Bris Variant

[Secretariat]

Does anyone know of a way to extrapolate the Bris variant from the Speed or Pace figures? I know keeping ones own figures are superior, but I don't have time anymore for that.

Any ideas?

[Big Bill]

Sec,

I can suggest what I have done to extract the track variant from the BRIS speed rating, but it may take you as much time as making your figs. I have converted BRIS ratings to Beyer ratings with a formula that was once posted on this forum. After the conversion to a Beyer rating I've extracted the variant in accordance with Beyer's instructions on how to do it in his book, Beyer on Speed, page 134.

If this interests you let me know and I'll try and supply more information.

Big Bill

[cj]

It could be done, but would take some work. First, you would have to figure out the value of time per second, or 1/5 of a second, or 1/100 of a second, or whatever, at each distance. You can do this by finding winners from the same race card that raced at the same distance. If Horse A wins in 1:10 and gets a 100 and Horse B wins in 1:11 and gets a 90, you know that 1/5 second is equal to two points. I'd look for several examples to eliminate possible split variants and smooth out rounded values.

Then, you need to figure out the value of a beaten length at each distance. I think this has been posted on this site a few times.

Once those are done, you only need to find out what final times at each distance BRIS considers equal. This would be the toughest part, but isn't that hard if you have a bunch of old PPs/results.

The last part would be to construct BRIS "speed charts" similar to what Beyer published in his book Picking Winners using the standard times for each distance and the value of time.

If all that is finished, assume a horse is given an 80 beaten 2 lengths at 6f in 1:12 2/5. First, just add the value of two lengths to the 80 to get the race figure. Now you just consult your new speed charts at 6f to determine the value of 1:12 2/5 seconds. If it was a 90, you now know the track was considered 10 fast.

If anyone could post any of the above info and we could start making it a group project.

[Light]

Sec


What's worse is Bris doesn't even have a variant listed for some poly tracks like DMR,OSA,GG. Not enough data yet. But they make the speed figures based on factors that include the variant. Those speed figs may be suspect,but like you,I dont have time to make accurate ones. When I put them through my homemade program,I believe they are all recieving the same automatic variant.

[Tom]

I am finding that one SR point equal 0.1 seconds, but there are enough days that this is not the case to make we wary. The beaten lenghts works out to be 10 / distance, but round off - so 6.5 is 6, 3.75 is 4....it comes out better more often than not.

TSN has posted a list of what the 100 equals at each distance....this might be close enough - you could try it and see if your daily variant makes sense using those times. I'll see if I can find the link. I have posted it previously.

There is another easy way to make a PDQ variant. I'll get into it after lunch.

[Tom]

Clader today, race 2, as an example.

[bobphilo]

TSN has posted a list of what the 100 equals at each distance....this might be close enough - you could try it and see if your daily variant makes sense using those times. .

Tom, I went to the TSN site and can't find that list anywhere.

Bob

[socantra]

TSN has posted a list of what the 100 equals at each distance....this might be close enough - you could try it and see if your daily variant makes sense using those times. I'll see if I can find the link. I have posted it previously.


That particular list has floated around for some years, purporting to be from TSN. The place I remember seeing it 1st was at Jerry Stokes' a1handicapping site. I have NEVER seen any statement by TSN that claimed ownership and could never get word from Jerry identifying the source. The point values are mostly 10/#furlongs.

[cj]

I'll try to put something together over the break. It really isn't very hard.

I currently do the exact same thing Bob does on the rare occasion I need it, I just convert to Beyer scale and do it that way.

[bobphilo]

I'll try to put something together over the break. It really isn't very hard.

I currently do the exact same thing Bob does on the rare occasion I need it, I just convert to Beyer scale and do it that way.

If the TSN scale is identical to the Brisnet scale, I had the methodology right under my nose all along. I found that that 3 Beyer points were equal to one of my standardized points (and 1 T-graph or Sheets points). I was also using a rather simplified scale where 2 Brisnet Speed points were equal to 1 of my std points and 1 Bris Class points, but am currently revising my scale in accordance with CJ’s more precise scale in which,
Beyer = Bris – 41.1 / .612, or
Bris = .612 Beyer = 41.1.


By the way, CJ, have you found a formula for converting Bris Class to Bris Speed other than the 1 to 2 ratio I’m using, or perhaps a conversion formula from Bris Class To Beyer?

I find that by converting Bris Class and Speed to standardized ratings and averaging the 2, I get a useful Class/Speed rating.



Bob

[Secretariat]

Sec,

I can suggest what I have done to extract the track variant from the BRIS speed rating, but it may take you as much time as making your figs. I have converted BRIS ratings to Beyer ratings with a formula that was once posted on this forum. After the conversion to a Beyer rating I've extracted the variant in accordance with Beyer's instructions on how to do it in his book, Beyer on Speed, page 134.

If this interests you let me know and I'll try and supply more information.

Big Bill

Big Bill. Thanks a good idea. I've my own formula for converting Bris to Beyers and programmed it. It hadn't occurred to me use Beyer's extraction formula for that. The only thing is it is twice removed which adds even more possible error into it. However, your approach seems quite valid.

