Average odds of favorite

[highnote]

Does anyone know what the average odds of the favorite are?

I'm not interested in field size, class or surface. Just lump all races together then sum the odds of the favorite and then divide by the number of races.

I'd guess 2-1, but would like to know the actual answer.

Thanks.

js

[Overlay]

Without having the statistics at hand, I'd estimate that it would be 2-1 (representing the one-third of races that favorites have historically won), minus the average amount of the track take combined with the effects of breakage, which would put the odds at around 3-2, or perhaps slightly lower (7-5). (If the average payoff were in fact a full 2-1, you could break even just by betting the favorite.)

[Hosshead]

I'll take a wild guess. 8/5

[sjk]

I get 1.617

[highnote]

Thanks. I forgot about track take. That's the killer.

[Bruddah]

is their a significant difference between the mean odds and average odds?

[highnote]

No. Mean is another word for average.

Median is different than average. Median refers to the middle value of a set of numbers when they are arranged in order. For example:

Take the sequence of numbers: 1, 5, 7, 9, 10

7 is the median number.

6.4 is the average number.

Now take this sequence: 1, 5, 7, 8, 9, 10

7.5 is the median -- 7.5 is also the average of the two middle numbers 7 and 8.

6.67 is the average.

So when the sequence of numbers is odd, the middle number is the median.

When the sequence of numbers is even, the average of the two middle numbers is the median.

[Overlay]

"Mean" is what most people are referring to when they say "the average", but the median is also a type of average, just arrived at by a different procedure, and usually used when you want to discount the effects of "outliers" (unusually high or low values on either end of the range of values that you are calculating the average for).

A third type of average would be the mode, which would be figuring the number of favorites that paid off at each individual mutuel price (such as $2.10, $2.20, $2.40, and on up), and taking the price that the greatest number of favorites paid off at as the average.

If I recall, in a perfect normal-curve distribution (which you hardly ever see), all three of these values will be the same. The extent to which they are different for any given set of values highlights how those values differ from a normal distribution.

[Overlay]

Thanks. I forgot about track take. That's the killer.

And don't we all wish we could forget about the track take? :D

Also, correction to my last post: delete that parenthetical comment I made about the normal curve. Perfect distributions are in fact fairly common. (That's why it's called the normal curve.) The point I was trying to make was that calculating the various types of averages helps you analyze the ways in which the distribution of the data that you're working with differs from normal, so you can take that into account in your analysis of it.

[Bruddah]

I did not ask my question correctly. It was early morning when I posed it. Does anyone have the median odds. I just wondered if there was a big difference between the two. (avg. and median)

Thanks

[twindouble]

[QUOTE=swetyejohn]Thanks. I forgot about track take. That's the killer./QUOTE]


swetyejohn; No the biggest killer is inflation. A player today has to have a ROI a heck of lot greater than he did 40 years ago. For example, if I hit for $24,000.00 in 1965, I could have baught 3 new cars, today I get one. 1963 my father built a new cape home for $9500.00 and the contractor made money. For the average person to stay above water today you need two incomes. Anyone who brags about making $10 grand in any year on the horses is living in la la land.


Good luck,

T.D.

[highnote]

[QUOTE=swetyejohn]Thanks. I forgot about track take. That's the killer./QUOTE]


swetyejohn; No the biggest killer is inflation. A player today has to have a ROI a heck of lot greater than he did 40 years ago. For example, if I hit for $24,000.00 in 1965, I could have baught 3 new cars, today I get one. 1963 my father built a new cape home for $9500.00 and the contractor made money. For the average person to stay above water today you need two incomes. Anyone who brags about making $10 grand in any year on the horses is living in la la land.


Good luck,

T.D.


TD,

Don't forget though, your two dollar bets are a lot smaller now, too, due to inflation. So when you bet $2 today, it's really like betting 50 or 60 cents in 1965 dollars. So it should even out to some degree -- shouldn't it?

[rokitman]

http://www.horsestreet.com/freestuff/statistics/winfavs.html

[twindouble]

TD,

Don't forget though, your two dollar bets are a lot smaller now, too, due to inflation. So when you bet $2 today, it's really like betting 50 or 60 cents in 1965 dollars. So it should even out to some degree -- shouldn't it?

