Please help to provide Basic Strategy

[Gronbog]

Please help to generate 8 deck BS for peek and no peek. Thanks

I will generate the strategy for you for NOPEEK assuming you lose all bets to a dealer blackjack (ENHC). You already have the post-split strategy for PEEK. The pre-split strategy for PEEK does not change from the normal basic strategy.

Sorry -- I forgot about this.

I have generated the post-split strategy for aces for NOPEEK. Here it is:

http://gronbog.org/results/blackjack....S17.ENHC.html

Some notes:

• Compared to the PEEK strategy, all of the doubles for less vs T and A are still doubles for less, however the critical fractions have changed.
• A,2 through A,5 vs T and A with PEEK were double for any amount, but for as little as possible. These hands now have a critical fraction. With NOPEEK, you must be able to double for less than the critical fraction of your original bet, otherwise you should stand. The critical fractions are now displayed directly in the table for convenience.
• The critical fractions for A,6 vs T and A are now lower, as expected.

[refinery]

Sorry -- I forgot about this.

I have generated the post-split strategy for aces for NOPEEK. Here it is:

http://gronbog.org/results/blackjack....S17.ENHC.html

Some notes:

• Compared to the PEEK strategy, all of the doubles for less vs T and A are still doubles for less, however the critical fractions have changed.
• A,2 through A,5 vs T and A with PEEK were double for any amount, but for as little as possible. These hands now have a critical fraction. With NOPEEK, you must be able to double for less than the critical fraction of your original bet, otherwise you should stand. The critical fractions are now displayed directly in the table for convenience.
• The critical fractions for A,6 vs T and A are now lower, as expected.

Your math skills are impressive.

[Gronbog]

<sarcasm>Wow. That's twice in one day</sarcasm>

[BughouseMASTER]

OP: why the hell are you playing at a place you cant double A/A or split or can only double for "little" or only "hit" like gronbog is saying..... find a REAL casino!

[James989]

Thanks for your hard work

[Gronbog]

Bughouse, you've missed the premise. This is a casino that lets you either double or stand after you have split aces. This is a small improvement over most casinos where you must stand after splitting aces. This is the strategy for deciding whether to double or stand and, if doubling, for how much.

[MGP]

You can use my BJCA to do those rules. Use a Forced Strategy on the CD soft/pairs tab and apply it to only post-split hands.


https://www.blackjackinfo.com/free-b...rial-analyzer/

[Phoebe]

I take this opportunity to thank you for this excellent program that I often use and recommend.

[NB10]

You can use my BJCA to do those rules. Use a Forced Strategy on the CD soft/pairs tab and apply it to only post-split hands.


https://www.blackjackinfo.com/free-b...rial-analyzer/

How does this program differ in terms of accuracy versus something that uses simulations to derive the correct strategy?

[Phoebe]

Accuracy is perfect.

[NB10]

Accuracy is perfect.

So better than using simulations, or the same/no real difference?

[Gronbog]

Comparing CA-generated strategies to simulation-generated ones:

• The strategy decisions generated by the simulation will be correct, provided that it has been run long enough (and mine has for the strategies posted here).
• The underlying EV's will become more accurate the longer the sim is run, but will never be accurate to the same number of decimal places as the CA. The EVs in the strategies I posted are accurate to 2 decimal places.
• For simple things like Basic Strategy, the sim takes much, much longer than the CA to resolve to the correct strategy.
• For complex things like index generation, only a simulator is practical.

The reason I use simulation to generate strategies is that a lot of my work revolves around multi-player tournament situations for which the CA becomes impractical. All of the CAs that I am aware of make compromises just to be able to handle splitting, let alone multiple players.

[lij45o6]

How does this program differ in terms of accuracy versus something that uses simulations to derive the correct strategy?

As Gronbog has stated: A Combinatorial Analysis of the game performs statistical/probabilistic computations for exact expectation.. That is, it will compute things like the expectation of the hand [T T] vs dealer [T] with a deck subset of say...: {0 0 0 0 0 0 0 0 4 0}. That is, what is our Expectation for standing, hitting, doubling, and in some cases splitting and what is the correct action based on those expectations. As per our example: based on the given deck subset, the *ONLY* correct action is to stand and our Expectation is 0.0000 with a probability of 1.00. Throw in an 9 : {0 0 0 0 0 0 0 1 4 0} and our expectation changes:

STAND: +0.500
HIT: -1.000
DOUBLE: -2.000
P([T T] & [T]): 0.400

CA's can also perform correct computations faster than that of simulators, especially Monte Carlo methods. Coupling that with correct values makes a CA a valuable tool when analysing the game of 21.


To get a better understanding of CA's and of what "goes on under the hood," I would recommend the Theory of Blackjack by Peter Griffin. 6th Edition. There Dr Griffin goes into details about the knitty-gritty maths that are involved with analysing 21. Much of what you know about card counting can be found in the book. However, it is not going to teach you how to count better, just how to do maths in the game.

Another book: Blackjack Attack by Don Schlesinger. 3rd Edition.