A simple toy problem for variant makers

[Sapio]

I have a collection of horses for which which I know their performance parameters. I simulate a race and provide you with a final time. To this final time I might or might not have a added a variant. This information is unknown to you.

As costumary you have access to each horse's past performances.

How do you determine the variant?

Once again I simulate the race and provide you a final time for the race. Again I may have added the same variant, different variant or no variant.

How do you determine the variant?

Thomas Sapio

[Sapio]

Hmm...not a single response. Am I to assume the question is difficult?

Thomas Sapio

[cj]

Hmm...not a single response. Am I to assume the question is difficult?

Thomas Sapio

I'd probably assume most people were sleeping.

[Robert Goren]

probably most people who do daily variants use par times. Doing a projection method looks like a lot more work than most horse players want to do.

[dannyhill]

Reads like somebody fishing to find out how to make a good variant.

[Sapio]

Reads like somebody fishing to find out how to make a good variant.

Hi dannyhill

I can assure you, I'm not fishing.

Thomas Sapio

[dannyhill]

Hi dannyhill

I can assure you, I'm not fishing.

Thomas Sapio
Hi Sapio

Best of Luck

[AndyC]

Hmm...not a single response. Am I to assume the question is difficult?

Thomas Sapio

Assume that we considered it a rhetorical question. It makes us look smarter.

I am eagerly awaiting your answer. Interesting question.

[classhandicapper]

I have a collection of horses for which which I know their performance parameters. I simulate a race and provide you with a final time. To this final time I might or might not have a added a variant. This information is unknown to you.

As costumary you have access to each horse's past performances.

How do you determine the variant?

Once again I simulate the race and provide you a final time for the race. Again I may have added the same variant, different variant or no variant.

How do you determine the variant?

Thomas Sapio

IMO you need more information, but maybe I'm missing something.

Assume these are horses than can run 6F in 1:12 adjusted for track speed.

If all you given is a time of 1:12 (and you don't know if it was adjusted or not) there's no way of knowing for sure.

IF you are given 1:12 as the time, the actual time could have been 1:09 adjusted to 1:12, 1:15 adjusted to 1:12, or actually 1:12.

However, if you know these horses are typically 1:11 horses or 1:14 horses, (anything other than the 1:12 you were given) then you at least know you have been given a raw time.

So you can say the track was 1 second fast, 2 seconds slow etc...

[cj]

I have a collection of horses for which which I know their performance parameters. I simulate a race and provide you with a final time. To this final time I might or might not have a added a variant. This information is unknown to you.



The key is knowing the performance parameter. Classhandicapper sums up the rest nicely.

[Robert Goren]

I am assuming that he has a range for each horse. say at 6f 1:11 2/5 -1:12 1/5 for horse A , 1:10 4/5 -1:11 4/5 for horse B etc. From those range, then come with a daily variant so that when the variant is applied to todays time that most most of the runners fit into their range. You could kind of eyeball a number, but there should be a way to have a program do it for you. When you add ranges instead of a hard number for a horse, things get a little more tricky, but you probably would gain some accurucy. I doubt if very many posters, if any, use this method. I would be very interested to know how much better the numbers are using this method. I thought about playing with it about 30 years ago, but decide it was too much work back then. The advent of the PC and Excel would make it much easier today. Ah, to be 30 years younger with today's tools.

[Sapio]

The key is knowing the RACE performance parameter. Classhandicapper sums up the rest nicely.

(I added "RACE")

Variant making approaches generally posit that the race is one way (e.g., a parameter has one particular true value), then conducts an experiment (either par tables, projection, etc) whose resulting conclusion -- no matter the true value of the parameter determines a variant.

Thomas Sapio

[tucker6]

(I added "RACE")

Variant making approaches generally posit that the race is one way (e.g., a parameter has one particular true value), then conducts an experiment (either par tables, projection, etc) whose resulting conclusion -- no matter the true value of the parameter determines a variant.

Thomas Sapio
If you started this thread in hopes of getting CJ to respond so that you could reply, why not simply send him a PM?

[cj]

(I added "RACE")

Variant making approaches generally posit that the race is one way (e.g., a parameter has one particular true value), then conducts an experiment (either par tables, projection, etc) whose resulting conclusion -- no matter the true value of the parameter determines a variant.