Ya sure, we'll all get rich playing two cent supers and penny stocks.

[highnote]

Ya sure, we'll all get rich playing two cent supers and penny stocks.


Don't forget nickle slots! :D

[arkansasman]

The mean of the favorite in my database is 1.694

[arkansasman]

The median of the favorite in my database is 1.70

[plainolebill]

1.59 over the past few years at the Socal tracks. (Excluding Fairplex)

[highnote]

1.59 over the past few years at the Socal tracks. (Excluding Fairplex)


Do you think your average is lower because of the smaller fields?

[plainolebill]

Short Fields sure could account for it.

[Grifter]

Dick Mitchell (Winning Thoroughbred Strategies) points out that the odds on the favorite go up as the day progreeses, because the degenerates among us chase after higher-odds horses. To get a number for average odds that has some real utility, you'd have to factor in field size, the number of the race (i.e., 1st race, 9th race, etc.), and perhaps type of race (e.g., less formful maiden claimers probably have higher-priced favorites).

Eventually, I'll run that regression analysis myself. It's Task #459 right now.... (sigh)....

--Grifter

[Rook]

Here are the average odds for favourites sorted by field size and class:

St Class Odds Num
6 CLM 1.48 18936
6 STK 1.1 393
6 MC 1.27 3890
6 MSW 1.24 2500
6 NW 1.32 5024
7 CLM 1.63 29144
7 STK 1.36 437
7 MC 1.43 7713
7 MSW 1.37 4537
7 NW 1.49 6752
8 CLM 1.76 28539
8 STK 1.57 431
8 MC 1.57 8722
8 MSW 1.51 5429
8 NW 1.65 6255
9 CLM 1.91 23189
9 STK 1.67 339
9 MC 1.71 8472
9 MSW 1.63 5158
9 NW 1.77 4593
10 CLM 2.01 24096
10 STK 1.88 235
10 MC 1.79 9136
10 MSW 1.71 6541
10 NW 1.93 4646
11 CLM 2.09 7022
11 STK 2.1 151
11 MC 1.89 4350
11 MSW 1.81 2506
11 NW 2.05 1325
12 CLM 2.21 6034
12 STK 2.12 114
12 MC 2.00 4316
12 MSW 1.91 2886
12 NW 2.19 1081
13 CLM 2.12 193
13 STK 2.59 34
13 MC 2.16 299
13 MSW 2.23 173
13 NW 2.43 56
14 CLM 2.28 148
14 STK 2.54 19
14 MC 2.21 252
14 MSW 2.02 140
14 NW 2.53 45

[RonTiller]

There are wide variations in the average odds of favorites and the favorite win percentage at various tracks, independent of field size. The following is from all races since 1997, with the caveat that I did not make corrections for coupled favorites.

Looking only at 8 horse fields across various tracks, I get the 2 highest average favorite odds at Canadian tracks:

Trk Avg W%
NP 1.96 29%
STP 1.96 29%

The two worst:

Trk Avg W%
RP 1.38 37%
RET 1.44 34%

There is no straight linear correlation between the average favorite odds and the favorite win% in this sample - that is, there really are good and bad tracks to bet favorites, independent of field size.

I was suspicious of the idea that odds of favorites go up throughout the race day, but, well, I was wrong. There is a general trend, the most dramatic being that race 1, no matter how I slice and dice, always seems to have the lowest favorite odds of the race card.

Only looking at 8 horse fields again, all tracks, by race #:

R# Avg W%
1 1.48 34%
2 1.57 33%
3 1.58 34%
4 1.61 33%
5 1.62 33%
6 1.64 32%
7 1.64 32%
8 1.61 32%
9 1.63 33%
10 1.63 34%

In 6 horse fields the odds peak mid-card then decline almost back to the starting point. 10 horse fields have a beautiful line, with the highest odds in race 10, but the same win%, 30%, as most of the other races.

Race 1 always has the lowest odds, even if one further subdivides the data by race type, but the win% doesn't vary much throughout the card.

Conclusions:

1. Bet Northland Park and Stampede Park favorites AFTER race 1.
2. Stay away from the hot dogs.

Ron Tiller
HDW

[highnote]

Good info everyone. Thanks.

[Grifter]

Rook / Ron -- Good stuff... thanks.

--Grifter