Don Schlesinger: Just what is "Basic Strategy"?

[Don Schlesinger]

> I was also surprised to find A2 v 5 listed by Wong as "hit." (I'm looking at the chart in "Blackjack Secrets.")

I've explained this quite often. Wong ran sims to determine his BS and then stuck by them, no matter what answer they gave. He knows that hitting is the "wrong" play, but didn't want to violate his own algorithm for such a borderline decision.

But, he's taken a lot of criticism for having knowingly supplied the wrong play. In my view, he should have changed it to double and then furnished an explanation, in a footnote, that his computer actually said hit. Instead, he looks needlessly "wrong" in the eyes of many readers.

In any event, there is no one, official "generic" basic strategy. Anyone can come up with one. Indeed, I've seen a few decent books, each with a different "one BS fits all" generic flavor.

And, of course, the moment you introduce either counter's basic strategy or risk-averse BS into the equation, you complicate matters even more, which is why I didn't even bring them up.

Each is not unique, even with a given set of rules. Each depends on the bet scheme used by the player, and so, trying to agree is quite hopeless.

Don

[Cyrus]

I couldn't just get up and leave for some reason , and that damn shoe would just never end !

[ET Fan]

In any event, there is no one, official "generic" basic strategy. Anyone can come up with one. Indeed, I've seen a few decent books, each with a different "one BS fits all" generic flavor.

Exactly why infinite deck basic is useful as common ground. It's easy to calculate, there's no disagreement on what it is, it solves all the ambiguities per pair splitting and all the rest, there's only one inf deck basic for a given rule set, and it's a perfectly adequate, accurate, and practical basic for attacking the shoes -- which you've always said are the games professionals are drawn to.

And, of course, the moment you introduce either counter's basic strategy or risk-averse BS into the equation, you complicate matters even more, which is why I didn't even bring them up.

I only mentioned them to illustrate that infinite deck basic is as viable and practical as any other. I'm not suggesting for a moment that counter's basic or r-a basic be chosen as THE basic.

Each depends on the bet scheme used by the player, and so, trying to agree is quite hopeless.

The plays I mentioned don't depend on the bet scheme.

ETF

[Don Schlesinger]

> I couldn't just get up and leave for some
> reason, and that damn shoe would just
> never end!

Nice to see you back, Cyrus. Why so damn long between visits?

I can't warm up to infinite-deck anything. None of the statistics have ever interested me. If i-d BS closely resembles correct multi-deck BS, then I'll learn the latter, thank you, and be 100% correct, instead of being almost correct.

It's not as if you get any prize for learning the i-d variety. It's basically useless.

Don

[ET Fan]

I can't warm up to infinite-deck anything. None of the statistics have ever interested me. If i-d BS closely resembles correct multi-deck BS, then I'll learn the latter, thank you, and be 100% correct, instead of being almost correct.

You can't do that because there's more than one (unless you choose to learn 27-d basic, which is identical to i-d basic, as well as 28-d basic, 29-d basic, and so on)! In fact, you will have four different versions to learn for 2-d through 8-d.

I must have missed the definitive answer to your original question: "Your answers to any or all of the above questions are eagerly solicited. In short, just what would you want us researchers and authors to call `basic strategy'?"

I presented a simple, elegant solution, and you can't warm up to it because, what, you're stuck at the table? Pretty silly, no?

ETF

[Cyrus]

> I would change that a little: "given
> knowledge of nothing more than the dealer's
> upcard and the first two cards!. After
> splitting you play each hand split as a new
> hand comprised of two cards against the same
> dealer's upcard" (I know it can be
> worded better.)

I'll try to word it just a little better, if I may.

Every hand played in Basic Strategy assumes that it is played after a shuffle, ie off the top of the pack.

Suppose we get a pair and we split it, by following BS, then we play out the first split hand. To play the second split hand, if we want to follow Basic Strategy instead of something else, we must assume that the second split hand is actually the first hand dealt off the top.

