MR. WIZARD PLEASE READ: Can you unbalance a balanced count?
I hope this helps. Good luck!
Nothing like watching two newbs go at it. And people wonder why the "newbie treatment" thread is going so strong.Quote: SonuvabishHello, NVR. I think you misunderstood the question. I do not have any problem with TC calculation. Also, it is not a hunch. I am not sure if you are an AP or not, but all APs know that playing efficiency and insurance correlation are not as high as betting correlation, with rare exceptions.
Unfortunately, this is one of the compromises of the simpler counting systems that tag the Ace as negative along with Tens.Quote: SonuvavishIn my example, I have an even count and then a bunch of low cards come up, which increase the mathematical probability that I should take insurance. How much it increases, I do not know. But my count wrongly indicates that the mathematical probability of the dealer having blackjack has decreased. If I were, for that hand only, to count each 7 as positive 1 and disregard aces, and divide by decks remaining, I should probably take insurance (assuming equal 7s and aces). So that's really the essence of the question.
A way around the "compromise" is to switch to one of the more powerful counts which use a side count for Aces.
Actually, if insurance is concerning you that much, then you need a partner at the table (or on the rail) using a dedicated insurance count.
Quote: SonuvabishI guess this is theoretical, since this is not a very practical example. Using hi-lo, you are late in the shoe at a full table, you have a zero count. The dealer gets an Ace, and every other card is either an Ace or a 7 (so the normal count is actually negative). Does it make sense to take insurance, mathematically? Every once and awhile, I get into a situation where the count says no, but my eyeball says yes.
It would depend on the number of decks. In a six-deck game you would need 24 non-tens to come out before insurance because a neutral bet. Assuming seven players at the table you would see 15 cards. In a double-deck game the cards would likely be dealt face down.
The problem is that you are using a count that has imperfect insurance correlation. The situation described is that you see a lot of neutral cards come out on a particular hand, therefore making it more likely that the dealer has a 10 in the hole.
The problem with this analysis is that the appearance of these cards makes it less likely that neutral cards came out earlier in the shoe. You can't really treat the earlier dealt cards as uniformly distributed and random if you would have noticed certain situations and not noticed others -- this introduces bias into your analysis.
