Brisnet -- DRF Speed Rate + Variant

[horseplayer]

Brisnet - carries the DRF speed rate and track variant and this can be printed with their past performance generator - 2 tracks I downloaded today Golden Gate and Laurel are missing the speed rate and variant for any race after march 18 - 2018 called Brisnet and they said they are aware of this .

[JohnGalt1]

I called them Friday and they said it was a glitch in data received from Equibase.

Tech is working on it.

[PaceAdvantage]

Maybe they left it out on purpose to see if anyone would notice...maybe they want to do away with it and this was their beta testing...

[JohnGalt1]

I asked that question. They said no, they still make them.

I looked at DRF charts and they still make a variant for all races.

[Denny]

If someone at DRF would just improve the Variant calculation, we could do away with the 'Beyers'.

But, everybody already knows that. Right?

[cj]

Originally Posted by Denny

If someone at DRF would just improve the Variant calculation, we could do away with the 'Beyers'.

But, everybody already knows that. Right?

Nope, not right actually. A problem just as big with the DRF Speed Ratings is that a fifth is counted as a fifth, regardless of distance.

I also don't think you can beat a variant done with human oversight with a strict math formula. There are a lot of reasons for this, but I'll start with timing and data entry errors.

[Denny]

There are errors built in to all speed figures.

An accurate variaint is high on the list. Especially when your working with so few races at the same distance and surface.

The 1/5 is minor compared. I'd be happy with races timed in 1/10 and adjusted with different speed for different lengths (1/100 is overkill to me). Plus it robs horses of track records for a small silly difference.

An improvement would be using VELOCITY instead, using actual distance covered. Then there's all the other things being ignored when figs are calculated.

The projections are pure bs to me. Humans make more mistakes than computers.

Speed figs are still in their infancy imo and have a long way to go.

But, I do consider yours an improvement.

[cj]

Originally Posted by Denny

There are errors built in to all speed figures.

An accurate variaint is high on the list. Especially when your working with so few races at the same distance and surface.

The 1/5 is minor compared. I'd be happy with races timed in 1/10 and adjusted with different speed for different lengths (1/100 is overkill to me). Plus it robs horses of track records for a small silly difference.

An improvement would be using VELOCITY instead, using actual distance covered. Then there's all the other things being ignored when figs are calculated.

The projections are pure bs to me. Humans make more mistakes than computers.

Speed figs are still in their infancy imo and have a long way to go.

But, I do consider yours an improvement.

I agree that they can and will be improved, but right now are limited by the data available in my opinion. Velocity is being used. Not sure what you are getting at with that one.

[Denny]

Where is velocity being used in speed figures? Other than what Sartin did.
Trakus shows velocity, put it's not part of PP's.
So where do I find a speed figure with velocity included?

[Dave Schwartz]

Originally Posted by Denny

Where is velocity being used in speed figures? Other than what Sartin did.
Trakus shows velocity, put it's not part of PP's.
So where do I find a speed figure with velocity included?

My new software will have it. Should be Beta in May.

[cj]

Originally Posted by Denny

Where is velocity being used in speed figures? Other than what Sartin did.
Trakus shows velocity, put it's not part of PP's.
So where do I find a speed figure with velocity included?

Point is they are the same thing expressed a different way. A 120 could just as easily be shown as 55.25 fps. What difference does it make?

[pandy]

Originally Posted by Denny

There are errors built in to all speed figures.

An accurate variaint is high on the list. Especially when your working with so few races at the same distance and surface.

The 1/5 is minor compared. I'd be happy with races timed in 1/10 and adjusted with different speed for different lengths (1/100 is overkill to me). Plus it robs horses of track records for a small silly difference.

An improvement would be using VELOCITY instead, using actual distance covered. Then there's all the other things being ignored when figs are calculated.

The projections are pure bs to me. Humans make more mistakes than computers.

Speed figs are still in their infancy imo and have a long way to go.

But, I do consider yours an improvement.

That's true, there are errors, and it's not just the track variant. We also use parallel time charts to compare times at different distances, various run up times that can change, track pars, which also change as track surfaces are groomed and changed, and creating these different aspects of the speed fig process is all somewhat of a best guess situation. This year is the first year of the "winterized" Aqueduct main track, which has played slower and more "sustained" that the regular Aqueduct main track. And I wonder if and when NYRA is going to change the track back to the regular main track or keep the extra material on it???

