Betting Multiple Selections, Database Related

[NormanTD]

I have about 3 good filters in my database that show a positive roi over some number of races. Occasionally, Filter_A and Filter_B pick different horses in the same race. Even less often, Filter_C will join in and I have 3 selections in one race. I, of course, have no real idea of what to do, but I've come up with 2 different possible solutions and would appreciate any and all comments on the merits of each.

Solution_A - Bet both horses that pop up (sometimes 3) but consider each bet as coming from a separate bankroll, i.e. Bankroll_A for Filter_A, Bankroll_B for Filter_B, Bankroll_C for Filter_C.

Solution_B - Bet the Filter that is showing the largest roi.

I lean toward Solution_A since to me that is the most simple and clear solution, however, I don't have any real experience at this since the filters have just been developed.

Thanks in advance.

[K9Pup]

I have about 3 good filters in my database that show a positive roi over some number of races. Occasionally, Filter_A and Filter_B pick different horses in the same race. Even less often, Filter_C will join in and I have 3 selections in one race. I, of course, have no real idea of what to do, but I've come up with 2 different possible solutions and would appreciate any and all comments on the merits of each.

Solution_A - Bet both horses that pop up (sometimes 3) but consider each bet as coming from a separate bankroll, i.e. Bankroll_A for Filter_A, Bankroll_B for Filter_B, Bankroll_C for Filter_C.

Solution_B - Bet the Filter that is showing the largest roi.

I lean toward Solution_A since to me that is the most simple and clear solution, however, I don't have any real experience at this since the filters have just been developed.

Thanks in advance.

It depends. When you determined that filter A,B,C showed a profit did you EXCLUDE races where 2 or 3 of them might have triggered? If not, then I say play as many as the filters select.

I'm not sure why the concern of "separate" bankroll????

[cj]

It should be pretty easy to test your solutions when you get more than one horse in the same race.

Personally, for me, unless both are more than 8 to 1, I just bet the highest odds and it works well. Your results may vary.

[punteray]

CJ,

Have you considered dutching the two or three horses?

Ray Scalise

[cj]

CJ,

Have you considered dutching the two or three horses?

Ray Scalise

Odds are too unpredictable in my opinion for dutching to really work, even at 0 mtp.

[Light]

Norm

The obvious is to play all filtered selections in a pk3 or 4.Your Roi will soar. Even if you hit 2 $20 payouts on win,so what. Maybe you invested $18. Profit $22. In a $1 pk3,the profit is at least 20 fold.

What would make this approach still stronger is to find the hit rate for specific class,sex,age,surface and distance that your program excels in. Then search for races where the pk3 or 4's contain those kinds of races.

Can you imagine your edge when you can key 1 leg knowing when you bet your top 2 selections on those kinds of races you have a 60% chance or bettor of hitting it. Now you can tell where to key in a scientific manner as opposed to a TVG guy or average capper whose reasoning for keying a race depends on their mood.

[NormanTD]

It depends. When you determined that filter A,B,C showed a profit did you EXCLUDE races where 2 or 3 of them might have triggered? If not, then I say play as many as the filters select.

I'm not sure why the concern of "separate" bankroll????

I was thinking in terms of separate bankrolls so as not to do the old "betting against yourself by betting more than one horse in a race" thing. With separate bankrolls, I could look at bankroll growth (or loss) for each filter as opposed to only using one bankroll.

[NormanTD]

It should be pretty easy to test your solutions when you get more than one horse in the same race.

Personally, for me, unless both are more than 8 to 1, I just bet the highest odds and it works well. Your results may vary.

I might be able to look at this. As simple as it is, I never thought of it.

[NormanTD]

Norm

The obvious is to play all filtered selections in a pk3 or 4.Your Roi will soar. Even if you hit 2 $20 payouts on win,so what. Maybe you invested $18. Profit $22. In a $1 pk3,the profit is at least 20 fold.

What would make this approach still stronger is to find the hit rate for specific class,sex,age,surface and distance that your program excels in. Then search for races where the pk3 or 4's contain those kinds of races.

Can you imagine your edge when you can key 1 leg knowing when you bet your top 2 selections on those kinds of races you have a 60% chance or bettor of hitting it. Now you can tell where to key in a scientific manner as opposed to a TVG guy or average capper whose reasoning for keying a race depends on their mood.

Whoa! I never thought of serial bets either.

[NormanTD]

Thanks gang. Some good ideas for me to pursue. I knew I came to the right place!

[Overlay]

I was thinking in terms of separate bankrolls so as not to do the old "betting against yourself by betting more than one horse in a race" thing. With separate bankrolls, I could look at bankroll growth (or loss) for each filter as opposed to only using one bankroll.

With multiple horses and bankrolls, I was just curious as to what criteria you would plan to apply for sizing your bets. Would you bet according to the degree of your demonstrated ROI edge (or some proportion of your edge) with each filter, or would you use some other ratio or measure? Would there be any odds ranges that you would consider unbettable?

[NormanTD]

Three words...

HMI

[K9Pup]

I was thinking in terms of separate bankrolls so as not to do the old "betting against yourself by betting more than one horse in a race" thing. With separate bankrolls, I could look at bankroll growth (or loss) for each filter as opposed to only using one bankroll.

From a database standpoint you should be able to look at them as separate or combined. Personally I don't consider betting multiples "betting against yourself". It's more like covering all the bases.

I'm curious, say you play a tri wheel of 1/2/345 do you consider the 1/2/4 and 1/2/5 bets as betting against yourself?

[NormanTD]

I like the covering all bases concept. Yeah, that's what I'm trying to do. I don't bet tri-wheels so I don't really have an answer for that.

[ryesteve]

The solution is to use your database to see what has happened in the past when you have multiple selections. Depending on how your filters work, it's possible that the presence of multiple selections might decrease their value (like for instance, if your filters all had an early speed component, multiple selections could suggests your picks will compromise each other's chances)