On Marginal Even Money Decisions

[Renzey]

You're playing a six deck shoe. It looks like 3.25 decks are in the discard tray -- maybe closer to 3.5 with the extra cards on board. Your Hi/Lo RC is +7. That makes the TC about +2.6 -- maybe 2.75.
You're playing two multi-unit hands and have 14 and blackjack. The dealer has an Ace up. EV-wise, you should blow off Insurance and Even Money. Variance-wise, I dunno. Taking either loses about three-quarters percent of several betting units. If it's an RA variance saver to take the Even Money, then what about Insuring the 14?

Who would make the "ploppy-look-alike" play of taking the Even Money while just gambling with the 14? Who would hold the variance down as much as possible by Insuring both hands? And who feels that variance be damned -- don't make a negative EV bet?

 

[Renzey]

You're playing a six deck shoe. It looks like 3.25 decks are in the discard tray -- maybe closer to 3.5 with the extra cards on board. Your Hi/Lo RC is +7. That makes the TC about +2.6 -- maybe 2.75.
You're playing two multi-unit hands and have 14 and blackjack. The dealer has an Ace up. EV-wise, you should blow off Insurance and Even Money. Variance-wise, I dunno. Taking either loses about three-quarters percent of several betting units. If it's an RA variance saver to take the Even Money, then what about Insuring the 14?

Who would make the "ploppy-look-alike" play of taking the Even Money while just gambling with the 14? Who would hold the variance down as much as possible by Insuring both hands? And who feels that variance be damned -- don't make a negative EV bet?

So Sorry! Can we please move this post over into the card counting section?

 

[Renzey]

That makes the TC about +2.6 -- maybe 2.75.
You're playing two multi-unit hands and have 14 and blackjack. The dealer has an Ace up. If it's an RA variance saver to take the Even Money, then what about Insuring the 14?

Is Insuring a 14 at +2.7 TC a variance reducer at all? There's still a pretty big window thru which to lose the Insurance bet and the hand too!

 

[zengrifter]

Depending on the health of my BR I would either pass on the insurance or only insure the BJ.
What am I missing? zg

 

[blackjack avenger]

Depends

You did not specify the cards of the 14. If the 14 did not contain a 10 I would insure. One can use the composition of the hand for these marginal calls.

:joker::whip:
good cards

 

[SleightOfHand]

Is Insuring a 14 at +2.7 TC a variance reducer at all? There's still a pretty big window thru which to lose the Insurance bet and the hand too!

HAND | EV w/o Insurance | Variance w/o Insurance | EV w/ Insurance | Variance w/ Insurance
14 | -0.61 | 0.57 | -0.65 | 0.69
19 | -0.11 | 0.86 | -0.15 | 0.52
20 | +0.15 | 0.85 | 0.11 | 0.27
21 | +1.04 | 0.48 | 1.00 | 0.00

(Taken from ECAA; I hope the author wont mind)

So it looks like insuring 14 will increase your variance for a lower EV. Don't do it. :laugh:

Personally, I insure 20s and BJs if the TC is <1/4 TC (zen) from the index.

 

[blackjack avenger]

Not the Full Story

HAND | EV w/o Insurance | Variance w/o Insurance | EV w/ Insurance | Variance w/ Insurance
14 | -0.61 | 0.57 | -0.65 | 0.69
19 | -0.11 | 0.86 | -0.15 | 0.52
20 | +0.15 | 0.85 | 0.11 | 0.27
21 | +1.04 | 0.48 | 1.00 | 0.00

(Taken from ECAA; I hope the author wont mind)

So it looks like insuring 14 will increase your variance for a lower EV. Don't do it. :laugh:

Personally, I insure 20s and BJs if the TC is <1/4 TC (zen) from the index.

One can still look to hand composition
:joker::whip:
good cards

 

[SleightOfHand]

One can still look to hand composition
:joker::whip:
good cards

Oh bja, why must we disagree so much these days? What will hand composition help with here? We are counting, so we already have the count of the deck including whats in your hand, which in this example still does not exceed the insurance index. "What am I missing?"

 

[Blue Efficacy]

You did not specify the cards of the 14. If the 14 did not contain a 10 I would insure. One can use the composition of the hand for these marginal calls.

:joker::whip:
good cards

Composition is irrelevant if you're already counting.

 

[Renzey]

You did not specify the cards of the 14. If the 14 did not contain a 10 I would insure.

Don't remember. But in either case, whether I held 10/4, 9/5, 8/6 or 7/7, it was the 10's or non-10's I held that brought the RC and resulting TC to +2.7 -- or whatever it was. If I had held A/3, it could be a different story, since the Ace and 10 are interchangeable in producing the same RC, yet one kills a 10 while the other does not.

This occurs with 16 vs. 10 at times when the RC reaches zero by virtue of the cards in your 16. If it got to zero with 8/4/4, I would stand. If it got there with 6/6/A/3, I would hit. The 6's and 4's are interchangeable in producing the same RC, but the two dead 4's reduce my "outs" while the two dead 6's increase them. This is an inherent weakness for mere humans in grouping different card ranks together with the same value. If a pentium chip in your head could track all ranks separately, your playing efficiency would be appreciably enhanced.