Thomas Sapio

That would be true if races were considered in isolation. They usually are not, though there are times where they must be.

I'll leave this discussion with this...if somebody wants to make figures based only on the times each day, and the person actually bet money on those figures, he or she will soon be broke. That is about a sure a bet as you will find. People that don't believe that either a) have never made figures, b) don't bet, c) both a and b.

[Sapio]

That would be true if races were considered in isolation. They usually are not, though there are times where they must be.

I'll leave this discussion with this...if somebody wants to make figures based only on the times each day, and the person actually bet money on those figures, he or she will soon be broke. That is about a sure a bet as you will find. People that don't believe that either a) have never made figures, b) don't bet, c) both a and b.

Hi cj

How you draw that conclusion based om my posts is perplexing. If anything I am suggesting greater uncertainity.

Thomas Sapio

[cj]

Hi cj

How you draw that conclusion based om my posts is perplexing. If anything I am suggesting greater uncertainity.

Thomas Sapio

How about being direct. What is your purpose for starting this thread?

[whodoyoulike]

I have a collection of horses for which which I know their performance parameters. I simulate a race and provide you with a final time. To this final time I might or might not have a added a variant. This information is unknown to you.

As costumary you have access to each horse's past performances.

How do you determine the variant?

Once again I simulate the race and provide you a final time for the race. Again I may have added the same variant, different variant or no variant.

How do you determine the variant?

Thomas Sapio

I've read your post and this thread a couple of times. What info or purpose are you attempting to convey?

I'm just not understanding. Thanks.

[classhandicapper]

Hi cj

How you draw that conclusion based om my posts is perplexing. If anything I am suggesting greater uncertainity.

Thomas Sapio


It's difficult to understand what you are trying to get at.

I would suggest you read some of Beyer's early books to become familiar with what everyone else is doing in a general sense (if you haven't already). That doesn't mean you have to accept the charts (which are outdated) or even the methodology. But it might make for easier conversation if you could point to something specific that we can discuss. Many of the methodology issues have been discussed to death on this and other forums.

[whodoyoulike]

How about being direct. What is your purpose for starting this thread?

And, I thought it was just me.

[PhantomOnTour]

Nineteen posts on this thread and the word "Bayesian" hasn't come up yet.

I am sure that's where this thread is going

[cj]

Nineteen posts on this thread and the word "Bayesian" hasn't come up yet.

I am sure that's where this thread is going

I figure as much, but I'm not closing it just yet.

[Sapio]

I figure as much, but I'm not closing it just yet.

Hi cj

If I may ask, why would the thread be closed?

Thomas Sapio

[headhawg]

If I may ask, why would the thread be closed?Are you trying to emulate Mad Scientist?

[cj]

Hi cj

If I may ask, why would the thread be closed?

Thomas Sapio

How about you answer this first:

How about being direct. What is your purpose for starting this thread?

[PaceAdvantage]

Hi cj

If I may ask, why would the thread be closed?

Thomas SapioBecause it's going nowhere...try getting to the point and stop delighting in trying to make people dance for you.

[BetHorses!]

Pet Peeve - people who sign their name after every post...

[Sapio]

I am assuming that he has a range for each horse. say at 6f 1:11 2/5 -1:12 1/5 for horse A , 1:10 4/5 -1:11 4/5 for horse B etc. From those range, then come with a daily variant so that when the variant is applied to todays time that most most of the runners fit into their range. You could kind of eyeball a number, but there should be a way to have a program do it for you. When you add ranges instead of a hard number for a horse, things get a little more tricky, but you probably would gain some accurucy. I doubt if very many posters, if any, use this method. I would be very interested to know how much better the numbers are using this method. I thought about playing with it about 30 years ago, but decide it was too much work back then. The advent of the PC and Excel would make it much easier today. Ah, to be 30 years younger with today's tools.

Hi Robert

You might also determine that often what appears to be a variant is merely performance variation or uncertainity in knowing the performance.

Thomas Sapio

[cj]

Hi Robert

You might also determine that often what appears to be a variant is merely performance variation or uncertainity in knowing the performance.

Thomas Sapio

Well, you've ignored the site owner and a moderator. So when you wonder why the thread was closed, that is your answer.