Making your own speed figs? It's multi-dimensional now or bust!

[Greyfox]

Originally Posted by raybo

Of course. In race A the horses ran the distance of a length faster than the horses in race B. So, a length in race A took less time to run than a length in race B. But, I was more talking about the fractional segments, independently. It's common sense, not rocket science.

I realize that both you and Cratos are theoretically quite right.
But adjusting a length time by 1/34* 1/5 sec * X lengths in the example given might not make much difference at the windows.
And yes, one could do it for fractional times where horses are running faster at the start of a race than the end. But once again I suspect it won't make your wallet much thicker.

[traynor]

The use of pace for wagering purposes requires that the values be predictive, not simply descriptive. The most excruciatingly detailed description of a past race is only useful for a bettor to the extent that it is predictive of what will happen in the future.

[traynor]

Originally Posted by Greyfox

I realize that both you and Cratos are theoretically quite right.
But adjusting a length time by 1/34* 1/5 sec * X lengths in the example given might not make much difference at the windows.
And yes, one could do it for fractional times where horses are running faster at the start of a race than the end. But once again I suspect it won't make your wallet much thicker.

A key point is not just a value describing how fast a certain horse was moving in a given segment of a given race, but rather the ability to make meaningful comparisons between rates of speeds in different length races, and of overall performance considering the rates of speed in differing segments of each race.

Those distinctions will definitely thicken your wallet. The proliferation of computer-assisted pace calculations has made it easier--not tougher--for anyone willing to devoid a bit of thought and analysis to the scenarios to stay ahead of the crowd, most of whom fail to grasp the difference between descriptions and prescriptions.

[Magister Ludi]

Originally Posted by clemkadiddle

The tool needs to be able to calculate that a faster a horse runs, the rate at which energy is expended occurs "exponentially".

For example, a horse running 56.2 FPS will expend twice the energy as a horse running 51.2 FPS over the same distance. Algebraically one may think that this is not an accurate assessment, but it is. My model will show this consistently.

Over the same distance, a horse running 56.2 f/s expends approximately 4% more energy than a horse running 51.2 f/s, not 100% more.

[DeltaLover]

Originally Posted by Magister Ludi

Over the same distance, a horse running 56.2 f/s expends approximately 4% more energy than a horse running 51.2 f/s, not 100% more.

So you say that energy relates to speed? My understanding is that the amount of energy needed to move between two points is not relative to the speed. Can you explain?

[classhandicapper]

I think worrying about beaten lengths calculations that will typically have virtually very little impact on your opinion is kind of like missing the forest for the trees.

If you compare speed figure sources there are sometimes differences of close to a full second for the same horses. That's HUGE. There are many differences of a couple of fifths. On top of that, it's common for beaten horses to be eased late or at least not persevered with fully to the finish.

So if you are worried about accuracy, IMO, you'd be way better off putting your extra energy into ensuring you have accurate track variants and have watched the races than worrying too much about minescule beaten length calculations, especially since other than Trakus tracks, it's all an estimate anyway (and even Trakus is not perfect). That doesn't mean you should working with the outdated and inaccurate 1/5 = 1 length, but if you are close to accurate that's good enough.

[cj]

Originally Posted by clemkadiddle

Here's the problem to your hypothetical: What if horse A was coming off the bench and ran a sparkling half-mile on the lead and then shut down? This is precisely the problem using speed ratings based on final time and parallel time.

Then there is no way to compare A with Q. Me...I would take A in a heartbeat because:

1. A was involved in the famous "pace race" theory described in Bob Heyburn's "Fast and Fit" horses.

2. A horse making his 2nd start off the bench...especially involved in a "pace race"...is a dangerous animal.

Of course there are a lot of reasons horses with lower figures win, but it doesn't excuse being sloppy with beaten lengths adjustments. My example was just showing how an error could schew comparisons.

[classhandicapper]

Originally Posted by traynor

The use of pace for wagering purposes requires that the values be predictive, not simply descriptive. The most excruciatingly detailed description of a past race is only useful for a bettor to the extent that it is predictive of what will happen in the future.

This is kind of what I was getting at earlier with Clem.

Horse A goes 22.8 22.8 24.4 for a total of 1:10

Horse B goes 22 22.6 25.4 for a total of 1:10

Everyone on earth knows that horse B ran the superior race because it ran a much faster pace. But in many cases they will run very similar fractions again next time and finish very close again.

It's not enough to measure past fractions and create a rating. You also have to know when those fractions deviated significantly from the horse's norm so that a change back might lead to an improved result.

