Finding a subset for favorites

[soflant]

I am wondering if there has been a study done or books where the research shows a subset where a profit is shown for betting favorites? I mean a profit of at least 10%-20% Did Burton Fabricand so such research? As we all know betting every favorite shows about a 10-20% loss at the end of a meet. Thanks so much in advance.

[pandy]

I don't think that's possible unless you could find an ADW that gives you 10% rebates on win wagers. But let's see if anyone answers your query with better information.

[romankoz]

There are three types of favourites: true, middling and false. There will never be a book or paper published that will show a profit of 10% on all three types as a combination.Where 10% can be made is by ONLY backing true favourites and/or backing 2nd favs against false favourites. You mention the word "subset" and this is where a major problem can occur if the sample figure is too low.

If I was in the US I would be recording all the parimutuel final dividends (SP for STarting Price) and start columns for a range of factors (distance, actual tracks, last stat winners, class .. whatever) and have some sort of Excel way of investigating what works where and how. I would do this on all horses that start at 4/1 and less or 3/1 and less or 5/2 or less or 2/1 or less ... etc. (you set the range)
I would imagine there would be some tracks that would be poison but in the finish the sample figure would have to be 500 plus for each factor. Some trainers and or jockeys the same.

I am positive this is being done ad nauseum across the US but it is the interpretation of the figures that needs to be well done.

[formula_2002]

Originally Posted by soflant

I am wondering if there has been a study done or books where the research shows a subset where a profit is shown for betting favorites? I mean a profit of at least 10%-20% Did Burton Fabricand so such research? As we all know betting every favorite shows about a 10-20% loss at the end of a meet. Thanks so much in advance.

yes he did
its all in his book "Horse Sense"
He based his play on the Theory Of Maximum Confusion using 15 factors.

Amazon has the book for about $100.

I never tested the system, but his statistical analysis presentation is quite clear and relatively simple.

it's a hallmark for every serious player or non-player

[sjk]

I had the book in the 60s. The sample sizes were around 3,000 races as I recall which is way too small to compare with what is possible today.


As someone who has bet money on tens of thousands of races in the years since I would not recommend the principal of maximum confusion working out.

[formula_2002]

Originally Posted by sjk

I had the book in the 60s. The sample sizes were around 3,000 races as I recall which is way too small to compare with what is possible today.


As someone who has bet money on tens of thousands of races in the years since I would not recommend the principal of maximum confusion working out.

He looked at all favorites, 6o88 races. He looked at 1478 system favorites and had successful z scores of 2.5 for 4 of the 6 incremental odd sets.
Benter's paper looked at 3198 races and as far as I can determined never produced a similar table of his system plays.

[PICSIX]

I believe Robert Goren posted about a sub-set where the M/L was below 1/1. I did an advanced search for threads he started but can't find it?

[headhawg]

Originally Posted by formula_2002

He looked at all favorites, 6o88 races. He looked at 1478 system favorites and had successful z scores of 2.5 for 4 of the 6 incremental odd sets.
Benter's paper looked at 3198 races and as far as I can determined never produced a similar table of his system plays.

And which guy made more money?

Your relentless pursuit of nothingness and then inserting your misguided "analysis" in too many threads is well beyond annoying at this point. And responses by respected members like sjk carry way more weight than any of your pseudo-statistical posts.

[PICSIX]

I found it. He was actually playing against "Bridge Jumpers"

http://www.paceadvantage.com/forum/s...ad.php?t=84416

[formula_2002]

Originally Posted by headhawg

And which guy made more money?

Your relentless pursuit of nothingness and then inserting your misguided "analysis" in too many threads is well beyond annoying at this point. And responses by respected members like sjk carry way more weight than any of your pseudo-statistical posts.

Sorry if the facts confound you.
As far as who made more money, I have no way of knowing.

[Johnny V]

Originally Posted by sjk

I had the book in the 60s. The sample sizes were around 3,000 races as I recall which is way too small to compare with what is possible today.


As someone who has bet money on tens of thousands of races in the years since I would not recommend the principal of maximum confusion working out.

The principal of maximum confusion is aptly named. The rules of similarity Fabricand puts forth with the many various subsets and exceptions is really almost impossible for most to really decipher. RPM has a software program called The Wizard of Odds that they claim is based on Fabricand's book. The description of the program reads like nothing described in Fabricand's book as far as I can tell. There is no mention of rules of similarity or anything like that, which is the central premise of the whole book. In their ad they mention Fabricand tested 200,000 races. Beware.

[headhawg]

Originally Posted by formula_2002

Sorry if the facts confound you.
As far as who made more money, I have no way of knowing.

Facts? I don't trust that you know what you are doing. You make stuff up as you go along, always changing the parameters of your "research" to make the results look more impressive. You are the ultimate backfitter. Unfortunately it is useless research as it can't be applied in real time with the same positive outcomes.

And you can't surmise that Benter made more money than Fabricand? Are you really that dense? Don't answer; it's clear.

[thaskalos]

Originally Posted by soflant

I am wondering if there has been a study done or books where the research shows a subset where a profit is shown for betting favorites? I mean a profit of at least 10%-20% Did Burton Fabricand so such research? As we all know betting every favorite shows about a 10-20% loss at the end of a meet. Thanks so much in advance.

You are looking not only for a profitable method that doesn't involve actual handicapping...but also for a profit of "at least 10%-20%". I am afraid that you are going to be disappointed in your search.

[formula_2002]

Originally Posted by headhawg

Facts? I don't trust that you know what you are doing. You make stuff up as you go along, always changing the parameters of your "research" to make the results look more impressive. You are the ultimate backfitter. Unfortunately it is useless research as it can't be applied in real time with the same positive outcomes.

And you can't surmise that Benter made more money than Fabricand? Are you really that dense? Don't answer; it's clear.

"Changing parameters " perhaps akin to Benter's statement "improvements to the model were made on a continuing basis, as were regular re-estimations of the model which incorporated the additional data accumulated. "

[Dave Schwartz]

Originally Posted by Johnny V

The principal of maximum confusion is aptly named. The rules of similarity Fabricand puts forth with the many various subsets and exceptions is really almost impossible for most to really decipher. RPM has a software program called The Wizard of Odds that they claim is based on Fabricand's book. The description of the program reads like nothing described in Fabricand's book as far as I can tell. There is no mention of rules of similarity or anything like that, which is the central premise of the whole book. In their ad they mention Fabricand tested 200,000 races. Beware.

Oh, there were pages and pages of rules.
And they were quite clearly written.

However...
A good friend and programmer coded Fabricand's rules.

Several of them were contradictory.

He spent quite a bit of time understanding them and eventually presented the evidence to me.

After several hours of working with him, I was forced to concur.

The rules were not completely logical.

Therefore, I would tend to agree with you.

For me, that put the entire body of work in question.