the "Averse Eight"?...quantifying R-A

[Adam N Subtractum]

As per zg's request, I took a little time to examine the data I have concerning risk-averse indices, to see if we can get an idea of the most important modifications to "ev maximizing" strategy adjustments. Keep in mind the data I have on this subject is limited (Karel Janeck's Blackjack Risk Manager will give you exact answers specific to your system/conditions), so consider this a rough quantification, and treat it as such.

To keep this synapsis (& my research brief I've assumed players will be using Catch 22 for 1D, Catch 22 + Fab Four for 6D, and do not want to add any additional indices to their set.

-Single Deck Risk-Averse Plays-

T v T
9 v 7
8 v 6

The indice for T v T and 9 v 7 will probably be raised by 1 to 2 TC points, and the number for 8 v 6 will probably raise by 3 to 4 points. I suspect A,8 v 5 & 6 will be altered as well, but I do not have the data at hand to make a conclusion.

-Six Deck Risk-Averse Plays-

T v T
T v A
9 v 7
8 v 5
8 v 6
15 v A (LS)

In the shoe scenario T v T and 9 v 7 should be raised more than in 1D, perhaps 4, or even as much as 5 points. T v A should only be adjusted upward by about 1 TC point, and 8 v 5 & 6 maybe 2 or 3 points. And finally, for the Late Surrender rule, we'll adjust the 15 v A indice DOWNward by about 1. See above concerning A,8 v 5 & 6.

From the info I have available to me at this time, these are the only plays affected by Risk-Aversion in the Catch 22 / Fab Four set. If one was willing to add on additional indices (or already use more) there definitely is more to be gained, but is beyond me quantifying at this time. Intuitively, I presume A,2 v 5 and 8,8 v T will be amongst the top contenders for immmediate addition.

Another avenue I'll briefly touch on here is the use of Risk-Averse INSURANCE indices. In other words, it is sometimes optimal to insure different hands at different times, in order to reduce variance, and in turn increase favorability. The following is a general rule of thumb.

-Composition Dependant Risk-Averse Insurance-

stiff hands = indice + 1
T,T or T,A = indice - 1
all others = usual indice

I wish I had some more conclusive data, but perhaps someone will follow-up with additional info to solidify/modify these estimations.

ANS

 

[Adam N. Subtractum]

Re: the "Averse Eight"?...further research...

...has uncovered some unlikely additions to our risk-averse set. With a little help from Michael Hall, it has come to my attention (kind of embarrassed I missed this) that risk-aversion does NOT just affect double, split, and surrender hands, as would be assumed intuitively, as R-A can have an affect on hit/stand hands as well. (Bonus question: can anybody guess why??? Not you T-H, let somebody else try

Also the rule of thumb I gave for insurance is being revised as we speak, and will be modified. Hopefully I will be able to post the follow-up sometime tomorrow.

ANS

 

[alienated]

Re: the "Averse Eight"?...further research...

"it has come to my attention (kind of embarrassed I missed this) that risk-aversion does NOT just affect double, split, and surrender hands, as would be assumed intuitively, as R-A can have an affect on hit/stand hands as well. (Bonus question: can anybody guess why??? Not you T-H, let somebody else try"

This is news to me, too. This is a wild guess, but is an example A,7 v A in S17 games? The hit and stand expectations are very similar for this hand, but hitting has more downside (and upside). I also wondered about A,7 v 9,T and A,6 v 7, but the expectations seem too far apart in those cases.

 

[Adam N. Subtractum]

Bonus answer...

"This is news to me, too. This is a wild guess, but is an example A,7 v A in S17 games? The hit and stand expectations are very similar for this hand, but hitting has more downside (and upside). I also wondered about A,7 v 9,T and A,6 v 7, but the expectations seem too far apart in those cases."

I think you're on the right track here, I haven't looked at these plays specifically yet. A similar example I can give is 16 v. T, another close play. Now the "stock" KO indice is +1, while the R-A # is +2. Waiting until +2 to stand is the more optimal play...why is this? Because there is a DECREASE in variance by _hitting more aggressively_ in this instance. The simple reason for this, which previously alluded me, is that by standing on a hand of less than 17 there is no chance of a push! By hitting these hands we have possible outcomes of +1, 0, -1....there is a chance that we will neither vary up, nor down, therefore REDUCING VARIANCE.

When this variance reduction is more beneficial than the ev loss is detrimental, it would be optimal to adjust the indice accordingly.

ANS

 

[Adam N. Subtractum]

Ted, ren the right track...A,7 v A

I checked the data on A,7 v A, and your intuition was correct Ted, the indice is adjusted. In this case we _already have_ a chance of pushing by standing, and reduce that chance by hitting, because there is the possibility of busting (obviously not immediately, after subsequent hits). The more we bust, the less we win and PUSH. So we would want to adjust the indice downward by one, in this case, so we will stop hitting sooner, and therefore lower variance.

Interestingly, the 16 v T example I gave in my prior post required a +1 increase for R-A KO, yet this apparently does not apply to Hi-Lo (from what I have available to me). What I do know is that this affects only a very few hit/stand plays, because most of the time the R-A difference is so minimal that we'd have to use fractional increases/decreases. Now my theory as to why 16 v T is adjusted with KO and not Hi-Lo, is that the +1 increase to KO is applied to the running count, which basically allows for fractional increases in TC, does that make sense?

The other plays mentioned were not affected as you presumed. I'll include all the affected hit/stand hands in my revised post.

ANS

 

[Adam N. Subtractum]

Re: A,7 v A...Correction...

I said:

"The more we bust, the less we win and PUSH"

This is obviously incorrect, as we wouldn't be hitting if we didn't win more often, but the main point still stands...we do push less by hitting in this situation.

ANS