Thx,
I basically have
better than perfect counting method at predicting next 1-2 cards (about 13% to 50% lift vs Perfect Counting(depends on the circumstances, i can predict the lift factor)). Tried simulate betting spread (not yet full optimized), to have 1.4%/1.6% edge over House (The Rules are quite bad with 0.993 edge when playing normal strategy and flat bets.).
But the problem is %Ruin before double at 12,7% to 20,5%(with 1000 betting units) (the latter numbers are agressive betting), when using Basic Strategy Table from
https://wizardofodds.com/games/black...gy/calculator/.
I seriously think I m doing something wrong...
The problem now i m facing is to make better Strategy table, basic on my prediction, and where to seat. (better betting at 1-2 table VS better prediction of D 2nd card at latter seats), so far I think the more weight is on better betting than better strategy.
Also the problem there is problem in my method that if at seats 1-2, the D 2nd card nad my 3nd card, can Influence each other, so for decision I would need to code alot of If and split....my head seriously hurts. I tried to draw schemas on paper.
..
Well sorry for my rambling, back to codding.
I think I will make it easy, ignoring BinominalProblem first :
Dealer second card= Pd=[pd0...pd9], then make "reward" based on decision versus that and make
sum (of pdX*RewardHitX )vs sum (of pdX*RewardStayX) vs DD vs Split...etc....
then i will try to get rid of the Bionomenal pair problem later with something...
Altrough, theoretically, I can have companion who can theoretically only specialize on busting the dealer, but I think Lift factor is not so great usually, that It could pay for his time and lost of basic betting unit each hand. But thx for idea
Bookmarks