Using two handicapping methods

[pandy]

I did two videos about class on my YouTube channel this week.

[Tom]

Thanks for the heads up!
Lots of interesting videos there.

[Tom]

Originally Posted by classhandicapper

It's called "Class of the Field" by James Quinn.

A really good method to combine speed and class with effort.

Each class has a rating, ie, Alw n1x might be generically a 60.
Now, you determine was the speed figute average, fast, or slow for the class?
Races are rated e ways, D A S.
Say the Beyer par is 92 and the race went n 97. That is fast..
Now you look at how competitive the race was, rating it 1-5, for non competitive to highly competitie. A race is a 3 unless there is good evidence
that it either softer or tougher. Say the horse ran on a lone, easy lead the whole trip and was never threatened, it might be called noncompetitive, a 1.
So the race will rate the lower than speed tells you, so you now have a SOR rating intelligently derived. There are specific guidelines for each level of speed and competiveness to guide you.

I can see this being helpful at NYRA for MSW, M590, Alw n1x, and Alw50000 levels.

[MJC922]

Originally Posted by classhandicapper

Speed figures measure how fast a horse ran given a specific set of conditions (pace, competitive nature of the race, quality of the other horses, how the track surface was playing, trip etc..)

Class is all of a horse's underlying tangible and less tangible abilities expressed as one

1. How fast he can run at top speed
2. How quickly he can accelerate
3. How much stamina he has
4. How long he can sustain at or close to top speed before cracking
5. How well he handles the pressure of other horses battling with him
6. How adaptive he is to various pace scenarios
7. How well he handles deep or hard surfaces
8. How willing he is to fight for the lead when tired
9. How fast he is out of the gate
10. etc

As you move up and down in class, you face progressively weaker or stronger opposition. So the demands of the race "tend" (not always) to get correspondingly weaker or stronger. Those conditions are part of what impacts the times.

What the class handicapper is trying to do is get past just the final times to the underlying ability of the horses. They will be closely correlated, which is why speed figures work so well. But class is more comprehensive, albeit more difficult to measure, especially when it comes to lightly raced horses that have not been tested yet. You don't know what a horse has in reserve until it's required. Some will wilt chasing horses quicker out of the gate and equally fast or faster and some will reveal they had more pure speed and ability to carry the battle further than they showed to that point.

(and of course, class/ability is not static. it changes)

IMO there are NO really good books on class. There are chapters in books that are nonsensical, chapters that highlight techniques for measuring it that are not good, and occasional solid insights, but no one has put it all together like Andy Beyer and others did for measuring final time.

I don't know if there ever will be a class rating book like Picking Winners was for rating speed. For one thing to make some sort of viable pen and paper class rating which measures the effort of a horse you would need good comprehensive class pars which almost nobody has, especially given all of the various permutations that exist. I could probably assemble such pars but they would take up an entire book on their own. The other thing is if you take the computer processing out of class ratings then you're likely going to have less accuracy by at least a few percentage points. Today at 30% wins on class we have probably one of the finest horse vs horse class ratings available in the game today but I'll admit there are still significant issues sometimes with rating lightly raced horses, well-beaten horses etc. With my technique one of the things I find the computer needs to be able to do in order to rate a winning horse's performance properly is it needs to rate every other horse in the race in a way that makes the most sense given the available information. That means well-beaten horses are forced to contribute to the assessment and if the computer were to rate several well-beaten horses let's say 'properly' like a speed figure might do then it would no longer be able to rate the winner properly so limits are sometimes enforced by the computer in order to maintain the needed accuracy of the winner's 'rating'. With speed figures, for the horses who get crushed in a race you can almost just ignore them in the calculations because (1) you have the clock to fall back on and (2) the end-users don't expect a horse that runs last in a race to have a speed figure which measured much of anything accurately. With class ratings though (at least with the process that I use) the computer can't allow some numbers to be that bad because if it did then it would lose accuracy for the better finishers in the race. Someone out there may have a better technique to handle these issues. I wish I had the time to look into it. Knowing first hand the complexity involved my hat is off to anyone who can write a book which does the class rating game justice.

[classhandicapper]

I know exactly what you are saying about well beaten horses and comparative handicapping.

I think the "depth" of a field is part of what impacts a horses finishing position and beaten lengths.

I like to use this theoretical example.

Imagine a race with 10 Spectacular Bids in it. One of them is going to finish 1st and one last. I can almost guarantee you they won't finish 10 across the track. The margins will in part be related to ground loss, pace, and bias, but there will also be more subtle positional elements to it like who just happened to get out of the gate best, what the jockey decided to do etc...

Especially in deeper fields, if try to you "systematically" rate the worst finishers relative to the winner you will tend to underrate the losers and if you rate the winner relative to the worst losers you will overrate the winners. If you try to adjust for trip, you'll do better, but you will still have that issue.

My preference is to look for common trips within the race and rate those horses relative to just each other.

For example, let's say I rated 3 horses as Grade 3 types coming into a race and they finished 7th, 9th and 10th with similar trips in a very deep Grade 2 race. Systematically and even speed figure wise they will probably rate well below Grade 3 because of their poor finish.

