final time or not?

[imofe]

Originally Posted by Dave Schwartz

When one looks for correlation between known factors and the tote board, it becomes pretty obvious right away that the public uses speed ratings very heavily in its handicapping.

Dave,

So does that mean you do not use them?

[traynor]

Does TM mean TrackMaster?

[DeltaLover]

Originally Posted by Ray2000

imofe

I use the trend of Final Times + 4thQ Time as a measure current Form. Form contributes about 20% to the final rating for each entry as does Trip and Driver.

All Final and Quarter Times are adjusted for Distance covered including 'racing wide' Track Speed Rating, Track Condition (Variant), Driver and Starting Post/Tier. Some lines get rejected and no lines older than 40 days are used.

Biggest problem... "Form" has peaked........but I live with it...

Ray,

I am a similar approach as your breakdown of your handicapping factors (Form, Trip etc) with one difference:

I am not using the absolute value of each factor but a derivative of it considering the competition.

For example, assuming that we use a form figure f, if all starters happen to have the same f then form is no longer considerable...

There are more than one methodologies to implement this idea, like zscores, robust zscores, logistic normalization etc. The important thing is to use a relative as opposed to an absolute value...

Following this approach your representation of the race will become a matrix N X M assuming N factors and M Starters

You need to multiply the matrix by a vector having N elements to generate a final composite rating that you can later transform to a probability or use it as it is....

[DeltaLover]

Originally Posted by imofe

Ok Ray

Do you use all the races in the last 40 days?

Or the horse's best time in that 40 days?

Or the line you think would best represent the horse?

Or something else?

These can become parametrical attruibutes to be optimized

[Ray2000]

imofe

I (the program) uses at least 3 and up to 5 (rarely 6) lines in the last 40 days. no Quals or "distanced" or Break lines. I then use a trend (linear regression) to project what the form number will be in the upcoming race. Works best with young horses moving up in class.


traynor... yes, TM is Trackmaster I use their average Class and Speed in a minor way.

DeltaLover

Good Point, I didn't give the units for the various factors because that takes some explaining.

I decided at the onset to try and normalize all factors to the standard length or 1/5 of a second. Times and lengths are easy, just use the number. Driver rating is a bit trickier. The old saw was "A switch to a good driver can take a second off a horses final time." I used that as the metric and tried to come up with other equivalents for the other factors.

You T-Bred Guys sure come up with a lot of interesting topics, I enjoy reading them.

[DeltaLover]

Originally Posted by Ray2000

Good Point, I didn't give the units for the various factors because that takes some explaining.

All data used for automated handicapping should be normalized using some method and this is what I consider the most challenging and important part rather the accuracy of the figure or the way it is processed..

With this I mean that an insignificant or inaccurate figure will be easily detected and either improved or completely removed assuming a sound normalization approach..

The normalization procedure largely consist of trial and error experiments as it is impossible to analytically decide which method will be more suitable for any specific metric...

[eurocapper]

I don't know about mixing unrelated quantities (and attempting to quantify drivers to boot). Some drivers better with some horses than others. Seems better to me to set a par value for some of the factors instead of mixing them and aiming for the ultimate number.

[Ray2000]

The reason for that approach vs Par constructions, was my goal was to produce an oddsline and not to separate out the top contenders. So I needed some meaningful way to combine the various factors so that each horse ends up with a numerical rating. Then a fair (at least usable) odds line would be given by.

(Starters-1) * e^(-1*(Rating-mean)/StDev of Ratings)

Trouble is some factors refuse to be scaled.
How do you equate the tendency to break to any standard unit?

Still, the proof of the pudding is in the cashing.....

GL

[jefftune]

I do not pay any attention to final time. I figure last half times and only for horses who made up ground in their races. I think most harness players do look at final time so I want to zig when they zag.

-Jeff Tune
www.angelfire.com/pa/tune

[imofe]

Well, except for Ray and Delta, who only use final times as a part of their form handicapping, nobody seems to want to step up and defend the use of final times in either identifying contenders or as a tool when deciding which contender to bet. I was hoping to hear from someone that uses final times successfully. Maybe someone that thinks final times are better on bigger or smaller tracks or individual tracks. Or maybe how final times can be used but only when some value exists.

[traynor]

Originally Posted by imofe

Well, except for Ray and Delta, who only use final times as a part of their form handicapping, nobody seems to want to step up and defend the use of final times in either identifying contenders or as a tool when deciding which contender to bet. I was hoping to hear from someone that uses final times successfully. Maybe someone that thinks final times are better on bigger or smaller tracks or individual tracks. Or maybe how final times can be used but only when some value exists.

That sounds like really valuable information. It is unlikely that anyone smart enough to figure all that out could be coaxed into laying it in the street in exchange for a pat on the head and a little "Ya done good. Got anything else for us?" If anyone did have it, it is unlikely they would give it away. It is even more unlikely that they would sell it.

If they did have the information, the money would be easy to get. That is, if anyone did have the information.

[imofe]

Man that pat on the head really cracked me up . Thanks. I needed that after football today.

But I was talking more about a general belief or a suggestion that it can and is being done then a step by step lesson. As I mentioned before, I do not use final times. But I always try to keep an open mind.

[botster]

Some of these T-Bred guys have my head spinning.I don't think I can match their intellect and overall knowledge of their answers,but I can give you my two cents.Harness as everyone knows here is much more simplified.Lucky for me, because I am a man of limited mathmatical head smarts...hahaha.

Final times are important to a degree,but hold very little weight when handicapping a harness race overall.I liked one guys answer here,he said something like "Todays harness racing is like a mad dash of speed to get to the finish first".I liked it, because he was correct in this big respect.This is what mainly has hurt the game over the last 20 years.Mostly all the topnotch horses today in this game are racing "dirty" and their are many trainers in the sport today that continue to race dirty.This will cause the drop in average final times AND the individual horses final times racing behind
the winners in the sport.This ultimately has made final times inconsequential in many respects when mainly handicapping the better quality class of horse.

Times are only important to me, when I can see one "airing it out" or making multiple moves in a race.Saying this,I would say final times would have to be considered only, when I observe one doing the latter.

Good Fortune

[mrroyboy]

Ok here is my answer. I use TM speed ratings and final times to eliminate non contenders. But I don't make a final decision using them. I think that is the way many successful harness betters do it.I suggest you buy Pandy's book which details alot of this.

[imofe]

Originally Posted by mrroyboy

Ok here is my answer. I use TM speed ratings and final times to eliminate non contenders. But I don't make a final decision using them. I think that is the way many successful harness betters do it.I suggest you buy Pandy's book which details alot of this.

Thanks for the answer. Now from what you are saying, I'm taking it that horses that have ran slower times (or TM speed ratings) regardless of trip, finish position, or class are eliminated. I guess I could see that at tracks where the times are fairly consistent. Would not want to use that at half mile tracks where horses can be victims of others setting the pace or bad posts.