I bought your book "Professional Blackjack," and it was very helpful. I have a math (pure) and statistics (econometrics) background, so I find learning the underlying reasons for the different systems as useful. It gives me confidence in play.
True Count vs. Running Count.
In practicing to go to Vegas, I have been playing at Atlantic City. Over 3 weeks, I learned that I could count Hi-Lo without trouble at the table, and incorporated the two main changes to basic strategy you recommend (hard 16 and hard 15 vs. 10). Not too tough. Rewarding it seemed.
I bet using Kelly's criterion, in essence, but modified from the way most books would recommend. When the 8 deck shoe had 5-8 decks, I used the advantage computations for a running count (with its higher variance) and, when the shoe slipped below 5 decks, I used a single weighted average true count (an approximation, to be sure, but one that is easier to keep in mind than a constant change to the true count).
It turns out that the advantages merged, or were very close, so that bets at common running count numbers were fairly standard. I averaged to make them entirely standard, and tried when the count was high to spread bets over two spots.
If the data at Casino Verite is correct, I do not see the problem (other than increased variance) from using this approach. But I'd like to see it simulated (a Monte Carlo simulation) for the hybrid betting strategy -- easier to remember. Perhaps as a running count becomes more engrained in my mind and less mental work, there is no point but it greatly simplified calculations (which I could show you on a simple spread sheet) about when to bet extra units based on my modest bankroll (I'm conservative by nature; the game being more of a mental challenge).
Have you seen such a system modeled? (The problem is that in a multi-deck shoe, the number of times you vary your bets in the first four decks is very small if you only bet based on the true count, yet the running count at numbers of +7 or more starts to offer real advantages over the house -- albeit with greater variance than a true count.
Your thoughts? Anyone who would run a monte carlo experiment for me using this method?