I won't change my evaluation.

I'll continue to rate them as Grade 3 types because they all ran similarly with similar trips. I don't care what the figures, finishing position, and margins say.

I also won't suddenly think the winner is a superstar because some Grade 3 horses were well beaten.

That style of class handicapping where you are incorporating class, trips, and a kind of key race analysis is virtually impossible to code for. It requires a LOT of experience classing horses with good trip information to get right. Handicappers tend to overvalue some trips and underrate others in those comparisons. But when you have a strong opinion, you will sometimes come up with a horse that's way better than any commonly used figures will show. You can get a very good price on a very likely winner because the class is hidden.

[ranchwest]

A few observations:

1) It is almost impossible to separate speed and pace. Most fast times have fast lead fractions throughout.

2) Horses that start and finish at the rear should be no consideration as to class of the other participants because the rear horses typically didn't affect the final time of the race at all.

3) Predictably slow pace races cause some of the horses to run outside their element, on the lead, and will often have slow final times, often slower than prior speed ratings might suggest.

4) Speed and class correlate, but are not consistently joined at the hip.

[MJC922]

Originally Posted by classhandicapper

I know exactly what you are saying about well beaten horses and comparative handicapping.

I think the "depth" of a field is part of what impacts a horses finishing position and beaten lengths.

I like to use this theoretical example.

Imagine a race with 10 Spectacular Bids in it. One of them is going to finish 1st and one last. I can almost guarantee you they won't finish 10 across the track. The margins will in part be related to ground loss, pace, and bias, but there will also be more subtle positional elements to it like who just happened to get out of the gate best, what the jockey decided to do etc...

Especially in deeper fields, if try to you "systematically" rate the worst finishers relative to the winner you will tend to underrate the losers and if you rate the winner relative to the worst losers you will overrate the winners. If you try to adjust for trip, you'll do better, but you will still have that issue.

My preference is to look for common trips within the race and rate those horses relative to just each other.

For example, let's say I rated 3 horses as Grade 3 types coming into a race and they finished 7th, 9th and 10th with similar trips in a very deep Grade 2 race. Systematically and even speed figure wise they will probably rate well below Grade 3 because of their poor finish.

I won't change my evaluation.

I'll continue to rate them as Grade 3 types because they all ran similarly with similar trips. I don't care what the figures, finishing position, and margins say.

I also won't suddenly think the winner is a superstar because some Grade 3 horses were well beaten.

That style of class handicapping where you are incorporating class, trips, and a kind of key race analysis is virtually impossible to code for. It requires a LOT of experience classing horses with good trip information to get right. Handicappers tend to overvalue some trips and underrate others in those comparisons. But when you have a strong opinion, you will sometimes come up with a horse that's way better than any commonly used figures will show. You can get a very good price on a very likely winner because the class is hidden.

Yes when I dig deeper into what the computer does end up 'doing' in these situations it's very similar to what you mention here. Like for example some of Flightline's races when I looked back at horses at the rear there was very little difference in their performance numbers as far what the computer assigned to them even though the margins between them were sometimes substantial. It's kind of the computer's way of disregarding their impact on the estimate. I wasn't entirely at ease with it being a speed figure guy a long time ago but I do understand the why behind it. It literally HAS to sometimes bucket horses like this together in order to maintain the accuracy of the other performances in the race. If it doesn't then the process skews and races can't be rated properly. Folks may take issue with the similar numbers for those types of efforts and I totally understand where they're coming from but it's a whole different process vs how speed figures are calculated. When people understand what's going on they can evaluate accordingly. In short there are situations to be aware of that do result in a degree of reduced accuracy for those types but as long as there's an awareness it's not a deal breaker, at least for me personally.

[classhandicapper]

Originally Posted by MJC922

Yes when I dig deeper into what the computer does end up 'doing' in these situations it's very similar to what you mention here. Like for example some of Flightline's races when I looked back at horses at the rear there was very little difference in their performance numbers as far what the computer assigned to them even though the margins between them were sometimes substantial. It's kind of the computer's way of disregarding their impact on the estimate. I wasn't entirely at ease with it being a speed figure guy a long time ago but I do understand the why behind it. It literally HAS to sometimes bucket horses like this together in order to maintain the accuracy of the other performances in the race. If it doesn't then the process skews and races can't be rated properly. Folks may take issue with the similar numbers for those types of efforts and I totally understand where they're coming from but it's a whole different process vs how speed figures are calculated. When people understand what's going on they can evaluate accordingly. In short there are situations to be aware of that do result in a degree of reduced accuracy for those types but as long as there's an awareness it's not a deal breaker, at least for me personally.

It’s a bit tough to wrap your head around if you have always been a time oriented handicapper.

As I started appreciating what was really going on out there and further appreciated all the technical problems with time based figures, I got progressively more skeptical of time. But time based figures remain useful to me because there are occasions I don’t have a good qualitative line on the horses and time at least gets me into the ballpark.