Soft doubling -- BS doesn't make sense? ??

[Renzey]

BS for H17, 6D, DAS:
Why wouldn't an A,2 or A,3 be a double against a 4? With an A,2, the only cards that actually help you are 4, 5, 6, 7, or 8. A total of 5 cards. With an A,3, the only cards that help you are 3, 4, 5, 6, 7. Again, total of 5 cards.
Now, with an A,4, the only cards that actually help you are 2, 3, 4, 5, or 6 -- again, 5 cards.
By looking at the 3 hands (A,2 / A,3 / A,4) we realize they are all the same hands, because there are only 5 cards that help each of them.
So, what causes the difference in basic strategy?

Others here have alluded to the reason why you should double a hand like A/4 vs 4, but not A/2 vs 4, even though just 5 cards help either hand. The determinant is that with the A/2, doubling loses more "playing latitude" by restricting yourself to one hit, since twice as many hit cards will leave you wanting a second hit (A/2/3/4). Thus, there's more playing efficiency available with the A/2 by reserving the right to take extra hits -- enough so, that hitting the A/2 brings a more favorable result than doubling.

[ZenMaster_Flash]

Kudos to Renzey for a perfect clarification.

[8675309]

I also have a question about the A2 vs 5 hand....in BJA3 on page 425, table A32...the basic strategy EV shows it is ever so slightly favorable to double the hand in the dd h17 game....but in BJA3 on page 520, table D15 for h17, the dd m2 value is -0.2992 in the EOR table which would indicate it is better to not double, correct? The mean for single deck shows a favorable double but every other mean shows negative values, so am I looking at something wrong or does this information contradict itself as I think? I know that A2 vs 5 is not a strong play by any means but did not expect this contradiction....thanks

[RS]

Ah, thanks guys, What Renzey said makes perfect sense. The result of hitting A,2 will result in either a soft 14-17, which you'd want to hit again. Anything else (S18+ or H12, H13) you wouldn't want to hit. But, if you have something like A,6 and hit, you're not going to want to take another hit, because either you'll have a S18+ or an H12-H17, both of which you'd stay.

Grosjean put it best when he wrote, "BS is something more to be memorized than understood... There is nothing to understand!" Players are always wanting to understand it, but all that one needs to know is that each BS decision "produces a higher expectation than any other option."

I agree...to an extent. Like many (or all?) things in blackjack, nothing needs to be quite understood, but known or memorized. You don't need to understand why you have an advantage at a +4 TC...although it definitely helps why. Plus, I like to know why I'm doing what I'm doing. To me, memorizing some chart and not having a clue why I'm doing what I'm doing seems a bit....wrong or foolish.

[NT21]

I can understand that. My approach is just different. I have little interest in the theory of why a hand is played a certain way. Of course, usually there's common sense that can dictate why. But it really doesn't interest me. I'm more concerned with developing my game, knowing my index numbers, getting quicker at TC conversions, perfecting my craft, and maximizing my spread. To me, needing to know the "why" for a BS play is like a basketball player needing to know that the diameter of a hoop is 18" and the diameter of the ball is 9", and that two basketballs can fit thru the hoop at once. That's all well and good, but it doesn't make them a better player.

[Renzey]

To me, needing to know the "why" for a BS play is like a basketball player needing to know that the diameter of a hoop is 18" and the diameter of the ball is 9", and that two basketballs can fit thru the hoop at once. That's all well and good, but it doesn't make them a better player.

Sometime, somewhere, somehow, knowing the whys and wherefores of BS and other game technicalities will come in handy and enable you to make an advantageous "outside-the-box" move. One example that arises from time-to-time would be when the player next to you doubles for half with 9 against a 7 at a neutral count. Should you take the other half?
Or maybe that player has 7/7 against a 3, but doesn't have any money left. Should you partner with him on a split?
How about when you have 17 against a 5 on a misdeal, and the floorperson asks if you want in or out of the hand?
Etc.

[RS]

Yes

No

Get out of the hand

Was thinkin' the same thing. Is this stuff you just know, or are there charts for it?

[PinkChip]

Yes, I would participate in the doubling of 9 versus 7 cos I assume the EV of this play is positive (probability of winning the hand is > 50 percent). The fact that Basic Strategy recommends hitting rather than doubling is irrelevant here because both plays have positive EV, doubling is just a bit less "lucrative" than hitting with neutral count, if count is high then you should double anyway. And if the play is sub-optimal, it's the player's problem and not mine :-) I remember that Fred Renzey has described such examples in Bluebook II, am I right?

