I dug up the old data that I ran before. This was done in the Fall of 2000, about 2 years ago:
May Double/Split May Not Double/Split
0 0.000 -0.62% 1.13 -2.37% 0.98
1 1.000 -0.11% 1.13 -1.96% 0.99
2 2.000 0.40% 1.14 -1.55% 0.99
3 3.000 0.90% 1.13 -1.16% 0.99
4 4.000 1.39% 1.15 -0.76% 0.99
5 5.000 1.91% 1.17 -0.36% 0.99
6 6.000 2.45% 1.16 0.04% 0.99
7 7.000 2.99% 1.17 0.42% 0.99
8 8.000 3.52% 1.17 0.79% 0.99
9 9.000 4.07% 1.17 1.17% 1.00
10 10.000 4.61% 1.17 1.54% 1.00
Here is how you read this table.
At a TC of 10, the EV is 4.61% and SD is 1.17, if you play where you allowed to double or split, but only once. If you can't double or split, then your EV is only 1.54%, with an SD of 1.00 Note that 2*1.54% is 3.08%, which is still much lower than 4.61%.
This will change at higher true counts. At a TC of 52, the deck consists of only Aces and 10s, and so doubling will not be an issue. But this an extreme case. In practical BJ, you very seldom see counts higher than 10.
The method here was to use representative packs for the various true counts. A started with a 10 deck Shoe, and add one of each high card and removed one of each low card per TC point. I did a CA which used optimal strategy against that pack. This is only an approximation, but I believe that it is sufficiently accurate for the purposes here.
> there must be a reason. I would like to
> know.
> I catch an Ace for the first card several
> times everyday, how much should I bet if I
> have only 2 untis?
Bookmarks