 

[Renzey]

HAND | EV w/o Insurance | Variance w/o Insurance | EV w/ Insurance | Variance w/ Insurance
14 | -0.61 | 0.57 | -0.65 | 0.69
19 | -0.11 | 0.86 | -0.15 | 0.52
20 | +0.15 | 0.85 | 0.11 | 0.27
21 | +1.04 | 0.48 | 1.00 | 0.00


So it looks like insuring 14 will increase your variance for a lower EV. Don't do it.

Sleight, it looks like those figure are at a neutral count. Still though, it doesn't seem as though the situation would improve for the 14 at a higher count, unless the 14's EV drops more slowly than the dealer's likelihood of a 10 in the hole goes up. This was basically the input I was looking for. Thanks much!

 

[blackjack avenger]

A Small Voice Standing on the Shoulder's of Giants

A,A
A,10
10,10

All are -2 hands in hi lo; correct?, yet they would shift the insurance indice. So while not running the exact math of Renzy's example. Depending on the hand composition the indices would be different.

For just hi lo ins is 3?
For just halves ins 3.33?
Both are correct, depends on what you are considering.

With more info, like the composition of the hand, things can change. It's not worth much but also not hard to implement. It is interesting to only insure one hand and have it be the correct play. Like not insuring your 20 but insuring your 3,2 or is it vice versa? hmmmmm

For generic insurance what others have said is correct, you have counted your hand and that is all you need to know.

Renzy is bringing up a marginal play, so I am pointing out some marginal considerations. I think we need more posts like this.

No one should take anything I write to heart. I like the dialogue.

Here is another one; fairly easy to see, 12 v 2 at the margin. If a 10 in hand hit, if not stand. The reason your 10 in hand is one less that will break you and one less that will break the dealer so you better try to improve your hand. If you don't have the 10 in hand then that is one more 10 floating out there that could break you or the dealer so you better stand. It's not worth much.

good cards
:joker::whip:

or is it a blowhard standing on a soapbox lol

 

[blackjack avenger]

Not Blowing

I believe the RA for insurance is a higher indice and can be composition dependent.
:joker::whip:
good cards

 

[Renzey]

And another thing!

Sleight, this was basically the input I was looking for. Thanks much!

BTW, what's so integral about reducing variance on one hand -- vs. reducing variance between two consecutive hands -- or on two hands a half hour apart for that matter?? It's not as though this is the only hand you're going to be playing. Isn't there some inherent flaw in compartmentalizing results on an endless journey? Shouldn't EV reign supreme in your decision? It's not as though you're trading stocks on margin over your head and need to hedge.

The two-hand Insurance scenario I described has happened to all AP's many, many times over the long haul. If you've always eschewed the Even Money, getting burned as you often will, I still suppose you've saved money over the years by doing it that way.

 

[London Colin]

If insurance for less is available, but only even money (i.e. full insurance) can can be taken on a blackjack, then insuring the 14 for a lesser amount could be preferable to taking even money on the blackjack.

That's assuming it is actually possible to figure out the desired amount of insurance to take. I know there is a formula in TOBJ for how much insurance to take on a BJ, based on the precise proportion of tens remaining and the fraction of the bankroll that has been bet. But is anything known about how (or whether it's feasible) to apply the same approach to a count and betting ramp, in a practical way?

(I speak as one who lives in a country where insurance is not normally available, apart from even money, so it's very much an academic question from my point of view. )

 

[blackjack avenger]

Greek, Not to Me

for hi lo
A,A right at insurance indice, do you take it?
10,10 right at insurance indice, do you take it?
Yep, the cards are counted.

first one you would
second one you wouldn't

:joker::whip:
good cards

 

[aslan]

Yesterday, I was playing 6D, S17 NS $25 min. The count was +4 RC KO (the recommended insurance point in KO is +3 RC). I was playing two hands at $200 each and had a twenty and a stiff, and the dealer had an A. I took the insurance and the dealer had blackjack.

The next hand, same bet, the count was +2 RC KO and I had two stiffs, I think a 14 and a 13, no tens or aces in my hand, and the dealer had an A. I took the insurance again even though +2 is below the recommended threshold for taking insurance (also, I realize I was over-betting the hand a little (usually I would bet $160 (or round it to $175) X 2 @ +2). The dealer had blackjack again, the second time in a row.

This was an $800 turnaround for me, but did I do the right thing? I had read that it is only a marginal difference between taking insurance at +2 over taking it at +3. I felt that what I did was the more conservative move, and I was down about $2,200, but should I have done it? I'd like to hear your reasoning and comments on this.

 

[Renzey]

Here is another one; fairly easy to see, 12 v 2 at the margin. If a 10 in hand hit, if not stand. The reason your 10 in hand is one less that will break you and one less that will break the dealer so you better try to improve your hand. If you don't have the 10 in hand then that is one more 10 floating out there that could break you or the dealer so you better stand. It's not worth much.

