Research/Data Question. How would I go about…

[FlintAtTheFetlock]

The only experience I have with data files is BRISNET PP's when at one time I put some effort into crunching my own speed figures using an Access database. With that said, how would I go about researching an antidotal observation I have where 'odds on' favorites in lower to mid-level races finish 3rd or 4th? The angle is to find a clear, isolated second choice, and play cheap TRI and SUPER's. If multiple horses close in odds for runner up, don't play.

TRIFECTA

ALL // Second Post Time Horse // Odds on Favorite

SUPER

ALL // Second Post Time Horse // ALL // Odds on Favorite

SUPER

ALL// ALL // Second Post Time Horse/ Odds on Favorite


The idea is to bet small and hope the angle hits. Even if only 5% one score could cover the losses when an odd's on favorite finishes 3rd or 4th and the second choice gets beat. The proposition does not present all that often so its not tossing money around . It could be a relaxed way to weather slumps when your best handicapping is failing miserably

So where would one go to analyze this criteria? I 'd like to see if my observation is just a hunch or there is some data to back it up.

Thoughts? Resources? I do have programming and database experience, but don't want to rack my brain relearning coding syntax.

Criteria:

PT Favorite <= 1.0
PT Second Choice >= 5-2
PT Third Choice >= 6.0
Horses >= 8
Class < = Allowance @40k

Results:

PT Favorite 3rd or 4th
PT Second Choice 2nd (or 3rd)

That's what I'm looking for.
Thanks,
Flint

[DeltaLover]

Don’t waste your energy with this approach as for several reasons it has no chance to succeed.

First of all you have no clue about the final odds before the race is closed for betting and more than this your criteria represent another static handicapping factor which is very easy to back-test against a database and find out that it has a negative expectation (EVEN if you were able to magically predict the final odds in a posterior fashion). Since you have some programming experience, the way to go is to build a complete historical database covering at least two years of charts and on top of ut develop the necessary machine learning processes to assist to in your betting decisions.