Class Ratings

[xfile]

I'm working on several programs right now for personal use. The one thing I find hard to quantify by a figure is class. Probably because there are so many different opinions about the subject. Can anyone please share some ideas on how you would come up with your own class figures? Any input would be greatly appreciated.

[kenwoodallpromos]

I'm not a sofware data guy, but in class you may want to look at purse value and distance. IMO the lower the purse, the more important the difference in price of the purse and in distancel that is why they are cheaper horses! Not great stamina, not as much ability to improve too much.

[midnight]

My concept of class is basically how good a field the horse has beaten or run competitively against in the recent past. Quantifying that isn't easy, as one field of 10,000 claimers can be vey different from another. Once quantified, you tend to end up with chalky horses anyway.

[xfile]

There are so many variations when dealing with claiming price and purse levels. Well the old school way is as you know - what was the level of his last win. However if his last win was 6 races ago alot could have changed. It's a real hard thing to put into a number. If he won a 10k race 3 races ago and last time out ran 3rd for 15k.....it gets crazy

[ratpack]

You have to be real careful with using purse levels with all the slot money going into purses at certain tracks

You will have 8000 claimers at Mtn outclassing 16,000 horses at Santa Anita.

You may want to use something like intertrack Class-level rating from Cynthia Publishing or Dave Schwartz

[xfile]

Yes at Delaware Park too. I might have to look at claiming prices instead of purse level until it rises to allowance class then look for purse level. It's a tricky thing every way you look at. I think I saw a $4k claimer at Mountaineer one night with a purse $12k....i was in shock. That's a decent allowance purse at most small tracks. The thing is i am trying to find a simple way of coming up with this number. It doesn't even appear it can be done the hard way. I wonder what Bris and others use for their class ratings. Some kind of wild algorythym no doubt. :bang:

[ratpack]

Yes at Delaware Park too. I might have to look at claiming prices instead of purse level until it rises to allowance class then look for purse level. It's a tricky thing every way you look at. I think I saw a $4k claimer at Mountaineer one night with a purse $12k....i was in shock. That's a decent allowance purse at most small tracks. The thing is i am trying to find a simple way of coming up with this number. It doesn't even appear it can be done the hard way. I wonder what Bris and others use for their class ratings. Some kind of wild algorythym no doubt. :bang:

You run into the same problem. If anyone thinks a 16K Clm race at Mountaineer is the same class as a 16K Clm race at Sar, I have a nice 3 Bedroom condo in downtown Baghdad for sale

[xfile]

there are still condos left standing over there? lololol

[Niko]

I've been trying to do it myself to look for certain class moves. Can't figure it out. I always end up double checking the PP's because of the problems mentioned. You can try using track codes in relation to purse levels and pars but that's a pain in the a... for a non-programmer like me. Closest thing I've found that works but tough to do

[xfile]

I've been trying to do it myself to look for certain class moves. Can't figure it out. I always end up double checking the PP's because of the problems mentioned. You can try using track codes in relation to purse levels and pars but that's a pain in the a... for a non-programmer like me. Closest thing I've found that works but tough to do

There are so many variables it drives me nuts. Not only are there inflated purses at these tracks with slots/casinos but then there's the inflated purses for state bred allowance/stakes races. Especially in NY even though NY breds are getting much better these days alot of these horses who are running 3rd in a 50k allowance state bred race have a hard time beating 30k open claimers a few races later. I could use track codes for purse levels though. I know my programmer can have the program look at the track code then adjust the purse of the race it's looking at. I know I saw a chart on the average purse levels somewhere on the net. Do you know where this chart is?

[Niko]

I've seen the charts in the past but haven't looked for one recently. Right now I'm just eyeballing track codes and purses

[Light]

You also have to figure Nw2L,Nw2X etc. Classes within a class. I am also experimenting with class in my program. Alot of times,I find the sharp horse moving up in class,makes class look like Class = Form.

There was a $74 winner @ Mnr's 9th (purse $85K) last Sunday whom I played using my Class/Speed/Form angle.On Speed my computer program had him ranked 6th. On purse/Class I had him ranked 9th. When I started looking at my program's top selections,I could see the top ranked horses were not in form.So going thru a process of elimination,the longshot looked to be the main improving and in Form horse. Only question was class(9th ranked).Didn't matter. This race proved class=form. You can't play class blindly without considering form(i.E. does the horse seem fit and ready to improve or repeat a good race). I think it's impossible to program form into any software therefore I resort to judging class by purse for programing simplicity,knowing it's inherent flaws I just described.

[mikeb]

In my opinion there is only one kind of class figure that stands up.
"Current Class" available at tsn or brisnet.
Like the guy in the commercial say's "Try it you'll like it"

hwnn.

[sealord]

Quinn wrote the penultimate book on class in his book "Class Handicapping." It provides a scoring system relating class of race to adjusted final time to amount of effort and/or 'heart' shown. Its well proven that the majority of winners are in the top half of each race's class ranking. Its worth checking out, and could be applied to a computer program without too much effort I would think.

[RXB]

Penultimate means next to last, so I'm a little unsure about the phrase 'penultimate book on class.'

I was always skeptical about mechanical class quantifications, but Ken Massa at HTR has come up with one that hit 24% top-ranked winners with an ROI of .90 over a large sample, so it's possible to do it properly. Don't ask me how, though.

[sevenall]

Quinn wrote the penultimate book on class in his book "Class Handicapping." It provides a scoring system relating class of race to adjusted final time to amount of effort and/or 'heart' shown. Its well proven that the majority of winners are in the top half of each race's class ranking. Its worth checking out, and could be applied to a computer program without too much effort I would think.

