Originally Posted by
DSchles
You seem to be misunderstanding what I'm saying. If you use true count, I'm fine with using a single index regardless of penetration. But, you seem to be talking about using a single RUNNING count, regardless of level of penetration. What sense does that make? Do you propose using the 3/6 RC for when you're at 1/6 or 4.5/6?
Don
I'll try explain my methodology.
In order to compute probability of drawing each rank for an already input counting system:
1. First required input is number of cards remaining to be dealt
2. Second required input is the running count
3. Third input is cards that for one reason or other the user has determined should be specifically removed from the shoe
4. User can optionally decide to side count any ranks or even create subgroups which to side count. Generally this is not used but if taken to an extreme all ranks could theoretically be side counted and in that case there would only be 1 possible shoe composition and number of cards remaining possible.
Notice that so far there is only consideration of running count. In fact there will never be consideration of true count until the user decides it is prudent to calculate it from a running count/cards remaining pair.
Let's say user wanted to decide whether to insure a hand of T-3 with 26 cards remaining dealt from single deck and a HiLo running count of +2. Input is cards remaining=26, RC=+2, specific removals = T,3,A (A because insurance requires up card of ace.) Output of rank probabilities are below. The probability of drawing a ten is .34646, so insurance should be taken.
Code:
Count tags {1,-1,-1,-1,-1,-1,0,0,0,1}
Decks: 1
Cards remaining: 26
Initial running count (full shoe): 0
Running count: 2
Subgroup removals: None
Specific removals (1 - 10): {1,0,1,0,0,0,0,0,0,1}
Number of subsets for above conditions: 7
Prob of running count 2 with above removals from 1 deck: 0.092696
p[1] 0.069293 p[2] 0.071334 p[3] 0.0535 p[4] 0.071334 p[5] 0.071334
p[6] 0.071334 p[7] 0.081803 p[8] 0.081803 p[9] 0.081803 p[10] 0.34646
Press any key to continue:
What I have done recently is to take this further. Below is the combinatorial analysis of a player hand of T-3 for a running count of +2 and 26 cards remaining to be dealt from a single deck versus all up cards. A shoe is created with number of ranks present expressed in floating point values based upon the probability of drawing each rank. I changed my combinatorial analyzer to be able to process these.
Code:
Number of decks: 1 Count tags {1,-1,-1,-1,-1,-1,0,0,0,1}
Player hand composition: 0, 0, 1, 0, 0, 0, 0, 0, 0, 1: Hard 13, 2 cards
After player hand is dealt - Cards remaining: 26, Running count: 2
Subgroup removals: No subgroups are defined
Shoe comp (A-5): {2.2908, 1.8698, 1.4023, 1.8698, 1.8698}
Shoe comp (6-10): {1.8698, 2.079, 2.079, 2.079, 8.5906}
After up card is dealt - Cards remaining: 25
Running count (up card 1 to 10): {1,3,3,3,3,3,2,2,2,1}
Up card Stand Hit Double Split 1 Split 2 Split 3 Surr Strat
1 -76.073 -62.669 -93.209 hit
2 -27.855 -33.159 -66.318 stand
3 -21.789 -30.328 -60.657 stand
4 -11.566 -26.419 -52.838 stand
5 -3.987 -23.146 -46.293 stand
6 -11.665 -25.074 -50.148 stand
7 -48.421 -32.838 -67.954 hit
8 -53.715 -39.134 -82.297 hit
9 -54.028 -38.547 -79.197 hit
10 -59.167 -47.379 -87.233 hit
Overall hand EV vs all upcards: -35.1492
Press c or C for EV conditioned on no dealer blackjack, any other key to exit
When I generate indices I go through all of the possible running counts for a given penetration and record the running counts where strategy changes as running count indices. True count can then be computed from these and is dependent upon the number of cards remaining which has been input.
As I mentioned I decided to default to half shoe pen when generating indices to keep it simpler but it would be easy to give user the option to input pen. Anyway indices are always recorded as running count until they are converted to true count at the end.
Hope this helps,
k_c
Reply With Quote
Bookmarks