database folks-help with ML

[kitts]

Currently I am trying to determine methods that can show the winner of non-maiden races 80% of the time among the contenders selected. I have been relying on the Morning Line for years to make betting decisions but I would be the first to acknowledge the discrepancy between Morning Line and Final odds. I have made peace with that and now wonder: do the top five Morning Line horses have the winner among them at least 80% of the time?

[formula_2002]

Kitt, one of the top 4 public betting choices finishes 1st 81% of the time.

A dutch bet on a top four choices would return a loss of 19%.


Joe M

[hurrikane]

Kitt
It really depends on the track.
At AQU you can get 85% out of the top 4 mlo

At mnr you may as well throw it away.

[Dave Schwartz]

Depends upon field size.

[formula_2002]

Actually, it depends upon the odds, which takes the field size and everything else into account.

Sum (1/ (odds+1)) X (1-track out) for the top choice and it will win at that rate.

Add that rate to the 2nd choice, and that’s how often one of the two to will finish 1st.

Go right down the line to the last horse in the race and the sum of all the rates =100%.
So bet all the horses and win every race.
It’s all an approximation, but it’s as good as it gets.

There is a slight bias in the rate towards short priced favorites.

Joe M

[anotherdave]

Originally posted by formula_2002
Actually, it depends upon the odds, which takes the field size and everything else into account.


I agree with Dave. (Hey it had to happen) I guarantee you that in a 4 horse race, regardless of the odds, the probability of the winner being in the lowest 4 morning line odds is not 81%.

And I think Kitts is asking about the lowest 4 odds on the Morning Line not the lowest four odds.

AD

[formula_2002]

I moved the converation to post time odds. It's the most realistic way of calculating things like what % of the horses wil ......(anything)


sum the (1/odds+1) x (1-track take out) and will come up with 1.

Joe M

[kitts]

I was talking ML odds and I am content with the reply that the top 4 have the winner among them a lot. I am truly aware how the field size impacts and I also realize that the database finding was based on a whole bunch of events that happened in the past and are not likely to recur. I have utilized the info I got early on and am comparing that to the Top 5 picked by a software program. Ideally the contenders would be among the top 4 ML favorites AND among the top 5 of the software. But, as we all know, that is just the beginning. I am checking out the approach I just developed. It is unlikely, as any semi-mechanical system is, to succeed, but I'll let you know.

[Whirlaway]

Kitts, based on a roughly 10,000 race sample:

Top 5 MLO: 87.44% winners

Top 4 MLO: 79.31% winners

Top 3 MLO: 67.21% winners

Top 2 MLO: 51.22% winners

Top 1 MLO: 31.19% winners

Average Field size: 8.39

[formula_2002]

Whirlaway
Thanks.Nice pice of information.

Let me ask if you can offer the following.

List the ratio of acutual wins divided by (1/(odds+1)) for each of the TOP MLO groups presented.

A ratio greater willl indicate a dutch bet profit.

ratio-1= % profit or loss

thanks again

Joe M

[andicap]

Does the top ML horse have a similar ROI to the betting favorite?

[hurrikane]

whirl....10000 races of what?

Probably, as Dave says...fld size is one of the most important factors.

Also...is the MLO fav 3-1 or 7/5 ?

Are you playing at SA or MNR

HUGE DIFFERENCES!!!

[kitts]

Whirlaway-
Just what I wanted. I am going to continue using ML odds in my handicapping so this stat is great for me. I know how final odds are more predictive and how field size and track will impact all this. But I just wanted the number and now I've got it. Netcapper tracks ML and it can be predictive. Give me a horse in the top 4 ML with a top last race 1st and 2nd call and I will take it from there.