Effect of not doubling Ax hands

[FrankieT]

Loss in EV from not doubling Ax hands

Does anybody not double down A2, A3, A4, A5, or only double them vs. 6's?

I been hitting some nasty streaks with these.

This seems to be one of the more risky double downs on high counts.

How much EV would it cost to not double down on these bastard hands?

 

[UK-21]

It'll depend on the number of decks in play.

I have a chart for the six deck game giving the EVs for all combinations of hands which I think I can post to the board.

I decided a long time back to drop doubling down at max bets out for Ax, or 9,10 v a dealer 2,3,4 to avoid the variance. Although you lose some long term +EV, with an 8 or 16 unit bet out it becomes a question of risk v return. I don't play to make money and so for me it's an acceptable compromise. For other players though, the loss of EV may not be.

 

[tthree]

Without side counts a lot of soft doubles. Important key cards are often not counted and the cards grouped as low and high have EOR values that cause a low correlation to the correct play. Look at the hand match ups and multiply your tag values by the EOR values for each card. The lower the number the worse the correlation. Different counts have different correlations for the hand match ups. Next look at the EOR values for uncounted cards. If they are very significant use the following formula to determine the relative value of these cards to your count.

N/(N-k)*E*SS/L

where
L = sum of the EOR values for the card times its count tag
SS = Sum of the squares of the count tags
E = EOR of the uncounted card(s) in question
N = number of total cards
k = the number of cards in the group of the uncounted card in question

This number produced by the above equation would be the side count adjustment to the RC. When these are high and the correlation (L) of the count is low it is a risky double with a slow increase in EV after the index is exceeded. A low correlation shows the count is a weak indicator and the double is risky. Side counts can be quite useful in upping your EV and lowering risk for risky doubles. You can side count counted cards but you must deduct its tag value for the card from its calculated adjustment. HILO is particularly bad for most soft doubles. The ace and T are opposite in their EOR but counted the same. Soft 20, 19 and 18 are HILO hands were this affect is not seen. Once you hit soft 16 or below this affect is high. Soft 13, 14 and 15 can be very dependent on uncounted cards.

 

[UK-21]

Wow. I haven't looked into it to this degree, but with the HiLo count that I use, it was soon apparent that doubling against a dealer 2,3,4 at high counts came with a public health warning. I still double the 11 v x as at =>TC+4, with the odds in favour of drawing a ten card and hitting 21.

Also, doubling a 9 v ??, and pulling a 2 kinda leaves a sour taste . . .

 

[tthree]

Wow. I haven't looked into it to this degree, but with the HiLo count that I use, it was soon apparent that doubling against a dealer 2,3,4 at high counts came with a public health warning. I still double the 11 v x as at =>TC+4, with the odds in favour of drawing a ten card and hitting 21.

Also, doubling a 9 v ??, and pulling a 2 kinda leaves a sour taste . . .

Yeah hilo has a poor correlation for soft doubles against 2 or 3. The 4 is pretty bad but the correlation for soft 18 and 19 is strong and soft 17 isn't to bad.

Hilo is very strong for doubling any 9 that the index calls for. Don't fret that decision.

 

[blackjack avenger]

consider

Risk Averse indicies
Addresses all the issues mentioned

 

[HsiaoDi]

Risk Averse indicies
Addresses all the issues mentioned

Are there published indices for different counts, or most have to generate their own?

 

[FrankieT]

Risk Averse indicies
Addresses all the issues mentioned

I just looked in one of the charts provided in the appendix in blackjack attack on A2, A3, A4, A5 and the value of doubling as opposed to hitting is significantly less than 9,10,11, A6, A7. Doubling these hands vs a 6 seems worthwhile.

Maybe they grow in value with a higher index, but it can't be as significant as the growth in value from 9,10,11,A6, and A7 because although the increase in 10's and A's increases the chance of dealer bust out, it also increases the chance of a gutterball for you're hand.

 

[FrankieT]

Without side counts a lot of soft doubles. Important key cards are often not counted and the cards grouped as low and high have EOR values that cause a low correlation to the correct play. Look at the hand match ups and multiply your tag values by the EOR values for each card. The lower the number the worse the correlation. Different counts have different correlations for the hand match ups. Next look at the EOR values for uncounted cards. If they are very significant use the following formula to determine the relative value of these cards to your count.

N/(N-k)*E*SS/L

where
L = sum of the EOR values for the card times its count tag
SS = Sum of the squares of the count tags
E = EOR of the uncounted card(s) in question
N = number of total cards
k = the number of cards in the group of the uncounted card in question

This number produced by the above equation would be the side count adjustment to the RC. When these are high and the correlation (L) of the count is low it is a risky double with a slow increase in EV after the index is exceeded. A low correlation shows the count is a weak indicator and the double is risky. Side counts can be quite useful in upping your EV and lowering risk for risky doubles. You can side count counted cards but you must deduct its tag value for the card from its calculated adjustment. HILO is particularly bad for most soft doubles. The ace and T are opposite in their EOR but counted the same. Soft 20, 19 and 18 are HILO hands were this affect is not seen. Once you hit soft 16 or below this affect is high. Soft 13, 14 and 15 can be very dependent on uncounted cards.

Thanks for the great post.

Yeah obviously when it comes to doubling, the increase in EV by count with A2, A3, A4, A5 isn't as steep as the increase in EV for 9,10,11,A6, and A7. Unlike 9,10,11,A6,A7 - an increased count withA2, A3, A4, A5 increases the chance of a gutterball for you're hand.

From the charts in the back of Blackjack Attack (which I never thoroughly examined unfortunately), it seems like doubling A2-A5 vs 6 and A6 against 4-6 seems like the best bet.



Which side count (talking about every single sidecount) is worth the extra burden if one had to choose just one?[/B] How much extra EV, percentage-wise, would this side count produce? I've always thought that most people who utilize a side count count the ace.