Basic Strategy Philosophy???

[ChefJJ]

I am new to these message boards, but not necessarily new to blackjack or basic strategy. Hell, in fact, I have taken a little bit of time to do my own Excel tables to come to figure out the dealer's probabilities of drawing/busting--so I have a good understanding that the ol' "ten in the hole" assumption is garbage.

What I have come to the group about is this question:

When basic strategy is determined, what "philosophy" or philosophies are used to determine what the player should do vs. a particular dealer upcard? For instance, do you look for situations where the player bust % is less than the dealer? What I am saying is, I understand the probabilities of everything and the strategy table, but what criteria are used to determine what action should be taken?

Take, for instance, a player having 13. You have only about a 38% chance of busting, whereas the dealer (in an infinite deck situation) with a 4 showing has about a 39% chance of busting. Why would you not hit?

Thanks in advance for the insight.

 

[Automatic Monkey]

...Take, for instance, a player having 13. You have only about a 38% chance of busting, whereas the dealer (in an infinite deck situation) with a 4 showing has about a 39% chance of busting. Why would you not hit?...

Because if you both bust, only you lose. And you will both bust around 15% of the time. Remember, you have to go first

 

[ChefJJ]

Good point. BS says to hit 12 vs. a dealer 2 and 3 (31% player bust vs. 35% and 37% chance the dealer busts), but not 4 (39%). Why is the line drawn there? What criteria is used for the decision of what the player should do vs. the dealer's upcard?

If you think about it strictly in terms of who will bust, the player has just less than a 70% chance of drawing safely (not necessarily 17 or above) with a 12. But the dealer really never has more than a 42% chance of busting, and that is with a 6 up. But one would look really crazy to hit a 12 vs. dealer 6, right????

 

[MGP]

The conventional strategies are determined by looking at the maximum EV for a given hand. This is determined by simply looking at (WinEV-LoseEV) for each hand. It is slightly different than either a max-win strategy or a max ev/bet strategy. It is different from the latter because when you double/split you don't discount the fact that you have double the bet or more out on the table.

 

[ChefJJ]

Interesting. What do you mean by EV? And is there some sort of "personal" tolerance built into it, like aggressiveness?

 

[MGP]

EV = expected value = WinEV - LoseEV

It's just as simple as that - pure math - no personality...

 

[ChefJJ]

How does that then apply in the example I set out? 12 vs. 2 and 3, but not 4? If there is no personality, then why does some BS say 12 vs. 2, but not 3?

How does one calculate expected value? I appreciate the terms, but can you tell me how you use those terms?

Thanks

 

[MGP]

How does that then apply in the example I set out? 12 vs. 2 and 3, but not 4? If there is no personality, then why does some BS say 12 vs. 2, but not 3?

How does one calculate expected value? I appreciate the terms, but can you tell me how you use those terms?

Thanks

Many posted strategies are simply incorrect. Even here and on the Wizard of Odds site there are incorrect strategies/EV's because they are based on estimations or simulations for splits or unusual rules. Don't forget too that the startegies can differ a lot depending on the rule set.

I'm not sure how else to explain how to calculate EV - it's simply the WinEV-LoseEV.

For example - if you have 21, and the dealer has a 6. You're EV is 1-p(Dealer 21) because you win every hand except against a dealer 21 which is a push.

If you had a total 16 or less - you're EV is simply (DealerBustProb - (1-DealerBustProb)) because you'd lose everytime except when the dealer busts.

You should check out Eric Farmer's BJCA program since you said you wrote your own macro - it shows you how it's done. Splits are the only estimate he has for SPL>1. You can find it at bjmath.com

 

[halcyon1234]

I'm sure I'll be corrected if I'm wrong about this, but I believe the philosophy behind making the bs tables come down to the "best of all possible worlds".

Someone will make a computer program that will evaluate every possible hand versus every possible dealer upcard. When a hand is evaluated against an upcard, it does something like this:

What could the dealer have in its hole card?
For every possibility
Does it result in a win?
Does it result in a lose?
Does it result in a push?
If not, what if the dealer hit?
For every possible card the dealer could draw, does it win, lose, push, if not, draw another.

And so on. So in the example of 12 vs. 2, if you were to take a deck and play out every possibile of every combination of cards you could draw/stand on vs. every possible comination of cards the dealer could draw/bust with, you would find the following:

If the player stands, there are more situations which the dealer will win with.
If the player hits, there are more situations which the player will win with.

That is also the reason why different bs tables exist for different numbers of decks. In a single deck game, if you have T2222, there's no way to draw to a 20 with a single hit anymore. But if you were playing 6 decks, the possibilities are different. And those differences might mean that there are more wins than loses in certain situations.

Of course, this all assumes a fresh deck. The odds change dramatically as the deck depletes. After all, if there were no 10s left in the deck, you couldn't bust hitting a 12. But that's why card counting is for. Since it wouldn't be humanly possible to memorize a table for every combination of cards that has left the shoe, you just keep track of an idea of what the shoe's like, and modify bs based on it.

So that's the philosophy behind bs charts.

 

[tedloc]

BS Chart

Good point. BS says to hit 12 vs. a dealer 2 and 3 (31% player bust vs. 35% and 37% chance the dealer busts), but not 4 (39%). Why is the line drawn there? What criteria is used for the decision of what the player should do vs. the dealer's upcard?

If you think about it strictly in terms of who will bust, the player has just less than a 70% chance of drawing safely (not necessarily 17 or above) with a 12. But the dealer really never has more than a 42% chance of busting, and that is with a 6 up. But one would look really crazy to hit a 12 vs. dealer 6, right????

I don't know if you know this but the BS charts, that are generated ,are only based on 3 cards out of a full deck. Once the cards come out of the deck ,strategy changes. If you had a chart that constantly updated itself, you would see many of the don't hit hands, change to HIT. Thus card counting is the only way to take full advantage. As to your question about hitting 12 vs dealer four. Even though BS says NO, it is a very close call. Renzy says to BS players, "if you can see 5 or more 10's on the table, you might want to hit it" Let me give you an example, that shows a definite change in BS. Everyone knows, not to hit your 12 vs dealer 6. What if all the 10's are gone from the deck. An up to date chart would say to hit.

 

[MGP]

I don't know if you know this but the BS charts, that are generated ,are only based on 3 cards out of a full deck. Once the cards come out of the deck ,strategy changes. If you had a chart that constantly updated itself, you would see many of the don't hit hands, change to HIT. Thus card counting is the only way to take full advantage. As to your question about hitting 12 vs dealer four. Even though BS says NO, it is a very close call. Renzy says to BS players, "if you can see 5 or more 10's on the table, you might want to hit it" Let me give you an example, that shows a definite change in BS. Everyone knows, not to hit your 12 vs dealer 6. What if all the 10's are gone from the deck. An up to date chart would say to hit.

That's not completely true. Many charts are but mine and a few others are not. When I give a TD strategy it is based on a weighted avergae all pre-split hands with a given total. The strategies are however based on the top-of the deck. In addition, if a fixed number of rounds are played than the EV is the same for every round on average.

 

[ChefJJ]

I thank you all for the great insight on this issue. Basically, the BS chart all depends on outcome predictions--and from the sounds of it, most all of the BS versions are similar with some slight variations. Where personal preference such as aggressiveness comes in to play may be where marginal doubles or splits are called for/not called for. I look forward to looking at the situations with Excel, which is tedious, but after Christmas.

The discussion in place really emphasizes the importance of card counting (tracking the deck) and paying attention to what is going on with other cards on the table.

Thanks
Merry Christmas