It's been awhile since I read that book...but it seemed like there was quite a bit of subjectiveness to rating the horse's "heart" and what it's effort was against a particular class of horses. I remember being excited when I first started reading the book...thinking that it might be the answer to a solid "class rating"...but the process seemed extremly time consuming.

I'm thinking it might be difficult/impossible to program that into a software application??

[xfile]

I think it's impossible to program form into any software therefore I resort to judging class by purse for programing simplicity,knowing it's inherent flaws I just described.

Are you familiar with Tom Hambleton FORM POINTS?

[xfile]

This race proved class=form. You can't play class blindly without considering form(i.E. does the horse seem fit and ready to improve or repeat a good race). I think it's impossible to program form into any software therefore I resort to judging class by purse for programing simplicity,knowing it's inherent flaws I just described.

I certainly would not play class blindly. However I do want to come up with a class figure. I already have a great form figure. I believe I can have the tracks with inflated purses adjusted and state bred purses adjusted rather easily. Probably a 30% curve down for state breds and 40% to 50% for these casino/slots tracks. There is no perfect way to write a class number. I'm just trying to come up with one that is "good" enough.

[xfile]

It's been awhile since I read that book...but it seemed like there was quite a bit of subjectiveness to rating the horse's "heart" and what it's effort was against a particular class of horses. I remember being excited when I first started reading the book...thinking that it might be the answer to a solid "class rating"...but the process seemed extremly time consuming.

I'm thinking it might be difficult/impossible to program that into a software application??

Yes EXTREMELY time consuming. I have his "Handicapper's Condition Book". Is his other book similar to it do you know?

[Light]

Hambleton's Form points are a bit of a generalization. For example a myth I totally disagree with is a form point deduction for 1st route. I love that angle. It really depends who is stretching out. There are other rules I disagree with and others I actually incorporate that Hambleton mentions like giving points for class drops as long as they aren't negative drops.

Another aspect I think should be incorporated into software(re:class) is what constitutes the top class horse in a race. If a horse ran in a G3 and lost by 20 lengths and the others are allowance horses,the G3 horse is not the top class horse in the field cause he was not competitive.Therefore if a program gives points for running in higher class races,it's going to reward an inferior horse and detract from a lower class horse who may win the race.IMO,you can only assign a class rating to a horse from a race he ran competetively in.

[xfile]

Hambleton's Form points are a bit of a generalization. For example a myth I totally disagree with is a form point deduction for 1st route. I love that angle. It really depends who is stretching out. There are other rules I disagree with and others I actually incorporate that Hambleton mentions like giving points for class drops as long as they aren't negative drops.

Another aspect I think should be incorporated into software(re:class) is what constitutes the top class horse in a race. If a horse ran in a G3 and lost by 20 lengths and the others are allowance horses,the G3 horse is not the top class horse in the field cause he was not competitive.Therefore if a program gives points for running in higher class races,it's going to reward an inferior horse and detract from a lower class horse who may win the race.IMO,you can only assign a class rating to a horse from a race he ran competetively in.

I agree totally. If the horse ran 5th (or beat half the field) in a grade 3 and/or beaten only 5 lengths or less I would call that competitive. He would have a slight class edge on a field of allowance horses, if it was a condition allowance, who have no graded experience. If it was a NC Allowance his competitive grade 3 race might not mean a whole lot. However a grade 3 isn't really a good example. These days horses are running in 80k optional claimers one month and a grade 2 race the next month. Class has changed in this industry.

[JackS]

Considering that all horses at any Claiming price are running at a level that is slightly inflated. Horses at the 10k level are usually worth less than the asking price. If this weren't true, all horses would be claimed.
Perhaps the best guage would be consistancy within any claiming class.
Horses who win 1 race in 10 in average 10 horse fields, are winning an exact fair share of their races.
Horses at this level who win 2 races in 10 horse fields are wining twice their fair share and would be worth more than those who only win their fair average.
Placing a value on claimers would probably have more to do with an ability to beat the average.
Since the majority of claimers are not claimed, the winner of one race within a claiming class should probably be judged by his consistancy.
Consistance is only the first factor i.e- A 10k claimer who is 3 for 18 might look to be better than average but what if the average field size was 6.5 horse per race? Now he appears to be pretty much average.
I think if I were to attempt to quantify true class at any level, I'd start work along this line of thinking rather than some of the others mentioned.

[xfile]

Considering that all horses at any Claiming price are running at a level that is slightly inflated. Horses at the 10k level are usually worth less than the asking price. If this weren't true, all horses would be claimed.
Perhaps the best guage would be consistancy within any claiming class.
Horses who win 1 race in 10 in average 10 horse fields, are winning an exact fair share of their races.
Horses at this level who win 2 races in 10 horse fields are wining twice their fair share and would be worth more than those who only win their fair average.
Placing a value on claimers would probably have more to do with an ability to beat the average.
Since the majority of claimers are not claimed, the winner of one race within a claiming class should probably be judged by his consistancy.
Consistance is only the first factor i.e- A 10k claimer who is 3 for 18 might look to be better than average but what if the average field size was 6.5 horse per race? Now he appears to be pretty much average.
I think if I were to attempt to quantify true class at any level, I'd start work along this line of thinking rather than some of the others mentioned.

That could be programmed easily. Look at how many starters in the race a horse won. A horse winning a 10k claimer with 6 horses starting would get a lower class rating than one who beat 10k platers with 10 starters. Well it is food for thought.