A good way to make sure that the horse you're betting has an reliable figure is to compare figures from several different sources. If they don't all agree that your horse ran a fast race last time, proceed with caution.

[pandy]

Originally Posted by Denny

There are errors built in to all speed figures.

An accurate variaint is high on the list. Especially when your working with so few races at the same distance and surface.

The 1/5 is minor compared. I'd be happy with races timed in 1/10 and adjusted with different speed for different lengths (1/100 is overkill to me). Plus it robs horses of track records for a small silly difference.

An improvement would be using VELOCITY instead, using actual distance covered. Then there's all the other things being ignored when figs are calculated.

The projections are pure bs to me. Humans make more mistakes than computers.

Speed figs are still in their infancy imo and have a long way to go.

But, I do consider yours an improvement.

Velocity ratings, FPS, for instance, have their own inaccuracies. First of all, with any rating that incorporates the fractions of the race, and not just the final time, that means you're using times that were created by whoever was calling the charts. So you're relying on someone with binoculars who is trying to call twelve horses at once.

And any velocity ratings I've seen are more dependent on the actual raw time than speed and pace figures. Some see that as a bad thing, although sometimes it is an improvement. For instance, I've seen horses run 6 furlongs in 1:09 and get a speed figure of 80, because whoever did the speed figure thought the track was very fast that day. On another day, a horse ran 6 furlongs in 1:12 and also got a speed figure of 80, because the track was slow that day. In the pps, if these two horses get into the same race, the speed figure is going to show that they are equal. But the velocity ratings, even though they're adjusted for track speed, will always make it look like the horse that ran the 1:09 was clearly faster than the horse than ran 1:12. Whether or not this is a good or bad thing is debateable. I tend to prefer to bet on horses that actually ran fast, as opposed to horses that may have run fast if the track was faster, but others disagree with that way of thinking.

[cj]

Originally Posted by pandy

Velocity ratings, FPS, for instance, have their own inaccuracies. First of all, with any rating that incorporates the fractions of the race, and not just the final time, that means you're using times that were created by whoever was calling the charts. So you're relying on someone with binoculars who is trying to call twelve horses at once.

And any velocity ratings I've seen are more dependent on the actual raw time than speed and pace figures. Some see that as a bad thing, although sometimes it is an improvement. For instance, I've seen horses run 6 furlongs in 1:09 and get a speed figure of 80, because whoever did the speed figure thought the track was very fast that day. On another day, a horse ran 6 furlongs in 1:12 and also got a speed figure of 80, because the track was slow that day. In the pps, if these two horses get into the same race, the speed figure is going to show that they are equal. But the velocity ratings, even though they're adjusted for track speed, will always make it look like the horse that ran the 1:09 was clearly faster than the horse than ran 1:12. Whether or not this is a good or bad thing is debateable. I tend to prefer to bet on horses that actually ran fast, as opposed to horses that may have run fast if the track was faster, but others disagree with that way of thinking.

If you are making a track variant and adjusting the times by the variant, why would the velocity ratings be different?

[pandy]

Originally Posted by cj

If you are making a track variant and adjusting the times by the variant, why would the velocity ratings be different?

If you are only using final time, they would be the same but velocity ratings are usually based on a combination of several points of call, and the horses that ran faster usually have the lowest, and best, velocity ratings...unless the times are similar. But say you're using the Sartin/Brohamer Average Pace, in a sprint, that is the first, second, and third fraction added and divided by three. Say a horse runs 6 furlongs and goes :22, :44, 1:09 and earns an 80 speed figure (very fast track --15 variant), and another horse runs :23, :46, 1:12 and earns an 80 (average track zero variant)....the 1:09 horse would clearly be ranked as the faster horse using Average Pace, even after applying the track variant. Normally if the variant is -15, you're going to split it for the three fractions, so that would change 22 to 23, 44 to 45, and the last quarter 25 to 26. The other horse is 23, :46, 26. So the horse than ran faster has a second quarter split that is a full second faster, thus, the faster Average Pace even after applying a full second variant to slow down each quarter.

I've run these type of ratings thousands of times and the horses that ran faster raw times get better ratings using the Brohamer/Sartin type of numbers.