Since Speed is the Long Pole in the Tent

[DJofSD]

I think the main intent of the thread is to find the one ring which rules them all.

[classhandicapper]

IMO good class figures outperform good speed figures mildly in win percentage and ROI. Most likely the difference in ROI is that good speed figures are easily purchased and good class figure require a lot of initial research and work that very few people are willing to do.

[DJofSD]

Originally Posted by classhandicapper

IMO good class figures outperform good speed figures mildly in win percentage and ROI. Most likely the difference in ROI is that good speed figures are easily purchased and good class figure require a lot of initial research and work that very few people are willing to do.

Are you saying class figures perform better in both win percentage AND ROI?

It seems there is always a trade off between those two metrics. You can either have a higher win percentage but with a lower ROI, or, a lower win percentage but at a higher ROI.

[classhandicapper]

Originally Posted by DJofSD

Are you saying class figures perform better in both win percentage AND ROI?

It seems there is always a trade off between those two metrics. You can either have a higher win percentage but with a lower ROI, or, a lower win percentage but at a higher ROI.

Yes.

I'm testing my own systematically generated class figures, but only where I have good classing information. So you might say the test is biased towards what I already know I am pretty good at. I doubt the figures would do as well at tracks or at class levels I am less familiar with or where classing horses is more difficult.

They are outperforming speed figures in BOTH win% and ROI on both dirt and turf so far, but the samples are not large yet.

I'm not using any subjective judgment in creating the figures like I do in my gambling. It's all pre programmed and generated by rules. My goal was simply to answer some questions about class/speed and their relationship once and for all.

[cj]

Originally Posted by classhandicapper

IMO good class figures outperform good speed figures mildly in win percentage and ROI. Most likely the difference in ROI is that good speed figures are easily purchased and good class figure require a lot of initial research and work that very few people are willing to do.

If they are kept private, sure. But if they are shared like speed figures are these days, the same thing will happen.

Do you have an example of publicly available "good class figures"?

[classhandicapper]

Originally Posted by cj

If they are kept private, sure. But if they are shared like speed figures are these days, the same thing will happen.

Do you have an example of publicly available "good class figures"?

I'm not sure there are any "good" class figures available to the public. I've read some of the documentation on the Bris class figures and I like what I've read, but I never analyzed or tested them. So I'm not qualified. They aren't anything like what I am doing now.

[cj]

Originally Posted by classhandicapper

I'm not sure there are any "good" class figures available to the public. I've read some of the documentation on the Bris class figures and I like what I've read, but I never analyzed or tested them. So I'm not qualified. They aren't anything like what I am doing now.

That is my point though on the ROI side. You can't really compare something that everyone has to something very few have and say it is better...it is just not in public use so it isn't reflected on the tote.

The same goes with the win percentage. You are picking spots. There are plenty of situations where speed trumps class if I want to cherry pick.

[classhandicapper]

Originally Posted by cj

That is my point though on the ROI side. You can't really compare something that everyone has to something very few have and say it is better...it is just not in public use so it isn't reflected on the tote.

The same goes with the win percentage. You are picking spots. There are plenty of situations where speed trumps class if I want to cherry pick.

No doubt, speed figures are at a disadvantage in ROI. Whatever everyone is doing is at a disadvantage.

Here's what I am doing in the test.

I'm not looking at surface changes, layoffs, off tracks, distance changes, or anything else. I look at last race and that's it.

One reason I am being selective is that to duplicate what I am doing right now on a national basis would take years. I don't have the time to look at all tracks. I consider that irrelevant.

Another reason is that I don't see the point of trying to class maiden races with loads of FTS or turf races with loads of first time turfers. So I exclude some race types. Speed figures would have similar issues in those races.

I'm only comparing speed figures and class figures in the exact same sample of races. They are races where speed figures do very well also. So there should be no inherent advantage to class over speed in my sample that I can think of.

Here are the results so far.

If the top class figure goes off the favorite, the loss is a little better than the track take.

If the top class figure does not go off the favorite, there are huge ROI profits.

Dirt is better than turf, but turf is also profitable (barely).

Overall, the figures are mildly profitable as of last week's results.

I was hoping to find greater value at a slightly lower win percentage than speed figures. I assumed that any chance of competing with speed figures would require a more detailed subjective analysis. But I'm generating both greater value and a higher win percentage on automated class ratings. That makes me enthusiastic.

I coded the system rules at the start and froze them there. I just keep adding to the sample. I haven't even gotten to the point where I start tweaking the rules to try to improve the results for turf vs. dirt, sprint vs, route etc..

Maybe as the sample gets bigger the results will deteriorate a bit, but it's hard to not be encouraged.

[Capper Al]

Originally Posted by pandy

He is talking about predictability. That was the original idea for the thread, using the top predictor as a starting point and improving on it.

You got it.

[Capper Al]

Originally Posted by Tom

Pace is the key factor - every race is decided by a pace match up.
When a class horse wins, it wind by setting or overcoming a pace.
When a form horse wins, it does by setting or overcoming the pace.
When a speed horse wins, it does by setting or overcoming the pace.

You can see it in the charts, but before the race, it is not always easy.
Especially if you play where jockeys like to grab.

For me, pace has made a better elimination than a predictor. Primary factors, again for me, work better as predictors. Secondary factors work best in eliminations. The horse set the best pace because of his class vs the horse isn't a contender because he's an E type and can't get within 5 lengths at the quarter pole.

[Capper Al]

Originally Posted by DJofSD

I think the main intent of the thread is to find the one ring which rules them all.

Not so, but I can see how one could think this. The starting point could be seen as the most valuable ring.

[Capper Al]

Originally Posted by classhandicapper

I'm not sure there are any "good" class figures available to the public. I've read some of the documentation on the Bris class figures and I like what I've read, but I never analyzed or tested them. So I'm not qualified. They aren't anything like what I am doing now.

Using BRIS RR and CR has worked well for me. Yet speed is still my most predictive factor. The class figs just pay better.

[appistappis]

years ago speed was somebodies main pole and pace was probably someone else's main pole but in today's game you better have a multi faceted tent.

[Capper Al]

Originally Posted by appistappis

years ago speed was somebodies main pole and pace was probably someone else's main pole but in today's game you better have a multi faceted tent.

The jury is still out for me on this one. Multi faceted. or what I call comprehensive handicapping using more than a couple of factors, is more predictable. Yet, my best ROI streak did come from using just Giles' pace. I'm waiting to finish my rewriting my app before testing this again.

[classhandicapper]

Originally Posted by Capper Al

The jury is still out for me on this one. Multi faceted. or what I call comprehensive handicapping using more than a couple of factors, is more predictable. Yet, my best ROI streak did come from using just Giles' pace. I'm waiting to finish my rewriting my app before testing this again.

My experience is that a multi faceted approach works best for predictability.

If you are a strict line maker I think you want to take a multi-faceted approach.

If you are more selection oriented, meaning you look for horses that are underrated or that possess certain characteristics that tend to be undervalued by the public, then adding more factors has a tendency to improve the win%, but lower the ROI.

That's even what I am finding in my own tests on class. When I added speed figures to the formula just to see what would happen, I was able to find a weighting between class and speed that improved the win% over either of the factors as a stand alone. But the ROI plummeted because speed figures are so built into the odds.