Baccarat AP play

[johndoe]

As I recall, perfect nonlinear counting (knowledge of all subsets) has been thoroughly analyzed, and the prospects were still poor; the limitation doesn't seem to be with linear count systems, the limitation seems to be the lack of enough profitable opportunities to make the game more than marginally beatable. And that's with perfect knowledge.

Sure, there are some deck compositions that provide a large advantage, but they seem to be so rare as to not be worthwhile. I've yet to see any evidence to the contrary, but would be happy to be proven wrong.

 

[zengrifter]

As I recall, perfect nonlinear counting (knowledge of all subsets) has been thoroughly analyzed, and the prospects were still poor; the limitation doesn't seem to be with linear count systems, the limitation seems to be the lack of enough profitable opportunities to make the game more than marginally beatable. And that's with perfect knowledge.

Sure, there are some deck compositions that provide a large advantage, but they seem to be so rare as to not be worthwhile. I've yet to see any evidence to the contrary, but would be happy to be proven wrong.

That was Thorp's criticism - 'not worth the effort', as I recall. zg

 

[GBV]

That was Thorp's criticism - 'not worth the effort', as I recall. zg

Thorp has never published anything concerned with nonlinear counting, that quote concerns a linear system.

 

[Automatic Monkey]

That was Thorp's criticism - 'not worth the effort', as I recall. zg

On the Wizard of Odds site they discuss a baccarat sidebet called 4-5-6 where you bet on the combined number of cards used in the hand. It was rumored to be offered at the Atlantic City Hilton.

I almost spent myself as I was doing the numbers on that one! Unfortunately, the sidebet doesn't exist.

 

[johndoe]

Thorp has never published anything concerned with nonlinear counting, that quote concerns a linear system.

It might have been Griffin.

Here's the nonlinear (perfect count) analysis:

http://www.bjrnet.com/member/archive/baccarat_hall.html

Given perfect nonlinear counting, there's an earn opportunity of $16.30/hr for a $1 million bankroll.

Where is the flaw in this study?

 

[GBV]

It might have been Griffin.

Here's the nonlinear (perfect count) analysis:

http://www.bjrnet.com/member/archive/baccarat_hall.html

Given perfect nonlinear counting, there's an earn opportunity of $16.30/hr for a $1 million bankroll.

Where is the flaw in this study?

No flaw, but you can find much better than 25-card penetration.

Ironically Michael Hall was the first guy I ever discussed the possibility of beating baccarat with, he was a very talented and successful advantage gambler.
I'm breaking a confidence here but this was fifteen years ago and so far as I know he is completely retired from gambling altogether, if even still alive. The report you link to was a heavily censored version, the uncensored version has much the same data (at least on counting) but is much more optimistic about potential returns as it acknowledges much deeper penetration exists.

Hall was well aware that the game could be much more profitable provided there was deeper penetration and even suggested a partnership at one stage, since I was well placed Europe to attack much better games. He knew that he could train himself to estimate advantage remainders accurately and develop a heuristic.
In the end I think he took the money he made from casino advantage play, which was a lot, and vanished. He reappeared after a few years at twoplustwo with some interesting work on poker but appeared to have completely lost interest in house-banked casino games, and then vanished again, this time permanently.

 

[moo321]

Can we describe all of the potentially profitable situations? I know all tens, all evens, or 3 ranks would all help the tie.

Once we describe them, how frequently do they occur, and what are they worth? Basic frequency distribution.

And is there better than 10 card pen available? It seems like baccarat should be quite beatable if dealt to, say 6 or 4 cards.

 

[GBV]

Can we describe all of the potentially profitable situations? I know all tens, all evens, or 3 ranks would all help the tie.

Once we describe them, how frequently do they occur, and what are they worth? Basic frequency distribution.

And is there better than 10 card pen available? It seems like baccarat should be quite beatable if dealt to, say 6 or 4 cards.