Thanks much.

[kgonzales]

Interesting idea extracting the BRIS Speed variant via a Beyer equivalent. Does anyone know if BRIS calculates different variants for pace figs and speed figs? I have a feeling they might since the Speed figs seem a little more reliable than the pace figs.

Haven't had a chance to figure this out yet, but hoping someone can save me some time. Does anyone have a formula for figuring the unadjusted beyer (ala charts in his books) from plugging in the raw final time and race distance? I'd imagine there are separate ones for 1 and 2 turn races.

Thanks,
Kristian

[bobphilo]

Haven't had a chance to figure this out yet, but hoping someone can save me some time. Does anyone have a formula for figuring the unadjusted beyer (ala charts in his books) from plugging in the raw final time and race distance? I'd imagine there are separate ones for 1 and 2 turn races.

Thanks,
Kristian

In the back of "Beyer on Speed" there is a chart that gieves the unadjusted Beyers for each time for every distance. Their are seperate charts for one and two turn races.

Bob

[kgonzales]

Thanks Bob. I have the charts.

What I was looking for was a formula I could plug in excel that takes the final time and distance and spits out the beyer, so I could automate this variant extraction deal without manually looking it up in the chart. I just haven't had time to solve the relationship myself.
I suppose I could just make the charts into a a table of static lookup values as well.
-KG

[Tom]

I have a formula to make a Beyer from raw time, depending on the distance. Let me look around and I'll post it later.

[cj]

I have a formula to make a Beyer from raw time, depending on the distance. Let me look around and I'll post it later.

I'm sure you know this Tom, but there are many differences at the same distance depending on the track for Beyer. I have no idea if BRIS has the same differences, but I doubt it.

[Tom]

Yeah, I know :mad:

Not only that, but he treats one and two turn miles differently - I made up my own charts for Belmont for 8.5 and 9....just too many variables to make it easy, but I figure if I get it close enough.....maybe it will cancel out the beaten length errors! :D

This used to be a major pain for me to make my own figs......then YOU came along! ;)

[Tom]

So with all the normal caveats, here is what I have used in the past....the "x" is the final time, in this format - 1:35.4 = 95.4, 109.6 = 69.6.
One Turn Races
5.0 y = -16.94 x + 1078.7
5.5 y = -15.353x + 1083.9
6.0 y = -13.879x + 1077.6
6.5 y = -12.751x + 1079.6
7.0 y = -11/755x + 1078.7
8.0 y = -10.162x + 1078.3
8.5 y = -9.5054x + 1077.7
9.0 y = -8.9199x + 1076.2

Two Turn Races
8.0 y = -10.1440x + 1085.8
8.5 y = - 9.5054x + 1086.7
9.0 y = - 8.9199x + 1085.2
10.0 y = - 7.9747x + 1089.4

[kgonzales]

Thanks Tom!

I'll try those out and see how it goes.

-Kristian

[Tom]

Been working on this and came up with some new values for a point of BRIS speed.
At 6 furlongs, I get 1 point = .12 seconds and at 1 mile, 1 point = .17
I had been using 0.1 for all distances, but got to many really odd times.
Theses two values are not really that constant either - lots of variation, or BRIS splits variants almost every other day.

I need to feed some more days in to my db to do the other distances.

[socantra]

Been working on this and came up with some new values for a point of BRIS speed.
At 6 furlongs, I get 1 point = .12 seconds and at 1 mile, 1 point = .17
I had been using 0.1 for all distances, but got to many really odd times.
Theses two values are not really that constant either - lots of variation, or BRIS splits variants almost every other day.

I need to feed some more days in to my db to do the other distances.

Tom,

I don't have a database, but your numbers seem close to what I've been using. I always heard that the old BRIS BBS gave a point value per fifth second as 10 / # furlongs.

10/6=1.6667

Multiply the fifth of a second valus by five to get points per second.

1.6667*5=8.3335

And take the reciprocal of that to get seconds per point.

1/8.3335=.12

Nice linear results and always seemed to fit fairly well. The BRIS/DRF tie in to a fifth second length always seemed appropriate.

5f = .10
5.5f= .11
6f = .12
6.5f= .13f
7f = .14
7.5f= .15
8f = .16
8.3f=.1667
8.5f=.17
9fr =.18
9.5f=.19
10f =.20

[Tom]

That is looking very close to what I am getting tonight...thanks!

[socantra]

That is looking very close to what I am getting tonight...thanks!

They've always seemed reasonable for general use. The problem I've always had is establishing the base, or par to use when extracting variants.

CJ uses the "near" universal Beyer pars and does a great deal of his track to track adjustment thu the variants themselves. It makes extraction much easier. I believe that BRIS may well do their track to track adjustments thru the individual pars.

That's what Trackmaster does, but they make it much easier by supplying the par time for each individual race and the variant itself. BRIS only gives you the DRF variant and the par for today's race in terms of their own speed figures rather than actual times.

I'm sure its doable, but the hurdles have put me right to sleep every time I started playing with it.