Improvement On Beyer Method (Universal Pars)

[RaceIsClosed]

Perhaps CJ or another speed handicapper could comment here.

Back in the day before the Beyers were in the Form, I used to make my own numbers. To do this, however, I needed to establish the route-sprint relationship for each track. This was not usually difficult, as most tracks card long meets, but at places like Keeneland, Saratoga, and Timonium, it was almost impossible because of the relatively small sample of races. To compensate for this, I tweaked Beyer's formula a little bit.

To get a route-sprint relatrionship (i.e., how many Beyer points to add for two-turn races), Beyer suggests averaging out the difference in average sprint and route figures for each class: i.e., MSW, NW1, and all levels of claimer. This, however, meant keeping several separate categories for the claiming races, when according to Beyer's own Universal Claiming Pars this isn't necessary.

Beyer claims a fixed relationship between claiming prices that applies to all (or most) tracks, whereby a $5,000 claimer is (or was) 28 points weaker, on average, than a $50,000 claimer, and so forth. For the route-sprint relationship, this was very useful information, because it allowed me to lump all claiming races into one group by "converting" each claiming race into a $50,000 claimer. Instead of nine different averages with a few races each to compare, I would get one average for sprint and route claimers, with dozens of races, and a much more accurate route-sprint relationship.

Has anyone else ever done or contemplated doing something like this? I no longer make figures but this was a very useful technique when I did.

[Whirlaway]

I simply average all the non-maiden races for each distance. Seems to work fine. If you're a little off, it will soon become evident (e.g. all the 1 mile routes may seem to come up a little high, so you change your parallel time chart to reflect that.) As always with figures, there's a little bit of art that goes with the science.

[RaceIsClosed]

I simply average all the non-maiden races for each distance. Seems to work fine. If you're a little off, it will soon become evident (e.g. all the 1 mile routes may seem to come up a little high, so you change your parallel time chart to reflect that.) As always with figures, there's a little bit of art that goes with the science.

I could see where that might work if they balance the sprints and routes.

In this case, however, Beyer himself effectively said that a $50,000 claiming time was usually 28 points better than a $5,000 claimer, that this was a relationship he had proven.

Given that, Beyer should have also recommended converting the claiming races as above, since we "know" that a $5,000 claimer running a 61 is the same as a $50,000 claimer running an 89, or a $12,500 claimer running say a 67, etc.

[osophy_junkie]

I do not worry about man made class when generating figures based on time. They are an arbitrary measurment which the horses know nothing about. When they run they are not running against every other horse in that class, they are running against a small subset of all horses at the track. When I generate track to track I am looking for the overall speed characteristics of the track and classiness of horses that runs there.

I generate a class rating for each race that is based on how the horses in the field have run in the past, and how their competitors have run since. I take the median average of these to generate a track par. I do track to track conversion by distance and not class. For the most part I do not do distance to distance conversions. If a horse hasn't raced at the distance or one very close to it, that fact is noted.

[RaceIsClosed]

I do not worry about man made class when generating figures based on time. They are an arbitrary measurment which the horses know nothing about. When they run they are not running against every other horse in that class, they are running against a small subset of all horses at the track. When I generate track to track I am looking for the overall speed characteristics of the track and classiness of horses that runs there.

I generate a class rating for each race that is based on how the horses in the field have run in the past, and how their competitors have run since. I take the median average of these to generate a track par. I do track to track conversion by distance and not class. For the most part I do not do distance to distance conversions. If a horse hasn't raced at the distance or one very close to it, that fact is noted.

This was for back in the day, when one had no basis for getting the ball rolling. You still have to know how much time the extra turn is worth at each track, and to do that, you need a set of sprint and route pars.

My adjustment showed how to generate accurate two-turn adjustments from a very limited sample of races, using Beyer's own published theory on Universal Claiming Pars.

[Whirlaway]

You can use the Beyers in the DRF to make your projections. I don't see the point of trying to keep a comprehensive database of speed figures when there are so many good ones out there for sale. The time isn't worth the effort.