Calculating the frequency of a given end-deck subset can be done by calculating the chance of those cards appearing from a full deck, eg for all-tens there are 128 ten-valued cards in an eight-deck pack, so 128/416*127*415...

The problem is that while favorable subsets are rare are very large numbers of permutations with all but very small card subsets, so you need a heuristic to whittle the subsets down to manageable proportions.

Additionally, and this is where it gets very complicated, you need to work out a good method of backcounting. It is frequently the case that if a lot of tens and even cards are eliminated it saves time to go and play a fresh pack since your chance of encountering a favorable subset is worse than it is off-the-top, even adjusting for the fewer hands that need to be played and/or shuffle time.
This is absolutely crucial, and something many of those who have studied this area fail to understand. It is insane to just sit there through every shoe regardless of the card distribution. You need to maximize the frequency with which you encounter favorable subsets.

I have played against packs that were dealt down to the last six cards. In some cases you can achieve similar penetration by asking to see the burn cards, because some cards are exposed during the shuffle, or one of half a dozen other tricks. The vast majority of games are not dealt down to this level deliberately.

 

[Dipsy]

I saw some interesting baccarat side bets at 188bet.com's live casino. Namely:

Player Odd Pays 0.96:1
Player Even Pays 0.90:1
Banker Odd pays 0.94:1
Banker Even Pays 0.94:1

BIG (5 or 6 cards dealt) pays 0.54:1
SMALL (4 cards dealt- naturals) pays 1.5:1

are they beatable?

if the penetration is CRAP - 60% is it beatable? if the dealer is not attractive, would it make it even harder to beat?

 

[GBV]

I saw some interesting baccarat side bets at 188bet.com's live casino. Namely:

Player Odd Pays 0.96:1
Player Even Pays 0.90:1
Banker Odd pays 0.94:1
Banker Even Pays 0.94:1

BIG (5 or 6 cards dealt) pays 0.54:1
SMALL (4 cards dealt- naturals) pays 1.5:1

are they beatable?

if the penetration is CRAP - 60% is it beatable? if the dealer is not attractive, would it make it even harder to beat?

Even without any formal analysis 60% is clearly way too little. Those bets are clearly pretty symmetrical when it comes to effects of removal. You can construct hypothetical examples which would create an advantage, but these would only occur at deep penetration.

 

[FLASH1296]

Fuggedabouit. 95% is not sufficient.

And just for the sake of clarity —

You said: “SMALL (4 cards dealt- naturals) pays 1.5:1“

You omitted something -- no draw is possible when there is a 7 vs. 6

 

[Sucker]

I have played against packs that were dealt down to the last six cards. In some cases you can achieve similar penetration by asking to see the burn cards, because some cards are exposed during the shuffle, or one of half a dozen other tricks. The vast majority of games are not dealt down to this level deliberately.

There are NO baccarat games that are deliberately dealt down to the last six cards. If this were true, then you would be able to frequently obtain advantages of up to 15%, just by looking in the back of the shoe & knowing the last card in the shoe.

 

[GBV]

There are NO baccarat games that are deliberately dealt down to the last six cards. If this were true, then you would be able to frequently obtain advantages of up to 15%, just by looking in the back of the shoe & knowing the last card in the shoe.

Imagine that. That "up to 15%" edge on the last cards would be a huge addition to the up to 500% advantages you would often encounter at that level without seeing the back card.

 

[Dipsy]

Fuggedabouit. 95% is not sufficient.

And just for the sake of clarity —

You said: “SMALL (4 cards dealt- naturals) pays 1.5:1“

You omitted something -- no draw is possible when there is a 7 vs. 6

thanks for the clarification. I'm not very experienced at baccarat myself. cheers.

 

[moo321]

Even without any formal analysis 60% is clearly way too little. Those bets are clearly pretty symmetrical when it comes to effects of removal. You can construct hypothetical examples which would create an advantage, but these would only occur at deep penetration.