(7,7) versus 3 is probably a defensive split (Standing and Splitting are both negative expectation plays), so I would not participate in that since I am not obliged to - it's not my hand to play.

Regarding 17 versus 5, I am not sure if this constellation has positive or negative EV for me. 17 is a mediocre hand (better than a stiff hand, no bust danger, but far from 21). Dealer 5 is good for me, his bust probability is 42 percent, so my probability of winning the hand should be greater than 42 percent but the question is if it is greater than 50 percent. I would not know the answer by heart.

[Renzey]

Yes, No and Get Out: Was thinkin' the same thing. Is this stuff you just know, or are there charts for it?

Like the A/2 vs 4 situation, it's next to impossible to "just know". The knowledge comes from understanding what's behind the EV charts, once you've seen and thought about them.
All three of those plays would usually be made wrong by a basic strategy purist.
And that same player would scoff at somebody who stood with 7/5/4 against a 10, even if right off the top of the shoe.

[Renzey]

Yes, I would participate in the doubling of 9 versus 7 cos I assume the EV of this play is positive.
(7,7) versus 3 is probably a defensive split, so I would not participate in that. I remember that Fred Renzey has described such examples in Bluebook II, am I right?
Regarding 17 versus 5, I am not sure if this constellation has positive or negative EV for me.

9 vs 7 wins 59% of the time when hit, and 53% of the time when doubled. As a bystander of a double for half, you can jump in on a 53% shot, or do nothing.
7/7 vs 3 wins 48% of the time on each 7 when split (incl. das), but only 37% of the time if stood on. Although splitting is better than standing, both plays are underdogs.
All this is derived from the EV charts, such as the ones in Wong's Professional Blackjack. Yes, Bluebook II has charts naming improper doubles which still carry a net advantage, and proper splits that carry a net disadvantage.

Regarding 17 vs. 5, the EV charts show it to be an 11-to-10 underdog (counting pushes as a half win and half loss).

[roliin]

What am I missing here? Why double 9 vs 7 at a neutral count when the index # is 3????? Is it something to do with only doubling for half?

[PinkChip]

Regarding 17 vs. 5, the EV charts show it to be an 11-to-10 underdog (counting pushes as a half win and half loss).

Yes, last night I though over this and realised I had overlooked that with my 17, I can only win if the dealer busts. But my overall winning probability is nevertheless more than 42 percent because I have to take into account the pushes (the cases the dealer also ends up with 17). I ran my self-written program to compute the dealer result probabilities and it says the dealer having a 5 upcard will end up with 17 about 12 percent of the time.

To be more precise, he will get 17 about 12.18 percent of the time and bust 41.84 percent of the time (6 decks, heads up play, S17), or get 17 about 11.81 percent and bust about 41.96 percent (6 decks, heads up play, H17). Which makes sense, because when hitting Soft 17, the dealer will bust more often and end up with 17 less often because he stands only with the Hard 17s, not the soft ones.

But 12 percent for pushes are too little because the push probability would require more than 16 percent for a positive EV for me (I would win 42 percent when dealer busts, push 16 percent and lose 42 percent, for an EV of zero). Rather than that, I would win 42 percent, push 12 percent, and lose 46 percent, so my EV would be negative. So the odds are (46 + 6) : (42 + 6) = 52:48 = 13:12 against me, which is pretty close to your 11:10 (maybe my numbers rounded to whole percents were a bit inaccurate).

[PinkChip]

What am I missing here? Why double 9 vs 7 at a neutral count when the index # is 3????? Is it something to do with only doubling for half?

The point is, for the player who hesitates to double down, the correct decision (according to Basic Strategy) would be to hit rather than double down. This is because both plays have a positive expectation, but to hit gains more on average (in the long run). But the player (erroneously from his perspective) wants to double down but has too little money, so he tries to double for less (for half in this case).

Now that the player has decided, the question for you is just, should you participate in this double play and lay down the rest of the money till it is a doubling for the full amount. The answer is yes, because from your perspective, the double down play is still advantageous (it will win more than 50 percent of the time, yielding a positive expectation for the double down wager, thus also for your share). Basic Strategy is optimal for the player who makes the playing decisions, but not necessarily for the bystander who was the option to bet on, but not to play the hand.