Index Interaction CVData

[NotEnoughHeat]

This is a bit more of a stats question than a software one, but the software form is more accessible so I am posting here.

Recently I was generating new RA indices. So I didn't have to leave the computer on throughout the night (noise from the fan) I tried to sim some indexes piecemeal but after isolating one index I wanted to be extra sure about (10v10), as well as adding another index I forgot to check off (9v4) I noticed that the index number was not consistent through all trials (and compared to my older index sims). Using the same parameters, but changing how many indices are generated, the same index play would appear to change value depending on which other indices were simmed with it.

I was wondering whether what I noticed is a result of index generation or if each index number is inherently interdependent on the total number of index plays you employ.

[Norm]

Some indices are dependent on other indices. If you want to do them in groups, do them in this order:

Hard hit/stand
Soft hit/stand
Doubles
Splits
Surrender

Or, you can do them by dealer upcard.

[Three]

If you want to do them in groups, do them in this order:

Hard hit/stand
Soft hit/stand
Doubles
Splits
Surrender

I am not doubting you Norm. But what is the logic of this order? I see the any number of card decisions first and then the three decisions based only on your first 2 cards next. I would think doing them in order they would be made at the table would be more effective. That is pretty much the reverse order. If you are going to surrender certain hands at certain counts those situations should not be used in the sim for any of the other decisions. Of course not every game offers surrender. A little elaboration on why this is the best order will no doubt teach me something here. Having never written software for this kind of stuff I never needed to understand how the black box came up with the induces.

[Norm]

It's the opposite order of your decision making. First, you decide whether or not to surrender. If not, can/should you split. If not, can/should you double. If not, hit vs. stand.

Look at surrender. You always know the EV if you surrender. You look at surrender last, because you must know the advantage if you don't surrender. To know this, you must have already determined what you would do if you don't surrender. So, you must examine the hands that may occur if you don't surrender first.

[Three]

Thanks Norm. I was thinking it was something like that. It just wasn't the intuitive approach for me. The more I think about it the more it makes sense.

[NotEnoughHeat]

I've found this all fascinating. The advice given is also tremendously helpful. You never know when you will do something that brings up a question you would have never thought of otherwise. I am always excited when there is a new aspect of the game for me to understand.

[BJFan]

I guess that there are different indexes depending on the rules, for example 16 v 10 may have a different index in the case you play ES10 or not, am I right?

[ZenMaster_Flash]

"I guess that there are different indexes depending on the rules,
e.g 16 v 10 may have a different index in the case you play ES10 or not ..."

OF COURSE, different rules e.g. H17 vs. S17, 8 decks vs 2 decks, DAS vs. NDAS, etc. alter indices

You mentioned 16 vs. 10 in an ES10 game. It would take a rather negative count to refrain from surrendering.

[DSchles]

"You mentioned 16 vs. 10 in an ES10 game. It would take a rather negative count to refrain from surrendering."

-5, Hi-Lo, multi-deck.

Don

[BJFan]

I was asking about Hit/Stand index of 16v10 (with three or more cards) in case you have ES10. It's clear that surrendering it's the obvious move.

[Three]

I guess that there are different indexes depending on the rules, for example 16 v 10 may have a different index in the case you play ES10 or not, am I right?

You would have a decision hierarchy. You decide on ES10 and for cases were you don't surrender you have the index. Now rules can affect your sim that generated the index. Obviously with a T up for the dealer the H17 and S17 index would be the same. On the other hand a 6 or A up you can expect some differences between S17 and H17. DAS/nDAS should only affect splitting indices. The number of decks used can affect indices. Anyway the H/S index shouldn't change with ES10. You will just surrender a lot of 2 card hands that you never get to contemplate the H/S index since ES is the first thing to consider in the decision tree. If you get to the 16vT H/S decision in the decision tree, ES10 doesn't change the index.