House edge vs 'perfect' player?

I was wondering if anyone has calculated what the house edge is vs a 'perfect' blackjack player, given normal casino rules / decks / penetration.

To elaborate on what I mean by perfect-
Basically what would the house edge be vs a super-computer (or someone science-fictionally smart like Hollywood's Rainman), where each of the player's decisions would be optimal vs the exact remaining deck composition.

Semi related question -
Can a real player who is only ever flat betting be a favorite over the house, given real world constraints like using a humanly possible counting system, not literally cheating by using a computer, etc?

How close can a real player get to 'perfect'?

Quote: tldr

Semi related question -
Can a real player who is only ever flat betting be a favorite over the house, given real world constraints like using a humanly possible counting system, not literally cheating by using a computer, etc?

Yes, provided only hands with a positive expected value (+EV) are played. Of course sitting out the vast majority of hands would get attention. Playing reasonably -EV hands and increasing bets on +EV hands is the winning balance when card counting.

Edit: Defined abbreviation since OP is a new member.

Quote: BleedingChipsSlowly

Yes, provided only hands with a positive expected value (+EV) are played. Of course sitting out the vast majority of hands would get attention. Playing reasonably -EV hands and increasing bets on +EV hands is the winning balance when card counting.

Edit: Defined abbreviation since OP is a new member.

I don't consider wonging in and out to be flat-betting. You are spreading from 0 to 1.

The answer to this question depends on the rules. In another thread a game in france that was dealt from a CSM and had a players edge off the top was mentioned, so that can be beaten with just basic strategy (the edge was very small). Even if a game is slightly -EV off the top, if it's close enough to 0 then playing variations could bring it to +EV. You might not be able to find a game with rules this good, but in theory it is possible. Your best bet would be a deeply-dealt single-deck game with good rules, if you can find one.

It does depend upon the method you use, Hi/Lo, Zen, Omega 2, and when you enter and leave the game. I can't answer your question.
But I do know this... even with Ace-5 and a wong- in/out at 0, you will profit at a dealer stands 17 game... in the LONG, long, run just like any other method you choose. So once one starts "Advantage Play" the advantage is in your favor albeit slim. And slim's holdin that ticket outta town.

deleted

Like I said, it does depend on what method of raising/retreating is used. Certain methods are more advantageous than others, I think Uston's count is the best of all availible, so go with that. And I think that early play is not a good idea, one wants say 1/2 the shoe gone to accurately predict composition. If you're playing the composition game, history is everything, but don't wait till the last hand. There is a "certainty" that has to be evaluated. IMHO with at least a deck missing from the shoe burned, certainty is low at any time, but gets better with history... the trick is at what pen do you jump in, and at what score. I don't think the regulars like me can predict that. A computer might say +3 hi/lo with 1 deck burned and one remaining, or Zen +4, or Uston APC at +7.

Quote: tldr

I was wondering if anyone has calculated what the house edge is vs a 'perfect' blackjack player, given normal casino rules / decks / penetration.

To elaborate on what I mean by perfect-
Basically what would the house edge be vs a super-computer (or someone science-fictionally smart like Hollywood's Rainman), where each of the player's decisions would be optimal vs the exact remaining deck composition.


How close can a real player get to 'perfect'?

To the best of my knowledge with flat-betting, computer-perfect play will decrease the house edge by about 1%, giving you a small advantage in a typical game. Oddly, betting correlation will decrease--I can't say why, but it probably has something to do with the way it is calculated, and the non-linearity of the perfect play type of multi-parameter counting. Professional cheaters use computers and bet spreads to achieve large advantages with perfect play and shuffle analysis. A human counter can beat any game by flat-betting by simply sitting out a large portion of the hands.

Knowing the exact composition of the stub is the most useful near the very end with deep penetration. I doubt you could get an edge at 2 decks or more.

Quote: anonimuss

Knowing the exact composition of the stub is the most useful near the very end with deep penetration. I doubt you could get an edge at 2 decks or more.

I just read that with good rules and 75% penetration, at single deck you would get 3%. At 2D, you would get 1%, and at multideck, you would still be at a disadvantage. So, perhaps you are more right than I was. Thank you for catching that.

Quote: Sonuvabish

I just read that with good rules and 75% penetration, at single deck you would get 3%. At 2D, you would get 1%, and at multideck, you would still be at a disadvantage. So, perhaps you are more right than I was. Thank you for catching that.

That's interesting. Maxim in Vegas years ago had a single deck game with good rules plus late surrender. It was slightly positive off the top and it was pre Rule of Six days so depending on the dealer they went well beyond 50% pen once in a while. They ran out of cards twice one night when I was playing there. There was another casino out near Henderson that had a game that I believe was slightly positive off the top also. There was an airplane hanging from the ceiling in one of the restaurants and it was a nice, good sized place but I just can't remember the name. One table, a card counter at every seat every time I was there. I walked in there one day and there were two guys from my home area on the game.

I use a more advanced count and swear by it. My playing decisions in real time at the tables are nearly as accurate as if you were punching it into a combinatorial analysis program. I enjoy the precision and besides that you need every little sliver of what you can get... there is only the tiniest bit of an edge for an AP that is strictly counting to begin with. It's been said (by Zender, I think) that for your typical Hi-Lo counter to be any threat to the casino, they need to be using at least a 1-12 spread, so if you wong in TC+2 or more flat betting over a long enough period of time you'll make enough to buy you a cup of coffee at Starbucks. Why is this? When the count is high enough you want more money out there for those serious EV doubles and snappers, so I hope this is all just a curious theorization rather than anyone thinking you're going to fare all that well flat betting no matter what the PE or how advanced your counting method is.

http://www.youtube.com/watch?v=LNhyDPno_Kk

Quote: anonimuss

That's interesting. Maxim in Vegas years ago had a single deck game with good rules plus late surrender. It was slightly positive off the top and it was pre Rule of Six days so depending on the dealer they went well beyond 50% pen once in a while. They ran out of cards twice one night when I was playing there. There was another casino out near Henderson that had a game that I believe was slightly positive off the top also. There was an airplane hanging from the ceiling in one of the restaurants and it was a nice, good sized place but I just can't remember the name. One table, a card counter at every seat every time I was there. I walked in there one day and there were two guys from my home area on the game.

You could jump from table minimum to table maximum. We may have crossed paths.

Tarzan's here?

Nice!