I make figures for my home circuit (SoCal) because (a) I want pace figures; and (b) I want to identify the races that were particularly fast or slow for the level. You'll find that the Beyer boys are often too conservative, particularly when it comes to claiming horses, e.g. they assign a 12.5K Claimer an 85 figure when he clearly earned a 93 because the 93 just doesn't "feel" right to them. These are the kind of horses who often rip off three straight wins or three out of four as they climb the class ladder.

Keeping figures on the grass keeps you in touch with how big the pace variations are in those races (much bigger than in dirt races) and how so many of the Beyer figures on the turf are really just made up "class" figures masquerading as speed figures.

[GlenninOhio]

You can use the Beyers in the DRF to make your projections. I don't see the point of trying to keep a comprehensive database of speed figures when there are so many good ones out there for sale. The time isn't worth the effort.

I make figures for my home circuit (SoCal) because (a) I want pace figures; and (b) I want to identify the races that were particularly fast or slow for the level. You'll find that the Beyer boys are often too conservative, particularly when it comes to claiming horses, e.g. they assign a 12.5K Claimer an 85 figure when he clearly earned a 93 because the 93 just doesn't "feel" right to them. These are the kind of horses who often rip off three straight wins or three out of four as they climb the class ladder.

Keeping figures on the grass keeps you in touch with how big the pace variations are in those races (much bigger than in dirt races) and how so many of the Beyer figures on the turf are really just made up "class" figures masquerading as speed figures.

And the converse (or opposite, or whatever it should be called) might also be true - if the winner of a graded stakes runs what should be an 85 maybe the number is goosed into the 90's because that "feels" better.

While I agree that the hypothetical "93 masquerading as an 85" horse you cite could be a great bet as he climbs up the class ladder, I also have a question for you and the group.

How is "class" defined in the context of speed figures? We've all seen examples of horses who can run a blowout 93 Beyer against $30,000 claimers and who time and again can't do an 80 against $50,000 company.

Is this simply a pace issue, where the class jump means the horse is going to have to expend more energy early than it is capable of sustaining? Or is there something less tangible going on in terms of horses deferring to others higher up in the pecking order?

[so.cal.fan]

AH HA! Glen!

"How is "class" defined in the context of speed figures? We've all seen examples of horses who can run a blowout 93 Beyer against $30,000 claimers and who time and again can't do an 80 against $50,000 company".


Because a good (better class) horse will always keep up a pace to beat a cheaper one...always.

Example, years ago in Calif. there was a long distance horse name of Hill Rise.
He once ran 6 1/2 on the turf.....everyone dismissed his chances, because he was in with a long distance horse.....he got there in time to win the race.
Can't prove it, because they are not going to do it, but let's take Lava Man....winner of the recently run Santa Anita Handicap.......put him in a race with Sprinters.....stakes race.........he'll win the race, because he will be a shade better class than the sprinters.
My husband once galloped a horse called Colonel Mac, he was entered in the Bing Crosby Handicap, at Del Mar,I believe in 1956....Col. Mac was a distance horse......he was entered against Moolah Bux, the fastest sprinter in the country.......John Longden used to ride Col. Mac....he took off him to ride Moolah Bux.....he said NO WAY Col Mac would ever beat Moolah Bux sprinting. Ralph Neves got the ride on Col. Mac. Col. Mac won the Bing Crosby Handicap that day. Till the day he died, Ralph Neves was a strong believer in CLASS. He used to tell his ex wife,
"Time is for people in jail".

[cj]

AH HA! Glen!

Because a good (better class) horse will always keep up a pace to beat a cheaper one...always.



SCF,

I would change the above to say:

Because a faster horse will always keep up a pace to beat a slower one...always.

If it was truly class, wouldn't older horses that slow down still be able to beat younger ones? Old horses don't lose "class", but they sure drop down the class ladder. That is because they get slower.

[so.cal.fan]

CJ, of course you are correct, I forgot to say, the horses in question must be in top form and top condition, that is obvious. I still hold the opinion, provided this is the case.
A former class horse who has injuries which have affected his stride, will obviously not be able to beat inferior class horses, although I've seen them do it often enough in cheap claiming races.

[Valuist]

I would say always make separate variables for routes and sprints. Needless to say, I'm not a big believer that there is a fixed relationship between a time at 6.5 furlongs and one at 1 1/16 miles. Making one variable for all makes sloppier speed figures for all.