[raybo]

Originally Posted by Greyfox

I realize that both you and Cratos are theoretically quite right.
But adjusting a length time by 1/34* 1/5 sec * X lengths in the example given might not make much difference at the windows.
And yes, one could do it for fractional times where horses are running faster at the start of a race than the end. But once again I suspect it won't make your wallet much thicker.

That's not the way I do it, but, whatever.

My point is: If you're going to go to the trouble to adjust times by beaten lengths, you should do it as accurately as possible. It costs you no more to do it better, unless you are doing it with pen and pencil (which I would not recommend for obvious reasons). In an app you only create the formulas once, then it's done forever.

There is no need to include 1/5 second in any of your calculations, as Cratos did. You just need distance traveled in feet, the average length of a horse in feet (8', 9', 10', doesn't matter as long as you use the same value every time), and the leader's time in seconds. Of course, it is important to have the daily variant, and I suggest breaking that down into fractional variants. Then, if you wish, convert those fps calculations to a traditional speed/pace figure scale.

Traynor is correct, the accuracy of the adjustments is very important, and definitely affects the thickness of your wallet, depending on how you are using the adjusted times. If your method is flawed, more accurate times will not matter, you will still lose money.

[Dave Schwartz]

Horse A goes 22.8 22.8 24.4 for a total of 1:10

Horse B goes 22 22.6 25.4 for a total of 1:10

Everyone on earth knows that horse B ran the superior race because it ran a much faster pace. But in many cases they will run very similar fractions again next time and finish very close again.

Actually, I disagree. I believe A ran the better race.

My reasoning is that horses who run faster early times are supposed to run faster final times. The fact that horse A ran 4 ticks slower means he should have wound up with a slower final time.

[Cratos]

Originally Posted by DeltaLover

So you say that energy relates to speed? My understanding is that the amount of energy needed to move between two points is not relative to the speed. Can you explain?

The main productive function of the racehorse is work.

Therefore the conversion of chemically bound energy (energy from food/medications) is by work (speed) into mechanical energy for the horse's muscular movement.

However the effect of the horse's speed on the horse's energy use does not appear to be linear.

[cj]

Originally Posted by Dave Schwartz

Actually, I disagree. I believe A ran the better race.

My reasoning is that horses who run faster early times are supposed to run faster final times. The fact that horse A ran 4 ticks slower means he should have wound up with a slower final time.

For me it depends on how the horse ran fast earthly, pressured or alone.

[Capper Al]

Originally Posted by traynor

That works best. At least it has for me. Fortunately, I am not corrupted by what I read and do not feel compelled to emulate the author(s). So I can read everyone's ideas, try them, and if they work, use them. If not, I can freely (and without regret) dismiss them as irrelevant to my goals. That saved me from many years of turmoil and losing that others seem to have suffered trying to emulate the portions of Beyer's writings they want to believe, while ignoring or rejecting those they do not want to believe.

It's game of what works for you.

[raybo]

Originally Posted by Dave Schwartz

Actually, I disagree. I believe A ran the better race.

My reasoning is that horses who run faster early times are supposed to run faster final times. The fact that horse A ran 4 ticks slower means he should have wound up with a slower final time.

I agree, for the most part. The final times were equal, so if no other factors entered into the running of the race (traffic, etc.) the horses had equal performances, they both hit the wire at the same time. A lot of this hinges on what the race means, regarding future races, with different field dynamics. That is where accurate calculations means something.

[Cratos]

Originally Posted by raybo

That's not the way I do it, but, whatever.

My point is: If you're going to go to the trouble to adjust times by beaten lengths, you should do it as accurately as possible. It costs you no more to do it better, unless you are doing it with pen and pencil (which I would not recommend for obvious reasons). In an app you only create the formulas once, then it's done forever.

There is no need to include 1/5 second in any of your calculations, as Cratos did. You just need distance traveled in feet, the average length of a horse in feet (8', 9', 10', doesn't matter as long as you use the same value every time), and the leader's time in seconds. Of course, it is important to have the daily variant, and I suggest breaking that down into fractional variants. Then, if you wish, convert those fps calculations to a traditional speed/pace figure scale.

Traynor is correct, the accuracy of the adjustments is very important, and definitely affects the thickness of your wallet, depending on how you are using the adjusted times. If your method is flawed, more accurate times will not matter, you will still lose money.

Ii used the 1/5 second metric because that what was under discussion. In the thread; in our model we use one-hundredth of a second and when we use Equibase/DRF data we always use 8 feet/length because that is the metric on their website.

However all of this is for naught because we do parametric modeling and not discrete calculations.