Beyer ART or SCIENCE

[PaceAdvantage]

Originally Posted by karlskorner

The Vet lady is back. If it looks like a duck, etc. etc.

That's not exactly where I was going...and I think you're probably wrong on that conclusion.

[Tom]

Originally Posted by Tom

Hey Raaaaalph!
We was right!

Hey Raaaaalph!
We was wrong!

[karlskorner]

Not yet, my friend, not yet. Don't put you finger in the dyke

[JulieKrone]

I'm real disappointed in both you guys' handicapping skills-- a duck trying to sneak back in would NOT quack-- maybe moo, or something else? Why would someone such practically unmask themselves in their last reply to the moderator(if that was the case)? That's too obvious to be so, like betting on a favorite dropping down ten levels.
BTW I heartily approved(will always) of the banning you allude to; in fact it needs to be exercised way more often imho.
And please don't even think of brooking quarel with me, as I never speak ill of anyone in any community where I post. The very worst you could pin on me is a possible dishonoring thru stated opinion of some bigwig, and does a horseplayer exist not guilty of this?

[NoDayJob]

MYTHBUSTERS!

[BMeadow]

Some general comments on speed figures, from an article I wrote for the March 2002 edition of Meadow's Racing Monthly:

Crafting speed figures is hard work. But even with a solid method of creating figures, the whole business is fraught with traps. If you are going to get into the speed-figure game (either as a maker or a user), you might want to consider:

* Projections vs. pars. Sometimes the four best $20,000 claimers on the grounds get into the same race. When the final time of their race exceeds par, figure makers who are simply using a par chart may be misled into thinking the track is faster than usual when it was simply this particular field that was fast.

* Ground loss adjustments. A horse going four wide around the turn loses valuable ground and actually outran a horse who beat him by a nose but saved all ground. But was he four wide around the whole turn, or just part? And maybe the inside horses got dirt in his face most of the way around while the wide horse was in the clear.

* Troubled trips. A horse who broke three lengths behind the field or who had his momentum stopped as he was trying to close might need to be upgraded--or not, depending on other factors.

* Perfect trips. A horse who coasted on an easy lead, or who sat behind three duelers and sailed by them as they tired, will wind up with an inflated number.

* Chart errors. While the final margins may be accurate, pace margins are notoriously inaccurate. Was that horse four lengths back at the half or six? Could make a big difference to your numbers.

* Weight adjustments. Two horses run 1:11. One is carrying 107 pounds, the other 122. Should they be awarded equal ratings?

* Biases. Some days a bias keeps a front-runner going, or propels a closer. Some days the rail is the worst place to be. How does this enter into your figure-making?

* Fliers. The clocking of a race does not begin at the gate, but starts once the first horse passes an electronic beam. But if a flier got a two-length jump on the rest of the field in the first few yards, everyone’s times will be askew.

* Timing errors. Remember the flap at Gulfstream several years back when dozens of races were found to have been mistimed? Good luck depending on the hand times at summer fairs.

* Swirling winds. A tailwind will propel a horse and a headwind will sap his energy. Strange internal clockings often result.

* Big win. Here’s a problem that figure makers constantly face: A horse wins by eight lengths, with the rest of the field bunched. Did he run spectacularly well, or was everyone else dreadful? Do you start your analysis with the 1:10.1 run by the winner, or the 1:11.4 run by the second-place finisher?

* Off tracks. Some horses will glide over a wet track while others get mired in it. Worse, some tracks get faster or slower as the day progresses. How about mid-card rains and winds? You might need to make individual variants for each race rather than a single one for the day.

* Lack of samples. What if there’s only one route all day, and the pace was weird? Do you try to relate it to your sprint number? Use it as is? Throw it out altogether?

* Turf races. About distances, rails up or down, course is soft or firm, grass has been recently cut or not, strange pace situations--it's a complete mess.

* Pace. How does it affect the final time? Do you credit a horse for closing off a fast pace, downgrade him because he passed tired horses, or neither?

Perhaps the biggest problem with speed figures is their inherent nature: They measure what a horse has done under a particular set of circumstances. How likely is he to get that same set again? Usually, not very. While he might be in the same class or even face a few familiar opponents, he (and they) may be at a different phase in their form cycle. The posts will be different, the distance might be an extra furlong, the track may be harder or softer, etc. Predicting a change is the key component in being able to outwit the crowd.

And none of the above discussion includes such questions as what was the horse doing in the race, how did he appear in the paddock before the race, where was he in his form cycle, what has happened to him since that race, etc.

[classhandicapper]

Barry,

You did an excellent job of identifying all the problems.

Another thing I see happen from time to time is the rough start where from the head on shot you can see horses ping ponging off each other. That impacts how quickly they get up to speed.

I'll add something else to the list for consideration.

What do you do when you look at multiple sets of figures and they disagree by several lengths among the major contenders?

That happens to me all the time because I often have 3-4 sources of figures. I go back to the races in question and analyze them for myself (I have experience with pace and final time figures myself), but very often it's not very clear who is right - even after the fact. Many of things you identified are subjective in their impact even when we have accurate information to work with.

I've always felt that all else being equal, a large edge in speed figures is usually conclusive, but given similar numeric ratings, you are better off looking at other factors to seperate them.

[kenwoodallpromos]

I will add 1 more by this question- which finishers are going all-out at the wire and which ones are saving energy while holding their finishing position?
___________
Should the projected finish time for the place horse in Secretaritast's Belmont run be based on 1 length= 1/5 second, or 2 seconds faster at 1/7 second (about 2:28, an average Belmont winner's time)?

[46zilzal]

Originally Posted by kenwoodallpromos

I will add 1 more by this question- which finishers are going all-out at the wire and which ones are saving energy while holding their finishing position?
_

Well since NO TWO races are the same, that is an individual situation

[kenwoodallpromos]

Correct! I belive Beyer when he says you have to consider some other variables along with speed figs. Figures get your horse in the ballpark often, but each race is different because horses are only human and not machines!

[TimesTheyRAChangin]

Originally Posted by kenwoodallpromos

horses are only human!

???????

Sorry P.A.,I just couldn't let that go by!

[Brian Flewwelling]

Barry that was an excellent discussion of the problems of handicapping with Speed Figures.

Speed Figures should not, an mostly don't, make adjustments for anything but the average effect of the Track on the race time. Then it is up to the handicapper to decide how the factors you mention, and more, should be applied to those PAST performances, then attempt to determine what the speed, relative to the other entrants, will be today.

So there are issues with Speed Figures, and issues with Handicapping with Speed Figures.

[rastajenk]

I think anyone who's ever made their own numbers, or anyone with a working knowledge of how they're made, has encountered the kinds of things Mr. Meadow has listed. But you know what? If they are used with confidence and generate some success, it doesn't really matter. You take it in stride, because you realize it's a good template for making your initial decisions.

Lots of other "systems" round data into averages of one kind or another, like earnings per start, or internal fractions, or stretch gain/loss, all kinds of things, but they don't receive a fraction of the microscope time that speed figures get. I guess that means speed figs are doing pretty good to have so many people constantly chipping away at them.

[Overlay]

That daunting list of all the unknowns in speed-point calculation and interpretation from Barry Meadow makes it even more impressive to me that at least some sets of figures do a good job of maintaining a consistent record of predicting actual performance (as long as you consider them in terms of bulk, long-term probabilities, as opposed to a guarantee of how any particular horse will run in a specific race).

[Overlay]

Originally Posted by Overlay

That daunting list of all the unknowns in speed-point calculation and interpretation...

Of course, I meant speed-figure calculation. (I guess I've been looking at Steve Klein's book a bit too much in the last few days!)