Best Strategy with Perfect Knowledge of Cards Played

[SiMi]

Hi, all
Can anyone help me find a resource that details how to determine the best BJ strategy if one had complete, perfect knowledge of the cards that have already been seen in a 6 deck (or other) shoe? I remember Taft built a BJ computer that, I think, did this but I don't know how he developed the strategy/algorithm. The stories about Taft make me think that such a strategy exists...
I wrote a BJ play simulator (like CVCX) years ago and use it almost daily for training. I've gotten bored over the years of using it and thought it would be interesting to modify the program so it can advise me what the best play would be IF I HAD KEPT TRACK OF EVERY CARD THAT'S ALREADY BEEN PLAYED rather than just keeping a HiLo count. I already have it keeping track of the number of Aces and 5's that have been played just for my information, if I want to see that at any point but, I'd like to try to understand what strategy would be best if I knew this information for ALL the cards that have been played.
For example, if there were only 30 cards left in a 6 deck shoe (deep cut) and all the 6's and 7's have already been played but there's an excess of 10-value cards and 9's remaining, and you have a 14 vs. a dealer's 10, should you hit? Stuff like that.
I did some web searching but couldn't find anything exactly on point. Griffin's TOB has a chart on p.28 that shows how much can be gained from "perfect play" based on the number of unseen cards. (When the number of unseen cards gets small, the gain is impressive.) He also refers to the idea of 'perfect play' a lot in the book but I guess I'm missing the big picture of how one could design an algorithm that produced the best strategy given the seen vs. unseen cards. The task seems Herculean, but very interesting.

Thanks in advance for any help and BEST to all!
SiMi

[Letangs]

While I understand the curiosity out of such informations, in practice that wouldn't be remotely possible to take advantage over these informations except if the player has some kind of gift like Eidetic Memory... Which is the case of what ? 0,0001% of the population ?

Also, that reminds me of the device that Ken Uston had played with, hidden in his shoes, that constantly gave him the best play to do in regards of every card that has been shown previously. These stats must exist somewhere if it was already available 40 years ago.

Subside question : Are there any professional counters that we know of, that had Eidetic Memory ? (Except Rain Man of course hahaha )

[ummagumma]

Here is an combinatorial analyser that should be able to help: http://www.bjstrat.net/cgi-bin/cdca.cgi

I think somewhere on his site there is a downloadable which you could include in a device if you were clever enough. PM me if you are...

[SiMi]

Hi, UG
Thanks! I'm looking at it right now.

Best,
SiMi

[Dbs6582]

While I understand the curiosity out of such informations, in practice that wouldn't be remotely possible to take advantage over these informations except if the player has some kind of gift like Eidetic Memory... Which is the case of what ? 0,0001% of the population ?

Also, that reminds me of the device that Ken Uston had played with, hidden in his shoes, that constantly gave him the best play to do in regards of every card that has been shown previously. These stats must exist somewhere if it was already available 40 years ago.

Subside question : Are there any professional counters that we know of, that had Eidetic Memory ? (Except Rain Man of course hahaha )

Wow, that’s quite an in depth question and comments. You even knew about Ken Uston and the device he played with for a while. You sound like a student of the game. Why are you only wanting to learn KO. I’d think you'd at least want to learn Hilo.

Are you sure you didn’t know how to read basic stategy for soft and hard hands?

[DSchles]

Hi, all
Can anyone help me find a resource that details how to determine the best BJ strategy if one had complete, perfect knowledge of the cards that have already been seen in a 6 deck (or other) shoe? I remember Taft built a BJ computer that, I think, did this but I don't know how he developed the strategy/algorithm. The stories about Taft make me think that such a strategy exists...
I wrote a BJ play simulator (like CVCX) years ago and use it almost daily for training. I've gotten bored over the years of using it and thought it would be interesting to modify the program so it can advise me what the best play would be IF I HAD KEPT TRACK OF EVERY CARD THAT'S ALREADY BEEN PLAYED rather than just keeping a HiLo count. I already have it keeping track of the number of Aces and 5's that have been played just for my information, if I want to see that at any point but, I'd like to try to understand what strategy would be best if I knew this information for ALL the cards that have been played.
For example, if there were only 30 cards left in a 6 deck shoe (deep cut) and all the 6's and 7's have already been played but there's an excess of 10-value cards and 9's remaining, and you have a 14 vs. a dealer's 10, should you hit? Stuff like that.
I did some web searching but couldn't find anything exactly on point. Griffin's TOB has a chart on p.28 that shows how much can be gained from "perfect play" based on the number of unseen cards. (When the number of unseen cards gets small, the gain is impressive.) He also refers to the idea of 'perfect play' a lot in the book but I guess I'm missing the big picture of how one could design an algorithm that produced the best strategy given the seen vs. unseen cards. The task seems Herculean, but very interesting.

Thanks in advance for any help and BEST to all!
SiMi

Last year, there was an extensive discussion on this in which Eric Farmer and I, among others, participated. Search for it.

Don

[Letangs]

Wow, that’s quite an in depth question and comments. You even knew about Ken Uston and the device he played with for a while. You sound like a student of the game. Why are you only wanting to learn KO. I’d think you'd at least want to learn Hilo.

