About Statistics?

[Flysofree]

Is there anywhere to find out how the percentage of winners the public picks in the various types of races (Claiming, Alws, NG Stks GStks, MSW, MCL)? If the stat of winning favorites is 30% for all races --would there be no difference in the seperate categories?
Of course, I am talking about the averages over the years and not one specific track.


Also, when does a percentage become statistically meaningfull? I have noticed that on the news broadcasts whenever they do a poll which points out whatever type of yes/no question being asked; they will say something about + or - 4%. Is that where a meaningfull difference is found for all stats?

Thanks

[BeatTheChalk]

IMHO ... I think one has to break out results....by
Type of race .. distance .. class etc. For example ..many many
Seven Furlong races .. are NOT won by the favorite. Obviously
I have been looking at this topic for a long time ( since my
screen name here is ...BTC ! ) Cheers E...

[Flysofree]

Good point Beatthechalk,

I agree, but am looking for the facts. "just the facts, mam"


It's remarkable that the winning % of favorites has held so constant throughout the years. I think that says a lot. Just wish I could see the numbers and knew something about stats. I'm still puzzled about the + or - 4 % mentioned frequently.

[syyamo]

Flysofree,

I have to differ with your last statement a bit.

Steve Collison in "The Claiming Game" refers to a stat from "The American Racing Manual" of 1927 where as favorites won at a rate of 41%

I'm amazed that with all this information we have these days, we as a group perform worse than our ancestors.

Think what the extra 8% or so would do to everyone's ROI.

[Flysofree]

41% in 1927!! That is impressive. But how many track were actually in business in those days. Even if 75% or more of today's were around--that has been a LONG time ago-. Wonder when it became 30%+

[Secretariat]

Curious what the normal field size was in 1927. That could have accumulated to a higher favorite win percent.

[Blind Pursuit]

That +/- 4% or whatever is often called the "margin of error" although the technical term is "margin of sampling error." In a well-designed statistical test, you pick a random sample of things to look at (say, horse races) and measure how often the thing you're looking for shows up in the sample (say, winning favorites). But that only tells you how often the thing showed up in this particular sample. What if you picked a different sample? Would the results be the same? The "margin of sampling error" tells you that if you went out and did the same test on a whole new sample, you'd get the same answer plus or minus that percent.

[Hosshead]

I was also wondering about field size in 1927. Match races?

[Flysofree]

Sampling Error of 4% +/-.

That must be some complicated formula that was used to come up with that number of 4%. But I see what you are saying Blind Pursuit.



Field Size would definately play a role in 1927 or any year as to the percentage of winning favorites--I would think. 5 horse field vs 10 horse fields would seem to yield a higher percentage of favorites, I would think. But am not certain. I have no idea what fiel sizes were like in 1927.

[raybo]

Here's something to consider, concerning this subject. The amount and quality of data available to the general public today is, in my opinion, much better than that back in 1927. Also, the trainers are better equipped in many different ways to produce better results from their horses. The game has evolved in much the same way as other professional sports. Everything involved, almost, is better, which again, in my opinion, has produced more competition among the competitors. If their are more true competitors today, then almost certainly the public will be wagering on more horses in a given race which might have an effect on who the eventual post time favorite might be. I think it's harder for a good horse to win today than in the past because of all the previously stated reasons. The post time favorite today may win less often because there are others just as good today. And the public is better able to see good horses that are not the favorites because they are better informed by better data. The money becomes more split between horses and the favorite may more often, not be the best horse.

[Steve 'StatMan']

Plus a large number of us using our new and improved information are mostly trying to bet against the favorite instead of on it. If we wanted to pound an unprofitable chalk, we could, but many of us try to beat it, bet the exotics (not the win pool favorite, and most exotics not available in 1927), or pass the race.

[kenwoodallpromos]

Too many bettors today too confused by information overkill! 41% also means not as much manipulation to equalize the competition back then.

[raybo]

RE:<Plus a large number of us using our new and improved information are mostly trying to bet against the favorite instead of on it. If we wanted to pound an unprofitable chalk, we could, but many of us try to beat it, bet the exotics (not the win pool favorite, and most exotics not available in 1927), or pass the race.>

Good point Steve, I for one, do exactly that. If I can't put the favorite off the top row in a tri or a super, there better be some good odds on the other rows, or I pass the race. I never bet win bets anyway, and certainly wouldn't be betting on favorites if I did.

[Flysofree]

All of this has been interesting. I am different from most of you here in that I do NOT bet the exotics, except on rare occassions.

I am a Win OR Place bettor. Notice I said Or and not "and". I don't bet both in the same race. Probably 50% of races that I handicap I find no bet at all. However, to each his own style. Not only is the track takeout higher on the exotics, but the chances of hitting an individual race are less of course. Also, when you do hit that occassional 300 to 1 and up price; you have a silent partner that demands his portion. (LOL).

What I am trying to find out is : Are there any statistical differences in the favorite's wins when breaking down race types.
It suggests from these comments overall that you don't think so.
I don't know for sure, but I think there is in Maiden races and Stake Races. However, I don't have enough data to back up that claim.

[LOU M.]

What is the name of Dan Popes' book.
Thanks, Lou M.