade or upgrade the figs, to better equalize the figs for a surface change.

Horses that don't like a certain surface would have to be dealt with differently, however.

Yes Raybo, great idea. You are correct in your assumptions. I do maintain pars, and your method may be a great way to deal with turf to dirt figs, or any surface switch for that matter. Thank you.

[Gapfire]

Hmmm I doubt you can have a clear idea at this point. It'll be more of an ongoing project imo. One thing you can invest in without any risk is infrastructure that allows to specify, store and swiftly analyze any new variables/measures that you wish to investigate. You need a reliable car for you travels ;) .

I agree, Gapfire.com is an ongoing project, and I continually look for ways to improve our figs, and results. I do have the infrastructure which allows me to store, and analyze very quickly. That's why I posted here...looking for great ideas to investigate. The people here really do know their stuff. :ThmbUp:

[Greyfox]

To be honest, Greyfox...I haven't come to a firm conclusion about this one.

...

I agree. I haven't reached any firm decision either. I know in So Cal that a runner who does well on poly can move to turf and perform similarly.
However, it doesn't work as well the other way around.
That's why when it rains and they come off the turf, several of the "turf" intended types are scratched. I'm wondering if the scratches offer a clue as to who might run well and who won't. But that might be hard to keep track of in a data base.

[mistergee]

my 2 cents and firm conclusion, handicap turf to turf, dirt to dirt , poly to poly, and sprint to sprint, route to route. Your figures have to be out of whack otherwise. Also it seems more recent races should be weighted even more favorably. the horses last race is sometimes 2 weeks ago and his 2nd race back may be a year ago, or his last race with a new super trainer, and 2 race back with joe shmo. This is the danger of automated figures but you can adjust.

[JohnGalt1]

I cross out turf races if handicapping for the dirt and vice versa when making pace figures.

I do the same when calculating Scott's Performance Class Ratings since turf races are usually run at higher classes.

$5k dirt claimers will run in $20k claiming turf races it that's the lowest class for grass races at a track.

I do look at form from recent races if going from one to the other surface.