Avenger, I think these considerations need to include interchangeable cards within the tag structure to be applicable. With that 12 vs. 2 at around +3 TC for example and using the Halves count, if it got to +3 with 10/2 in your hand, I might hit. But if it got there with A/2/9, I would stand. In both cases, the count is at the index, but one case killed one of your bust cards while the other killed one of your 21's. We don't know what other cards went by earlier to bring the count into this area, although it would be helpful if we did. Still, we do know these few here, and that's a little extra help.

Here's another example. You're four decks into the shoe and the RC is +8, making the TC +4 -- easy Insurance territory. The dealer has an Ace up and the only other player has a pair of Aces. You've got A/8. The Ace-reckoned RC has just dropped to +4, making the TC +2. Should you Insure?? I know I would! I don't know how many Aces or 10's beforehand made the TC +4, but I do know that it came down to +2 without burning any 10's.
Okay, there may be some conditional probability that says a bigger chunk of the big count was likely due to a more serious excess of Aces in the first place, but I don't know by how much. I say, Insure -- purely for EV purposes!

 

[SleightOfHand]

A,A
A,10
10,10

All are -2 hands in hi lo; correct?, yet they would shift the insurance indice. So while not running the exact math of Renzy's example. Depending on the hand composition the indices would be different.

For just hi lo ins is 3?
For just halves ins 3.33?
Both are correct, depends on what you are considering.

With more info, like the composition of the hand, things can change. It's not worth much but also not hard to implement. It is interesting to only insure one hand and have it be the correct play. Like not insuring your 20 but insuring your 3,2 or is it vice versa? hmmmmm

For generic insurance what others have said is correct, you have counted your hand and that is all you need to know.

Renzy is bringing up a marginal play, so I am pointing out some marginal considerations. I think we need more posts like this.

No one should take anything I write to heart. I like the dialogue.

Here is another one; fairly easy to see, 12 v 2 at the margin. If a 10 in hand hit, if not stand. The reason your 10 in hand is one less that will break you and one less that will break the dealer so you better try to improve your hand. If you don't have the 10 in hand then that is one more 10 floating out there that could break you or the dealer so you better stand. It's not worth much.

good cards
:joker::whip:

or is it a blowhard standing on a soapbox lol

Im not sure what your point is about hilo and halves, but yea, I forgot about aces lol. I guess composition could have some bearing on this, although if we are trying to squeeze out marginal plays, we should look at the entire board (and memory if possible) for aces and neutral cards. With the "best case" of A,3, the RC would increase by 6 (3 aces on the board), which would push the RC to +13 with 3 plus some decks remaining, says we are in the advantage.

Can someone add anything to what I said above? I am not too strong on this particular aspect of CC, and I could be wrong. Since aces are considered negative, although for insurance they should be considered positive, you add 2 to your count for each ace you see (and add 1 for neutral cards)? Wouldn't that mean you add at least 2 every time you take insurance due to the dealer's ace (perhaps the simulator already takes the dealer ace into consideration)? How would this work with a system like zen? Thinking about it more makes me feel less comfortable about how correct this is :laugh:

Now, although I know there are composition dependent indices regarding playing decisions, I am not so sure about insurance (aside from aces and neutral cards). After all, we are betting if there is a T in the hole and thats it.

 

[jack.jackson]

Im not sure what your point is about hilo and halves, but yea, I forgot about aces lol. I guess composition could have some bearing on this, although if we are trying to squeeze out marginal plays, we should look at the entire board (and memory if possible) for aces and neutral cards. With the "best case" of A,3, the RC would increase by 6 (3 aces on the board), which would push the RC to +13 with 3 plus some decks remaining, says we are in the advantage.

Can someone add anything to what I said above? I am not too strong on this particular aspect of CC, and I could be wrong. Since aces are considered negative, although for insurance they should be considered positive, you add 2 to your count for each ace you see (and add 1 for neutral cards)? Wouldn't that mean you add at least 2 every time you take insurance due to the dealer's ace (perhaps the simulator already takes the dealer ace into consideration)? How would this work with a system like zen? Thinking about it more makes me feel less comfortable about how correct this is :laugh:

Now, although I know there are composition dependent indices regarding playing decisions, I am not so sure about insurance (aside from aces and neutral cards). After all, we are betting if there is a T in the hole and thats it.

Not sure, how this would be done with Ace-reckoned strategies but heres a brief description from QFIT

I know in counts where the the ace is ommited you "technically" assign the ace the average weight for the positive card tags. Zen=1.5

As for Ace-reckoned counts, im not even sure you would use the same method to whether there was a surplus or shortage of Aces or not.

Say your playing single deck(zen) and 1/2 deck has been dealt with a RC of +3(tc+6) and no aces have been played. Now whether, theres a surplus or shortage of aces I still think you have to compare this to the number of aces there should of been on average distribution, which in this case is 2 Aces. And since 1.5 is the average weight for Zen you would subtract this from your RC, then make your TC calculation, which is 0 in this case.

Interestingly enough however, say all four aces have been played in the scenario above would you now be adding +3 to your RC or +5 since Zen tags the Ace as -1 IDK lol.