SPEED FIGURE COMPARISONS

[Dave Schwartz]

Originally Posted by bobphilo

Actually, studies have shown that horses win at about the rate their odds would predict so that playing them at random would pretty much would put you around the takeout rate. If one can improve on the win percentage of favorites or come up with a few longer priced horses than favorites, using figures or some other method or combination thereof, one can beat the takeout rate and even show a profit. However it can't be done using one simplistic factor such as top figure.

This is almost correct.

There is a skewing towards the lower odds horses. Thus, a horse at (say) 2/1 will win more races relative to the percent of pool wagered on it, while a horse at (say) 20/1 will win far less.

[bobphilo]

Originally Posted by Dave Schwartz

This is almost correct.

There is a skewing towards the lower odds horses. Thus, a horse at (say) 2/1 will win more races relative to the percent of pool wagered on it, while a horse at (say) 20/1 will win far less.

Right, that is also true. That is likely due to people shying away from low priced favorites making them underbet, while being attracted to high odds longshots thinking they represent value.

[ubercapper]

Originally Posted by formula_2002

I'm not measuring speed figures
I'm looking for anyone that provides speed figures to tell me their $1 return. set any criteria you wish.

For instance
From HTR Monthly Report April 2001
Quirin speed points:
dirt routes (no wet tracks or maidens tested)
8 Quirin points, … I need to read the entire article. a few caveats to consider.

I'll get back to you.

Please read post #9 in this thread and check out the link to this topic from last year.



The post from Ron Tiller most of the way down the page explains what happened and both his and my posts explained why it is unlikely to happen again.



My recollection was the preliminary results of the study were using ANY figure and the most common criteria (best last, best of last 3 or 5, average last 3 or 5) will yield the same results on a large sample set. The Bayle Study from many years ago showed using any figure from the four of the figure makers at the time (DRF, Equibase/Trackmaster, Bris, HDW) and using any of those criteria yielded a loss and the amount of the loss was insignificantly different between the figures.

[Tom]

But this thread also sheds some light on speed figures themselves.
That is, how do you use them?

Seems to bee three days without a single answer, and that is real life. Not everyone uses them the same. Rightfully so.
It is not just the fig (but best to use the best) but who is using it.
It is not as simple as who has the best last out numbers, but many claim that is exactly what some people do.

Just now on ATR, Andy Beyer described the last race before the Classic that Accelerator and West Coast came out of was "slow and lousy," One horse won and the other was last or so.

On on paper, using Beyers, it was two horse race when you consider proven best numbers (WC and ACC). Do you say Accelerate was a 100 or a 115 going into the classic? (Of course, I picked the two backwards! ).

Speed figures (and pace figures) are a tool. When you use any tool, you need to have a process. You don't just buy a power saw and start turning fine furniture, even though people CAN.

Speed figs are a very good tool, good enough that many people make money using them in less than the best ways.

It's just not as simple as some people make it out to be.

[bobphilo]

Originally Posted by Dave Schwartz

This is almost correct.

There is a skewing towards the lower odds horses. Thus, a horse at (say) 2/1 will win more races relative to the percent of pool wagered on it, while a horse at (say) 20/1 will win far less.

I wish I still had the study at hand but I seem to recall that the skewing towards low priced favorites does not occur until the horses go below even money to odds on. That seems to be the point that drives many bettors away though it is the group that actually outperform their odds and provide true value.

My memory may not be perfect but I would greatly appreciate finding that study since I've lost the book that contained it.

[classhandicapper]

Originally Posted by formula_2002

How about working it backwards.
"Here is the best value under x y z conditions"

I look at:

1. win%, average finish position, and ROI of the top choice.

2. win% and ROI of top choice when it was favored

3. win% and ROI of favorite when it wasn't the top choice.

4. top choice when it was 5-2 and above.

5. Top choice when it was even money or less

6. I pay close attention to outlier win payoffs that could have distorted the ROI.

I usually don't get too complicated in the testing. I look at last race with no other conditions. I consider the rest of it (which races to use, how many races to use, which races to excuse etc..) more of a handicapping question than a figure question. I just want to see the general accuracy and value in the figures I am using and go from there.

[Tom]

Andy Beyer on ATR first hour today discussing the fig for the Turf Sprint.

[classhandicapper]

Originally Posted by Tom

Andy Beyer on ATR first hour today discussing the fig for the Turf Sprint.

I never made turf figures, but when the track is yielding some horses are not going to handle it and the kick back of clumps of mud and grass may cause problems for others, especially because they aren't used to it.

The top 2 almost had to have run huge because they were so far ahead of the rest of the field, even if it was partially related to the surface it was still a big margin. But it would not shock me if all the time charts trying to equalize various distances were off on Saturday. 5 1/2F vs 12F on that track is not necessarily same thing as 5 1/2F vs. 12F on a firm highway.

My guess is that they both ran really well, but not an all time turf figure well.

[cj]

I think he is too high on that number, and I have it high for me as is. He's another 10 points faster. Differing opinions is what makes the game great.

[Tom]

Thank God I will never play a horse out of either of those races!

[jay68802]

Originally Posted by cj

I think he is too high on that number, and I have it high for me as is. He's another 10 points faster. Differing opinions is what makes the game great.

I estimated what the Bris figure would be and agree with you. My guess is that is about 8 points high when compared to what I think the Bris figure will be.

[classhandicapper]

Originally Posted by cj

I think he is too high on that number, and I have it high for me as is. He's another 10 points faster. Differing opinions is what makes the game great.

Yeah.

This is what I am getting at in the other thread on figures.

This is where coming at it from a non "time" based analysis could help.

The turf sprint division in the US is a 2nd or 3rd tier division. Horses move from limited winner ALW races to stakes races very easily and also from minor stakes to the best graded stakes company all the time.

Stormy Liberal is an established top veteran in that division, but he's no superstar. World of Trouble is an up and coming very good 3yo turf sprinter. Disco Partner is well known to not like soft turf and most of the rest are listed stakes or Grade 3 types.

I'm not buying that a 119 reflects the ability of those horses. IMO, these are not all time great turfers and that was not an all time great turf performance. They are very good turf sprinters and ran a big race but they are not 119s. Either the figure is off or there are some other developmental or other reasons that explain it.

Now if we are talking about Enable, that's something else. She's a multiple Grade 1 winner, double Arc winner, and she and the runner up trounced a field of legit Grade 1 horses. There may be reasons the margin was so huge there also, but those are spectacular horses.

[bobphilo]

Originally Posted by cj

I think he is too high on that number, and I have it high for me as is. He's another 10 points faster. Differing opinions is what makes the game great.

This is where the art part of figure making comes in. The figure maker must know when to cut a race loose for a particular race when the time does not make sense according to the horses in that race.
Those that use this as a criticism of figure handicapping must stop assuming that figures are made blindly and cannot be useful. They are attacking a straw man.

[JJMartin]

Anybody know if the generic speed rating in the Bris data file is an Equibase speed rating? It is the one that comes after Bris speed in field 856. What about the "DRF" speed in the HDW file?

[Tom]

BRIS SR is their own number,not EB.
Then comes the DRF SR and the DRF TV,
Not sure on the HDW files.

846- 855 BRIS Speed Rating
856- 865 Speed Rating (DRF)
866- 875 Track Variant (DRF)