odds line EVALUATION

[formula_2002]

I have attached an excel file which

COMPARE’S THE ACCURACY OF SEVERAL ODDS LINES.MY ODDS LINE, CALL QUANTUM, SOMEONE ELSE'S ODDS LINE CALLED "X PROB" AND THE PUBLIC'S ODDS LINE, CALLED "EXPECTED"
WHICH IS THE SUM 1/ (ODDS+1)

ACCURACY IS DETERMINED BY DIVIDING THE SUM OF THE PROBABILITY
BY THE ACTUAL WINNERS
"1" IS PERFECT.

QUESTION.
HOW CAN I BEST USE THE "QUANTUM" LINE TO FIND OVERLAYS??

THANKS
JOE M

[robert99]

formula_2002,

Not knowing where your data has come from and what take out allowances made it looks that horses that you think are 4/1 or more but the public bets to be longer odds are the ones to look for. Shorter than your 4/1 estimate the public on average is betting to shorter odds than your average estimate.

It is no point looking just for a horse that is at the greatest positive difference in odds with your and the public. The value horse has to have a good chance of winning for you to collect, so do not stray too far up for horses above 4/1 your odds.

You could look at it this way:

You say horse is 4/1 chance (20%)
public says it is a 5/1 chance (17%) so your bet might have a 20% chance of collecting a 3% edge bonus. If you are wrong for the "true" odds, you might only have a 17% chance of collecting zero bonus. The truth lies somewhere in the middle and can only be found by trial runs. You have the data so run a simulation.

You say horse is 5/1 chance (17%)
public says it is a 6/1 chance (16%) so you might have a 17% chance of collecting a 1% edge bonus etc. It does not look at first glance much different from the first example but it is a much poorer value bet long term.

[formula_2002]

I'll be testing the following against an entirly new data base tomorrow.
Play any race where the "Quantum" probability is >=45% .
In this data base, it returned "even" in 1428 plays.
Win % is >50%