Switching away from a 'Blackjack' ?
UK-21 said:
Hmmm . . . who programmed the simulator? Switching a 21 against a dealer 7,8 or 9 would be silly - it's virtually a guaranteed win. A 20 would be a likely win, but would leave a window open to lose, and an A,6 at a high count against a 7,8 or 9 is only marginally better than a 10,6. And not switching against a 6 at a high count??? By doing so you'd end up with a 20 and an A,6 double opportunity with a relatively high probability of the dealer busting?
I'm not a believer in challenging the maths with hunches, but sometimes hunches that are based on study and experience are not completely out of place and are worth following. In the example above, even if the sim is mathematically correct (again hmmm . . . ) there'll be very little in it.
Hi UK-21,
I wanted to check whether the simulator gave the correct decisions in the above example so I went to Mike Shackleford's site where he lists the ev's for each player hand verses the dealer upcard for 'Blackjack Switch'. Incidentally, the 'Switch' simulator was programmed by a highly regarded Blackjack programmer who was also responsible for the well known SBA product.
We are dealt A,10 and 10,6 and we can 'switch' to 10,10 and A,6. The simulator produced a 'switch' verses a dealer 7,8 & 9 and a 'no switch' for all other dealer upcards.
Looking at the ev's on Mike's site, the following numbers were attained for each hand and dealer upcard :-
A,10 obviously this is an instant winner so is worth 1.00 of the players bet (as it pays 1/1 and assuming dealer does not have a 'Blackjack').
Looking at 10,6 we get :-
Dealers Upcard 2 3 4 5 6 7 8 9 10 A
Value of 10.6 -0.44 -0.34 -0.30 -0.26 -0.21 -0.43 -0.48 -0.53 -0.56 -0.56
Adding the ev to the 'Blackjack' ev we get :-
BJ + 10,6 EV 0.56 0.66 0.70 0.74 0.79 0.57 0.52 0.47 0.44 0.44
Looking at the ev's for 10,10 and A,6 we get :-
Value of 10,10 0.48 0.55 0.56 0.58 0.59 0.71 0.73 0.70 0.55 0.55
Value of A,6 -0.15 -0.07 -0.03 0.01 0.07 0.00 -0.12 -0.19 -0.24 -0.26
10,10 + A,6 EV 0.33 0.48 0.53 0.59 0.66 0.71 0.61 0.51 0.31 0.29
Looking at the total EV for both hands against each dealer upcard :-
Dealer Upcard (A,10 + 10,6) (10,10 + A,6) Correct Decision
2 0.56 0.33 NO SWITCH
3 0.66 0.48 NO SWITCH
4 0.70 0.53 NO SWITCH
5 0.74 0.59 NO SWITCH
6 0.79 0.66 NO SWITCH
7 0.57 0.71 SWITCH
8 0.52 0.61 SWITCH
9 0.47 0.51 SWITCH
10 0.44 0.31 NO SWITCH
A 0.44 0.29 NO SWITCH
So, the correct play is to 'Switch' against a dealer 7, 8 and 9 and to stay against the other upcards which is what the simulator computed. A dealer 9 upcard is a close decision and I would imagine that high counts may affect the decision although there is a considerable gap in ev in the 'switch' decisions against a dealer 4, 5 and 6.