Are you sure you didn’t know how to read basic stategy for soft and hard hands?

As I said, I did read maybe 10 to 12 Blackjack books, so I kinda know about the story of the BJ golden age. I have always been fascinated by card counters and that’s why I did read books. Squeezed most of the charts tho. But I never had the motivation to actually learn systems to put it in practice, or even learn perfect BS. But I recently changed my motivations and I want to try to master card counting.

My question about BS might have sounded stupid to you, but multi-cards soft hands didn’t make much sense to me, that’s why I wanted to ensure I understood things correctly.
my lack of knowledge regarding the « playing » has nothing to do with the fact that I know about Uston’s device, as I read Million Dollars Blackjack...

as for KO vs HILO, I don’t feel I have the capacity to make true count conversion.. I want to master an easier system before switching to something complex... the ultimate goal is to play in a Casino, if I have a system that I’m not feeling comfortable with, I might just save my time and do something else ...

[SiMi]

Hi, Don
Thanks for the help. If I found the right threads ('Computer Perfect Play' and 'Card by card EOR and strategy'), looks like Eric's BJ7 has the CA algorithm needed to do what I asked about. I downloaded it and I'm looking at it now and it is very cool.
It also seems as if you were working on this issue late last year and it fizzled out? Also, it looks as if you were expecting the result to be very disappointing.
Despite that, as an intellectual exercise, I am willing to keep working on this problem because CA is an area I would like to understand better. Griffin's book has always been a real stretch for me and I would really like to understand this topic better.

Best to you,
SiMi

[Dbs6582]

As I said, I did read maybe 10 to 12 Blackjack books, so I kinda know about the story of the BJ golden age. I have always been fascinated by card counters and that’s why I did read books. Squeezed most of the charts tho. But I never had the motivation to actually learn systems to put it in practice, or even learn perfect BS. But I recently changed my motivations and I want to try to master card counting.

My question about BS might have sounded stupid to you, but multi-cards soft hands didn’t make much sense to me, that’s why I wanted to ensure I understood things correctly.
my lack of knowledge regarding the « playing » has nothing to do with the fact that I know about Uston’s device, as I read Million Dollars Blackjack...

as for KO vs HILO, I don’t feel I have the capacity to make true count conversion.. I want to master an easier system before switching to something complex... the ultimate goal is to play in a Casino, if I have a system that I’m not feeling comfortable with, I might just save my time and do something else ...

Makes sense. Thanks for the clarification. Million Dollar Blackjack was the first bj book I read many years ago. A friend gave it to me. I still didn’t get into blackjack after I read it since I didn’t live near any casinos during that time. But it did peak my interest so when I moved to an area closer to casinos, I spent more time researching and reading books on bj.

I’ve looked at the KO system, but I thought it was easier to go with Hilo. If you’re going to spend anytime counting, I think it’s best to go with Hilo since the majority use it, it’s easier to understand and people will be able to help you more. It’s not as hard as it seems. If you’re this much a student of the game, I’d think you’d want to spend the extra time learning it. Actually, the KO system seems harder to me and didn’t make as much intuitive sense. But I know sims have shown KO works and gets the money.

Its your call though. Just giving you a little advice based on what I think you can handle. Good luck with whatever you choose. By the way, you have to develop a thick skin here, but overall it’s a good group. I’ve enjoed it and leaned a lot.

[DSchles]

Hi, Don
Thanks for the help. If I found the right threads ('Computer Perfect Play' and 'Card by card EOR and strategy'), looks like Eric's BJ7 has the CA algorithm needed to do what I asked about. I downloaded it and I'm looking at it now and it is very cool.
It also seems as if you were working on this issue late last year and it fizzled out? Also, it looks as if you were expecting the result to be very disappointing.
Despite that, as an intellectual exercise, I am willing to keep working on this problem because CA is an area I would like to understand better. Griffin's book has always been a real stretch for me and I would really like to understand this topic better.

Best to you,
SiMi

Glad you found it. The results were pretty much in line with what I expected -- not mind-boggling, dazzling improvements, but obviously superior to anything else that can be done. I wanted to write it all up and perhaps add a chapter or new appendix to the new printing of BJA3, which I think would have been great and would have publicized Eric's fine work, but he wasn't amenable, so I dropped it.

Don

[Letangs]

Don, do you have any reminder of any professional counters that we know of, that had Eidetic Memory and therefore could use this kind of statistics ?

[Three]

Don, do you have any reminder of any professional counters that we know of, that had Eidetic Memory and therefore could use this kind of statistics ?

You don't just need to know the cards played, which is hard enough, you also need to know how to use the information. Does the term head exploding mean anything to you?

Back in the old SD days before the books on counting I knew people that did exactly that. They could play about 45 minutes and would be mentally exhausted and drenched in sweat and need a break.

[Three]

They used to have brain teasers about the last few cards from the all the cards dealt in SD to see if people were good enough to use the knowledge of the last handful of cards to make the right decisions to maximize EV for a hand matchup. I don't see them anymore since the days of that kind of penetration are long gone.