[Don Schlesinger]

> Suppose we get a pair and we split it, by
> following BS, then we play out the first
> split hand. To play the second split hand,
> if we want to follow Basic Strategy instead
> of something else, we must assume that the
> second split hand is actually the first hand
> dealt off the top.

Understood. But, apparently, when playing the first hand, we must also assume that we don't know that the other half of the split is lying to the left of us, and that bothers me a little bit, because that card belongs to my original hand (which, I understand, doesn't exist anymore)!

Don

[Richard Reid]

> I'll try to word it just a little better, if
> I may.

> Every hand played in Basic Strategy assumes
> that it is played after a shuffle, ie off
> the top of the pack.

> Suppose we get a pair and we split it, by
> following BS, then we play out the first
> split hand. To play the second split hand,
> if we want to follow Basic Strategy instead
> of something else, we must assume that the
> second split hand is actually the first hand
> dealt off the top.

I like your definition, Cyrus. This allows us to look at the cards we have in our current hand and to make a decision based on only these cards.

When dealing with splits, we look at the orginal hand first and decide to make a play based on it being played off the top of the deck. If we decide to split, then we look at the first of the split hands in isolation as if it now is being played off the top of the deck, and we make our play. Then we look at the next hand of the split hands in isolation as if it too were being played off the top of the deck, with no knowledge of the previous split hand that was just played. In other words, we treat each hand of a split hand individually with no knowledge of the other split hand(s).

I do like your definition. It allows for us to treat 7,7 vs 10 in single deck in isolation from the other 14's should we go as far as two card composition basic.

So, that being said, with your definition, the decision to go with total-dependant or composition-dependant basic is still an option.

In other words, "Do we treat 9,3 vs 4 different than T,2 vs 4, and different than 5,2,5 vs 4, etc. or do we treat them all as 12 vs 4?

Sincerely,
Richard Reid

[Don Schlesinger]

I'm delighted to see you posting here, Richard. I think readers should understand that, were it not for the fact that you have your own, wonderful site, I certainly would have invited you to be one of our Masters. As it is, your presence here will always be both welcomed and appreciated.

Although I grudgingly give in to this way of reckoning pairs, there is something always tugging at me that is bothersome. We define BS as the correct play off the top of the pack, because, as such, we really do have just three exposed cards to reckon, when we make that "correct" play.

However, when we have a pair, we need to make believe that we're seeing only three cards, when, in fact, there are clearly four (for the play of the first hand) lying on the table. You can make believe all you want that that other card isn't there, but a) it is! and b) it changes the EV of the hand you're about to play as if that split card didn't exist. And that, to me, makes the situation different and, therefore, worthy of discussion.

Don

[Cyrus]

"That being said, the decision to go with total-dependant or composition-dependant basic is still an option.

In other words, Do we treat 9,3 vs 4 different than T,2 vs 4, and different than 5,2,5 vs 4, etc. or do we treat them all as 12 vs 4?"

IMHO, we cannot properly define a BS that's applicable for every number of decks and/or every rule. A generic BS can be formulated, of course, and has been, but the inherent assumptions are too much for me. Way I see it, for every specific set of rules and number of decks, we must have a single, specific BS.

The direction given by Griffin, that (paraphrasing) no information available should be discarded, is unassailable. Accordingly, formulating a BS that takes into account everything we can legitimately see in front of us, card by card, until the round is over, is the theoretically correct BS. Full composition-dependent Basic Strategy, in other words. (I'm saying "theoretically" because even ETF would agree that this is just not practical. Practical BS for me is total-dependent, with the right exceptions allowed.)

The threshold between Basic Strategy, which worries not abt (dis)advantage, and constantly, a priori re-computing our advantage, aka card counting, is easy to define, isn't it? In the former we do not take into account the previous rounds dealt, if any, while in the latter we do. (That's why the former dictates flat betting. Because we "are" always on the 1st round, while the latter relates the level of betting to the level of every round's advantage.)

I don't think there can be a definition of BS that would accomodate the information provided by the previous rounds. (CBS is a hybrid : the player follows the same playing strategy, as if always on the 1st round, but takes into account the info from the previous rounds in order to bet.)