The evens vs. odds would be quite assymetrical. If there are lots of evens, bet even, if not, bet odd, ignore tens.

Although I agree that formal analysis is probably a waste of time with 60% pen.

 

[GBV]

The evens vs. odds would be quite assymetrical. If there are lots of evens, bet even, if not, bet odd, ignore tens.

That won't work. You need the deck structure to be almost completely composed of odd or even cards to get an edge. A smaller excess of odd or even cards makes almost no difference.

As a practical matter you would never see an edge on this bet with that pen.

 

[Egalite]

6 deck dealt to last 7 cards minimum

There are NO baccarat games that are deliberately dealt down to the last six cards. If this were true, then you would be able to frequently obtain advantages of up to 15%, just by looking in the back of the shoe & knowing the last card in the shoe.

Imagine that. That "up to 15%" edge on the last cards would be a huge addition to the up to 500% advantages you would often encounter at that level without seeing the back card.

In the UK 6 decks are used at the Baccarat tables, one particular casino chain burns no cards and all cards are dealt down to the the last 7 at a minimum. It can be less than 7 if the cut card comes out during the last hand, occasionally the cut card is the the next card in the tray after the current hand, this results in virtually all cards (exception is one card) may be dealt for the last hand.

The downside is, if the shoe turns out unfavourable mid-stream, there is no other table, to jump to, strictly a one table operation. I am of the opinion tracking for favourable "last hand sub-sets" would not present themselves often enough to be worthy of serious consideration. But am open to suggestions.

 

[moo321]

In the UK 6 decks are used at the Baccarat tables, one particular casino chain burns no cards and all cards are dealt down to the the last 7 at a minimum. It can be less than 7 if the cut card comes out during the last hand, occasionally the cut card is the the next card in the tray after the current hand, this results in virtually all cards (exception is one card) may be dealt for the last hand.

The downside is, if the shoe turns out unfavourable mid-stream, there is no other table, to jump to, strictly a one table operation. I am of the opinion tracking for favourable "last hand sub-sets" would not present themselves often enough to be worthy of serious consideration. But am open to suggestions.

James Grosjean agrees with you. I'm not totally sure...

 

[Sucker]

In the UK 6 decks are used at the Baccarat tables, one particular casino chain burns no cards and all cards are dealt down to the the last 7 at a minimum. It can be less than 7 if the cut card comes out during the last hand, occasionally the cut card is the the next card in the tray after the current hand, this results in virtually all cards (exception is one card) may be dealt for the last hand.

am open to suggestions.

At this casino the cut card is inserted 7 cards from the end? If this really IS so; then after you have made 10 posts to this site, you will be allowed to send me a P.M.; and I may be willing to send a reply with some of the suggestions you're seeking.

 

[Egalite]

Ideal conditions, but what's the critera?

At this casino the cut card is inserted 7 cards from the end? If this really IS so; then after you have made 10 posts to this site, you will be allowed to send me a P.M.; and I may be willing to send a reply with some of the suggestions you're seeking.

I'm getting a bit confused here, can't I PM until I make 10 posts? Yes I assure you this is the case, he cut card is placed 7 cards from the end.. The last hand will always overlap the last 7 cards.

This is a Blackjack site, I no longer play BJ. I play Baccarat and mainly post at baccaratforums, I have over 500 posts there, and over 4000 posts at gamblersglen under another ID. It would be hard for me to make 10 posts on this site, which is dedicated to a game I no longer play.

I gave BJ up due to one rough day many year ago. I've tried my hand at counting (Hi-Lo and KO). Most of the BJ games in the UK are dealt from CSM's, were the cards sometimes get returned to the machine after one round. I wouldn't go near them with your money .

So all of my action is at the Bacc' or Punto-Banco tables as they are called over here. I literally can't believe the opportunity that exists for a viable counting method if such a beast exists.

I can reply to this thread 10 ten times if you like