Norm Wattenberger: How often do you recalculate the True Count?

[Norm Wattenberger]

The Question

Question is, do we need to recalculate the TC before every decision.

The Sims

I ran a pile of sims this morning to take a look. There are four types of sims looking at four different levels of TC calculation effort as follow:

[*]Before every bet[*]Before taking Insurance[*]Before the first playing decision[*]Before every close playing decision[/list]

I also ran two sets of these sims as follows:


[*]Double deck, H17, DOA using AOII, full indexes, Ace side count, 4:1 spread[*]Six deck, S17, DAS, LS using Hi-Lo Illustrious 18 & Fab 4, 8:1 spread[/list]

For each sim, the following penetrations were simmed:


[*]Double deck: all penetrations from 26 to 52 cards cut off[*]Six deck: all penetrations from 26 to 104 cards cut off[/list]



The Results



The below links display two charts. Each chart plots penetration versus SCORE for the four different levels of TC calculation effort. At six deck, there is no gain at all from recalculating the TC after the first decision. That is, get your first two cards and hit or split and there is no need to count the drawn cards before making further decisions. In the double deck games there was some gain. Overall, the gain from recalculation was higher in six decks than I expected, despite the fact that only 22 indexes were used. I would also expect the spread to make a difference and have the data for all spreads.


[*]Double deck Chart[*]Six deck Chart[/list]



Sim details



All sims were 400 million rounds, four players, quarter deck resolution, no cover, no errors. Optimal betting ramps were calculated for every penetration.

[Don Schlesinger]

Great job, Norm! Fascinating study. In general, as a default for 6-deck, I'd suggest using a 1-12 spread, at the least.

Funny, but, to your original question of: "Do we need to recalculate the TC before every decision?" my immediate knee-jerk response was: "What else do we have to do?"!! :-)

Thanks for your usual scholarly effort.

Don

[Richard Reid]

Norm:

Just out of curiousity, why do the 6 deck results show that calculating before every close playing decision as being ever so slightly less optimal than simply calculating after the first playing decision? I would have thought the reverse would be true. Notice that the results for the Double deck chart look fine, but the 6 deck results, although close, seem reversed.

Sincerely,
Richard Reid

[Norm Wattenberger]

I didn't have time to look into that. The sims were run with CVCX because it quickly sims all penetrations at once and automatically applies optimal betting. But it doesn't show the deep data needed for such a study. I would need to rerun with CVData, probably 10-20 billion hands for such a minor difference. Then compare the results for each set of player cards vs. dealer cards to see where the differences lie. Also find at which counts the difference occurs. An example of this technique can be found at First Base Penalty. Then determine the cause. If I had to guess, two possibilities come to mind. A.) The indexes could be incorrect. Hi-Lo indexes from PBJ were used and they are not perfect. B.) With this small a difference, it is possible that the method of estimating the remaining decks and/or TC divisor resolution were not the same as that used in generating the indexes. That would favor recalculating or not recalculating depending on the difference.

regards,
norm