OK fellers, I'll go a bit of a long way for this one and I can only do Hi-Lo or KO at a given pen level, but I think the principle would extend to any other count as well.
In computing an index I think you want the average composition according to a given count. Say decks=2 and pen is 52 cards. (Dead link: http://www.bjstrat.net/99_v_10.html) _This image_ shows the average Hi-Lo comp probability of drawing each rank from the remaining 52 cards given a hand of 9-9 vs 10 and a running count of -4. If RC =-4 with 1 deck left then TC=-4. This is but one pen level out of many the could be input. The key is that the program shows the probability of drawing each rank considering all of the possible Hi-Lo count distributions at any given pen. What we want to do is reference a composition that as much as possible reflects these probabilities.
Code:
[u]From the program image, average number of each rank remaining in shoe[/u]
(2-6) .08593*52 = 4.47
(7-8) .07913*52 = 4.11
(9) .05935*52 = 3.09
(10) .28038*52 = 14.58
(A) .07235*52 = 3.76
Since we are not dealing in whole numbers, all we can do is use whole numbers that will approximate the above comp.
In tdca you could do the following:
Input 2 decks
Input 9-9 v 10
Remove 4 each (2,3,6)
Remove 3 each (4,5)
Remove 4 each (7,8)
Remove 3 more (9 - total of 5 removed)
Remove 17 more (10 - total of 18 removed)
Remove 4 each (Ace)
This is reasonbly close to the average comp proportions. It only adds to 51 cards because of rounding and using whole numbers and TC=-4.1. Other reasonable whole number approximations could be used and results would vary a little. Also, different pen might change things some. The idea is to do a calculation for something resembling an average comp and not some unlikely extreme comp.
Anyway clicking Calc will test the validity of a Hi-Lo index of -4 for that (~average) comp and a hand 9-9 v 10:
splitting: (-24.89% DAS) (-24.88% NDAS) [DAS < NDAS because doubling 11 according to BS is worse than hitting at this comp]
standing: -13.04%
So, no, do not split 9-9 v 10 at a Hi-Lo TC=-4 and pen of 52/104 cards (and probably at most any other pen.)
Maybe there is some positive index for Wong Halves where it is better to split 9-9 v 10. I don't know for sure but I would tend to doubt it. In any case the index plays would cover the average situation and not necessarily the extremes.
The point is that defining an index from some remote shoe comp doesn't make it right just because it works in that one case. It just may be a long way from what the actual index should be.
