Sorry about the typos again...Better version Begging Wong for count
There are days like that.
Better version of my response to George I.
Haze This! just advocates using the Griffin EORs, for doubling down and pair splitting, to find tables for the change in expected value, as a percent of your original bet, that each true count brings you. A simple formula of waiting until the true count is past the regular index enough to justify adding money, as if the double down adding of money were a NEW initial bet, is NOT the actual formula for risk adverse indexes or precise risk adverse playing, but it does allow adjustments for the actual bet you initially made (cover needs happen!) and it does allow two sets of numbers: the main index and these "delta numbers" to suffice no matter what the spread and what the starting advantage.
Regular risk adverse indexes, as provided by virtually all the simulators, assume that the true count at the time you make your playing decision is the same as it was at the time you decided on your bet amount (on AVERAGE it does but in the best deeply dealt games it diverges greatly) and cover needs, betting at other than your optimal spread, lessen the benefits of using them. Since they involve modifying your indexes, even though one final index is used for the play, you have to change them each time the starting advantage changes.
The above summerizes Haze This!
What is the aplha of your decisions points out how the assumptions you make about your overall variance per hand do not hold for the variance of a doubled or pair split hand. By speaking of alpha, rather than SCORE or N sub(0) for a decision, you can also point out that the index that leaves these measures unchanged is slightly lower than the usual index. The usual index is where the expected value of your play, doubling the amount bet but only getting one card, raises variance but does not add expected value. Because the variance assumptions for an initial bet are not the same as the variance assumptions of doubling or splitting, the zero impact index is actually lower than the zero added expected value or usual index.
The surrender decision is NOT based on whether or not to play out the hand using basic strategy, but is based on whether to play the hand using a count index or surrender it. The EORs for surrender (see bjmath.com's search engine) are developed on the basis of perfect play versus surrender. EORs are a single parameter perfect count for just ONE decision if you will.
In both surrender or doubling or splitting, with usual counts, or with EORs, coming close to actually playing by the remaining cards left in the pack, with the index, or the EORs, using the minus of the full deck expectation as the equivalent of a usual palying index, you are deciding between how the expected value changes between playing options, and except for usual surrender, are not playing to a specific expected value.
The math for finding an algebraic risk adverse index, the same way risk adverse indexes are found by SBA 6,7,8, CVX or etc., which is with Jalib's True Count Theorim, that the true count does not change between the betting and playing decision times, which is only an average trend, is recursive, even though it is still simple algebra. You have to find the spread, the amount the expected value changes for each true count and the ACTUAL edge of the hand, NOT just the edge changes for each hand.
But then the limits of the assumptions you use with usual risk adverse indexes still hold. I have already decided that the INACCURACY of my delta force tables is outwieghed by the simplicity of their use and the ability to try to be risk adverse in playing when you have to make a double or split decision when your actual bet is far from your optimal spread amount. That inaccuracy tends to be slightly playing doubles and splits too conservatively and waiting for slightly higher true counts, usually 1 true count more, than the usual simulators give. It turns out to be slightly too conserative, even though the Haze This! recomendations ignore actual variance for the play, because the zero impact index is actually lower and thus more agressive than the sero edge index, and the amount by which each true count improves your expected value is higher by the overall flat bet variance of the overall game. It is another case where 3 wrongs result in a final decision very close to right.
That is the basis of my answer to George I.
George should stick to usual surrender indexes or indexes he derives from usual surrender Effects of Removal. He is not really dealing with two decision recomendations as he thought when he compared both the surrender and count decision options to the basic strategy play.
Should he wish to combine the recomendations from different counts he should consider the value of having one player ten count for insurance while the other uses a regular point count. That combination is number two on the he said/she said division of counting duties.
The number one combination would occur when a game combines early surender against a dealer ten and the over/under bet. There are some casinos in Europe that I have been emailed about and invited to play (team solicitations). Then the suggested combination would be the regular point count with either the Snyder over/under level one or the Crush count (see Proffesional Blackjack 1995 edition!).
The number 3 combination would be to beg Wong to disclose the special count that was mentioned in his Blackjack in Asia, for the 5 card Charlies rule. He has in the past stated that the strategy recomendations are all in Basic Blackjack, but he never disclosed the count that he also had in BJIA. Some estimates from the cards that would change recomendations in the 4 card tables in PBJ or BBJ, would indicate where another player with some special count might double the gains that would come from the hi-lo 4 card indexes.
So George I. was on the wrong track for his thoughts on conflicting recomendations. But you can use the points raised in Haze This! and What is the alpha of your playing decisions? (just to still be ornery I didn't properly capitalize this mention!--it is mine so I will....) to show how the critical points you use to develope or simulate or algebraicly derive an index might be different than the usual assumptions. You might neglect something similar to the point in What is the alpha...?, where the zero long run impact index is highly agressive, and miss exactly what options you are weighing. You might also try to find an exact formula you could never use at the table, but ignore a simple formula that does better than a precise table of indexes that assumes two important premisses that are often not actually doable in real play.
So George I. was "out of the box," AND WRONG, and asking me to help because I had been out of the box myself in a similar fashion and had come up with correct findings. The rest should be explained by the above and I have added three situations where there are VALID ways to consider dividing up the counting chores, other than George I.'s idea.
But I had a bad day. Better to say I had a bad day than now go on and on about my now perfect errata IMHO.
So there!
