-----------------

The other B.S. play I have seen most ignored is hitting a soft eighteen against a nineteen or twenty. This is another play where it's hard to see the advantage on it's face. Some people also don't like to double eleven against a ten. I can't really blame them for that, since they are lessening their variance without giving up too much of the house edge.

If the question is what play are recreational players most likely to not make correctly, I would say 44 vs. 5 or 6, when double after a split is allowed. Even dealers often roll their eyes when I split them, and other players shake their heads, as if to say "what an idiot."

Quote:luckyjackgI doubled on a soft twenty against a dealer's 5 and was vilified by the dealer and anothor player. I thought it was the right play. It was a double deck game with the count +5. I lost the hand but the dealer and the player at the table thought I was nuts. Was I?

Probably. I'm not sure of the index number for that play but it is not a basic strategy play. That is similar to splitting tens.

Quote:KellynbnfFor some reason of all the parts of Blackjack Basic Strategy the thing that it seems that players are most reluctant to do is split a pair of 8s against a high dealer upcard (i.e. 9, 10, A). Often they say that they don't want to turn one losing hand into two losing ones, but the math shows that splitting results in an average lower loss than playing the hand unsplit. Does anyone know why people often think this way? Any other parts of BS you have heard people "contest"?

As you correctly pointed out, any time you start with one hand with a negative expectation and BS requires you to put out twice as much money (either by doubling or splitting) just to end up risking twice the money with a negative expectation. You have to trust the mathematics of all the possible outcomes rather than how it feels to you. But even the mathematics does not show a great improvement with the BS play.

It is easy for people to understand splitting 8's against a dealer 7. You start with a single hand with negative expectation, and end up with two hands both with positive expectation. I refer to this case as the greatest possible improvement in blackjack play, because the correct play (splitting) has the greatest increase in EV over the other three choices (stand, hit, double). It's a subjective name, because I am ruling out bonehead plays like hitting 20's.

Novice players want plays that feel like the greatest possible improvement play. Of course, if every choice was that way you wouldn't need BS. The reason you memorize BS, is how to choose the better play where both possible choices are giving you a negative expectation, and you want the least worst choice.

==========

I didn't see the movie 21, but there was a line where the MIT professor says that splitting 8's against a dealer 10 was a sucker bet. It may have been a set up line in the movie, because presumably if you are writing the script for a movie about blackjack, you would go to a store and spend $3 on a cheat card.

Quote:WizardIn terms of lost expected value, I think the greatest mistakes are made not doubling soft doubles against weak dealer cards. Seldom do you see recreational players do that. If the player never invokes correct soft doubles it increases the house edge by 0.09%.

If the question is what play are recreational players most likely to not make correctly, I would say 44 vs. 5 or 6, when double after a split is allowed. Even dealers often roll their eyes when I split them, and other players shake their heads, as if to say "what an idiot."

44 against 5 or 6 is a tough one especially at 3rd base. You draw a couple of 10s to stand on two 14s and you'll have the table shaking their fist at you when the dealer pulls an ten and then draws a 5 to make 20 or 21. You just gotta do the math though and stick to your game plan. I remind people when a questionable (yet correct) play has a happy ending, and jostle them back just as much as they mess with me.

Comparing to the Wizard's 2-deck H17 chart, I would not be making the following plays:

double 9 vs. 2

double 11 vs. A

double A3 vs. 4

double A7 vs. 2

double A8 vs. 6

split 44 vs. 5 (if DAS)

split 44 vs. 6 (if DAS)

split 66 vs. 7 (if DAS)

split 77 vs. 8 (if DAS)

I've gotta assume that these are very very marginal plays. But if I'm gonna use a chart, I may as well use the right one. What is the edge on the above plays?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

That's an estimate but I bet it's not too far off.

Quote:MelmanI've gotta assume that these are very very marginal plays. But if I'm gonna use a chart, I may as well use the right one. What is the edge on the above plays?

The Wizard calculate all the EV for every compositionally dependent play in Appendix 9 of the Blackjack section of the Wizard of Odds site.

You have to multiply them together using the probability calculation as well to know the true difference in EV. For example if a player has a sum of 9, the wizard calculates a line for 3,6 and for 4,5. You have to do the weighted average.

But yes, generally the plays where the decision alters because of small rule changes are marginal improvements in EV.

Quote:teddysThose are all borderline plays. I would say doing/not doing those plays would swing the house edge about .05%.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

That's an estimate but I bet it's not too far off.[/A]

According to this article (http://www.deepnettech.com/article1.html), it is .005% for the H17 vs. S17 rules... I use the H17 rules for either game, its not worth the added variance to add those doubles.

Quote:DrEntropyAccording to this article (http://www.deepnettech.com/article1.html), it is .005% for the H17 vs. S17 rules... I use the H17 rules for either game, its not worth the added variance to add those doubles.

Do you mean that you use the S17 rules? H17 has more soft doubles (A6 or A7 vs. 2; A8 vs. 6).

And as compared to the generic "LV multi-deck" chart in my book, Wizard has extra splits for DAS (2's or 3's vs. 2 or 3; 4's vs. 5 or 6; 6's vs. 7; 7's vs. 8). And some "single-deck" plays he uses with 2-deck but not with 4+ decks (split 6's vs. 2, double 9 vs. 2 and 11 vs. A).

Quote:HawksedI did the same thing at Hard Rock in Tampa and was dealt a two. The player to my left scornfully attributed my perfect hand to dumb luck after a stupid play. I probably won't do it again even if basic strategy dictates that this is the correct play. Assuming that a soft 19 vs. dealer 6 only comes along about every 300 to 400 hands, and there is only about a one percent advantage to doubling down vs. standing, it seems that the effect on the house edge is extremely small. And I am risking another $10 bet each time. I'm no math wizard but won't I have to bet an additional $30,000 or so over the long haul to net another $10.

Choices where you hit instead of stand where the EV is very close are not as difficult since you are not risking more money. It is difficult to double your bet in pursuit of a very small increase in EV. My friend calls it accepting a smaller EV in exchange for minimizing your risk of ruin.

Quote:pacomartinAs you correctly pointed out, any time you start with one hand with a negative expectation and BS requires you to put out twice as much money (either by doubling or splitting) just to end up risking twice the money with a negative expectation.

You are correct except about the part of BS having you double in a negative situation.

Quote:alternativeinvestorYou are correct except about the part of BS having you double in a negative situation.

There are no negative-EV situations where you'd double (in such cases hitting is always preferable to doubling, even if you'd be taking only one card such as 12 vs. 2 or 3).

The reason there are times you'd split in a negative-EV case is because with splitting you change your starting point (and in the case of a pair of 8s that's helpful since 16 is the worst possible total). That is also why 5s and 10s should never be split (in those cases you end up with a worse starting point than you were at before splitting).