BJ PERFECT PLAY COMPUTER

BJ PERFECT PLAY COMPUTER
Just wondering if any of you out there in cyber land can help me out. I would like to know what effect on the house edge of a typical bj vegas rules being

for example : 8 deck game double any 2 cards stand s17 double after split no resplit aces :

for a computer which plays a strategy not according to the count but according to the exact composition of the deck remaining.
( Perfect play knowing the exact cards removed.)

I am interested in various penetrations but lets say 25% 50% 75 % 85% of the cards dealt to the player.

Is this big enough to remove the house edge at a certain penetration level? If so what is it?


Thank you

Salthouse
DELETE
EDIT
LEAVE A COMMENT

Replied by PM. The thing is that, like counting, perfect composition play depends on wonging or bet ramping. It would also need sufficient bankroll and Kelly betting techniques to ensure a profit. With those aspects in play you would have a card counting machine on steroids. Would, of course be considered cheating in most jurisdictions and with an online casino would be easily dectectable and result in account closure with forfeiture of bankroll.

Salthouse,

Welcome to the forum. Your blog post is identical to this initial thread, and since you want to start a discussion, is more appropriate here, so I'm deleting your blog. Please don't make duplicate posts, including blog/thread creation. Thanks!

This discussion addresses computer-perfect play (changing Hit vs Stand vs Double vs split decisions) without bet variation and wonging out.

The concept of computer perfect play has been discussed several times before. I think the answer we came up with was that the advantage of computer-perfect decisions is probably less than 1% in edge unless you have extreme penetration.

The reason is that for a high fraction of all BJ hands the strategy decision is not very close. Like: 10-10 v 10 you stand, its not close; 10-9 v 6 you stand its not close, etc.

In order for computer-perfect play to make a difference:
- the hand you are playing must have a decision that is reasonably close
- the remaining cards in the deck must have a composition that is skewed in a direction that makes it more profitable to make the play that is not basic strategy for the hand (that is only the case 50% of the time)
- and the remaining card composition must be skewed sufficiently that making the non-standard decision is actually profitable (and significant profits occur only when the card composition is extremely skewed.) That will almost never happen in the early part of the shoe and is usually uncommon in the latter part of the shoe.

Look at it this way: how many hands do you think that you might be able to play differently because of your computer-perfect play? I would say that 10% is a generous amount.

And, on average, how much of an advantage over the normal basic strategy would you be able to achieve when you make a different play. I would guess less than 5%, on average, but let's say 10% to be generous.

So how much of an edge are you gaining? Well, according to the above assumptions, its an average of 10% on about 10% of the hands. Which is 1%.

Now, if you were to get to an extreme penetration and there are 10-20 cards left, then it would probably make a much larger difference - at that point.