Effects of splitting with Kelly Criterion

In Kelly Criterion, the % advantage is used in correlation with the % of bankroll used; but it's more ideal in straight forward bets (like if you had an edge in Baccarat) than in blackjack (with splits and doubles.)

All doubles are more than a 2% player advantage, so they should fall within Kelly criterion pretty well. But splits are often defensive. I looked at the numbers and played practice hands to see how they work in practice. Splitting 6's and 8's are the only hand that has a negative expectation overall, surrendering 8's v. T seems to solve that for the 8's, 6's are only split against bust cards and in practice is not as bad as the 2.5% disadvantage against all splits combined on average states.

However, the numbers I looked at were for infinite deck; what are the effects of splitting with +3TC in hilo?

I think norm has some pages that go in to more physical numbers for certain splits/hands at different true counts... but at TC +3 you're starting to introduce a few decent deviations that have some effect. However, there are still "defensive" splits as you put it. I would imagine at a higher true count everything would remain about the same "for the most part" except you might get more positive splits from deviations.

Quote: billionaireben

In Kelly Criterion, the % advantage is used in correlation with the % of bankroll used; but it's more ideal in straight forward bets (like if you had an edge in Baccarat) than in blackjack (with splits and doubles.)

All doubles are more than a 2% player advantage, so they should fall within Kelly criterion pretty well. But splits are often defensive. I looked at the numbers and played practice hands to see how they work in practice. Splitting 6's and 8's are the only hand that has a negative expectation overall, surrendering 8's v. T seems to solve that for the 8's, 6's are only split against bust cards and in practice is not as bad as the 2.5% disadvantage against all splits combined on average states.

However, the numbers I looked at were for infinite deck; what are the effects of splitting with +3TC in hilo?

billionaireben,

In August 2012 I posted an analysis of starting hands vs. dealer's upcard for a 6D H17 DAS game: if you have BJ21 Green Chip access, see this link:

https://bj21.com/boards/green-chip-forums/sub_boards/posts-of-the-month-archive/topics/aug-2012-positive-ev-hands

In particular, a starting hand of 8,8 is +EV at all HiLo TC's of -10 and up vs. upcards of 2-7. Against an 8, it is +EV only for +3 and above. It is never +EV vs. 9, X, or A.

A starting hand of 6,6 is +EV for all HiLo TC's of -10 and up vs. upcards of 4-6. Against a 3, it is +EV only for +2 and above. It is never +EV vs. 2, 7, 8, 9, X, or A.

Hope this helps!

Dog Hand

Quote: DogHand

...A starting hand of 6,6 is +EV for all HiLo TC's of -10 and up vs. upcards of 4-6. Against a 3, it is +EV only for +2 and above. It is never +EV vs. 2, 7, 8, 9, X, or A...

I'm confused by this. What if there are literally only 10's left in the deck and the dealer shows a 2 up and you have 6,6? Surely it's +EV to split your 6's when you know the dealer will bust? I guess the same could be said for 7's in the crazy case that there were only Aces left (perhaps end game strategy at foreign places that deal to the last card?). Just thinking out loud there must be "some" time when they're +EV.

Quote: Romes

I'm confused by this. What if there are literally only 10's left in the deck and the dealer shows a 2 up and you have 6,6? Surely it's +EV to split your 6's when you know the dealer will bust? I guess the same could be said for 7's in the crazy case that there were only Aces left (perhaps end game strategy at foreign places that deal to the last card?). Just thinking out loud there must be "some" time when they're +EV.

Romes,

Good point! I should have clarified that my work involved TC's in the range of -10 to +10, so I missed the TC= +52 situation you described.

Dog Hand