[osophy_junkie]

...You still have to know how much time the extra turn is worth at each track, and to do that, you need a set of sprint and route pars.

My adjustment showed how to generate accurate two-turn adjustments from a very limited sample of races, using Beyer's own published theory on Universal Claiming Pars.

My pars are generated for the fractional calls. Breaking them up by distance gives me a much more relivent data set to determine track section differences. How does breaking them up by a human made class do this?

[Murph]

My pars are generated for the fractional calls. Breaking them up by distance gives me a much more relivent data set to determine track section differences. How does breaking them up by a human made class do this?

There have been studies conducted that show an inverse relationship between pace and speed figures. As pace call times become faster at the same class level the final time figures fall. The same is true in reverse, when the pace is slower the final figure will increase.

Consider the demands asked of a runner moving from his maiden win into non-winners of one race condition. It is certain that to win the race the maiden winner will have to contend with a pace time that is much faster than was faced in the past. The horse has surely demonstrated his talent if he is able to contend with the faster pace of todays race.

If this runner wins his condition the first or second try, he may very well contend in his next condition. Beware allowance runners who have struggled to win their current condition level. The winning race after several tries may have come against the weakest runners at that level. The final figure may be inflated due to a slower pace demand.

Seperate allowance runners with the pace ratings, key the runners who look best to contend or improve on a faster pace time. Like that tuff closer who may finally get a :44 pace duel to run at down the stretch.


Man, I LOVE handicapping.
Murph

[GlenninOhio]

SCF,

I would change the above to say:

Because a faster horse will always keep up a pace to beat a slower one...always.

If it was truly class, wouldn't older horses that slow down still be able to beat younger ones? Old horses don't lose "class", but they sure drop down the class ladder. That is because they get slower.

cj -

I ask the following potentially dumb question looking to be educated, as I'm pretty new to the pace/speed game.

Are you therefore saying that the definition of a "fast" horse is one that can run a fast pace on its way to a fast final time, and the "fastest of the fast" can handle just about any pace demand and still win and/or put in a very strong final time?

And might these varying abilities to handle pace demands define, even in the claiming ranks, what many refer to as "class"? For example, the $10,000 horse not being able to beat the $25,000 horse even after winning at $10,000 with a final time that should be more than sufficient to move up and win at $25,000?

And does this distinction underly what many refer to as "cheap speed"?

Thanks.

Glenn

[cj]

I am saying, at least I think :), that class and speed are the same thing. However, you can't measure only final time as speed as many different factors can and do effect final time.

Even top class horses can be effected by fast pace. I'll give a couple examples:

The Sunshine Millions, won by Lava Man, was given a pretty slow final time figure by nearly everyone. The pace of the race was insanely fast however, fast enough to effect all the horses in the race. I don't know of any figures that gave Lava Man a higher figure than High Limit earned winning the Strub. However, when they matched up in the Big Cap, Lava Man looked like a standout to me.

The Derby, won by Giacomo of course, was also insanely fast. In this case, those anywhere even close to the pace, including Afleet Alex and Flower Alley, were adversely affected. Giacomo won because his jockey, whether through skill or luck, kept him where he could run his absolute best final time, while the others all ran well below their level due to being spent early. A horse doesn't have to be on or near the lead for this to happen. This race was about as extreme pace wise as you will ever see.

Class is speed, when but measuring it is extemely difficult. Every horse has limits on how fast and how far they can run, and the complex interaction of all the horses in a race create wide variances on what time a horse will cross the race in. Throw in that the horse may be in better or worse physical condition every time the horse races, and that is what makes this game such a challenge.

[Tom]

My pars are generated for the fractional calls. Breaking them up by distance gives me a much more relivent data set to determine track section differences. How does breaking them up by a human made class do this?

Say your par for 6 furlongs is 46 flat. Doe you call it fast 5 (or10) if a stakes horse runs 45 and slow 5 or(10) if a maiden runs in 47?
If you have a par day and run only three maiden races at 6, you might be calling the track slow when it is not.
When I make pace pars, they are technically man made, but based on long term statistics - this is what I statistically expect a defined sub set of horse to run (par).

If you average 6 furlong pace class at Finger Lakes and Belmont, you will not have a track to track adjustment if you use them raw.