[Cacarulo]

> However, when we have a pair, we need to
> make believe that we're seeing only three
> cards, when, in fact, there are clearly four
> (for the play of the first hand) lying on
> the table. You can make believe all you want
> that that other card isn't there, but a) it
> is! and b) it changes the EV of the hand
> you're about to play as if that split card
> didn't exist. And that, to me, makes the
> situation different and, therefore, worthy
> of discussion.

When you have your two cards in front of you against the dealer's upcard you have to make a decision based only on that info. Once you make that decision (stand, hit, double, split or surrender) you're done. What to do next? Well, you can follow a t-d strategy OR in the case of splitting you can use again a 2-card composition decision for each of the pairs.
Of course, we can generate a more accurate strategy for splitting. For example say we have: 8,8 vs 9
We can get the strategy for:

8,A vs 9 (given that an 8,8,9 have been removed)
8,2 vs 9 (")
........
8,T vs 9 (")

But, how many strategies are you going to learn? Is it worth learning them?

Sincerely,
Cacarulo

[Cacarulo]

> In any event, there is no one, official
> "generic" basic strategy. Anyone
> can come up with one. Indeed, I've seen a
> few decent books, each with a different
> "one BS fits all" generic flavor.
> Exactly why infinite deck basic is useful as
> common ground. It's easy to calculate,
> there's no disagreement on what it is, it
> solves all the ambiguities per pair
> splitting and all the rest, there's only
> one inf deck basic for a given rule set,
> and it's a perfectly adequate, accurate, and
> practical basic for attacking the shoes --
> which you've always said are the games
> professionals are drawn to.

I agree with you about using i-d for devising a "generic" strategy. If someone wanted to learn BJ I would recommend him/her studying this "generic" strategy first. Then he/she should have to learn the exceptions for the particular games.
However, I don't think we can define BS as a "generic" strategy. I think that what we are discussing here is about which definition fits best .

Sincerely,
Cacarulo

[ET Fan]

As I understand it, the "zero memory" pair splitting you, and many experts, embrace is zero memory for the strategy employed, but NOT for the EV -- which takes all card removals into account.

It just occurred to me (maybe not new to you) that the process of determining strategy (and hence, EV) could theoretically be just about IMPOSSIBLE, for some rule sets and # of decks.

The problem is that the decision to hit or stand on any group of cards is based on EV, and that EV should include EVs on that same group of cards while splitting under "zero memory." Eg. Suppose the CA program decides 7-2-3 v 4, for some # decks and some rule set, should be HIT, but then LATER, while analysing splitting 7s it turns out that the cumulative EV, after you include all those cases, favors STAND. But still LATER, while analysing splitting 2s, EV swings in favor of HIT again. And then STILL LATER, while splitting 3s. EV swings in favor of STAND.

So for EVERY c-d hit/stand decision, you must first fully analyse all splits which could result in those decisions! Or so it seems to me.

But, BEFORE YOU DECIDE WHETHER A SPLIT IS BS, don't you need all the EVs of all the resulting hit/stand decisions? So you determine these EVs, then decide to split or not FOR EACH CARD in the c-d group, then recalculate the EVs, then maybe reverse your hit/stand decision, then re-decide whether to split, and on and on ... recursive times three for 3 card groups, recursive times 4 for four card groups, etc. Just about UNIMAGINABLE. And this for the SIMPLEST model we have -- the zero memory model.

In many (most?) cases, the EVs would be so far apart, the program could HAZARD A GUESS that split analysis will not change a hit/stand decision, but that seems like anathema to the whole spirit of a CA program. And there's always a chance some new rule could throw the whole thing into a tizzy, making BS so close, and so unstable, it's literally intractable. As intractable as the solution to chess.

Or do we just throw up our hands and say pair splitting BS is zero-memory for EVs as well as strategy?

I remember you discussed "recursive" pair splitting algorithms with Steve on bjmath, but is this what you meant?

ETF