It seems to me like the frequency distribution on shoe counts should be such that the shape of the curve is that of a perfectly simetrical bell shape. In other words, in the long run (a billion shoes), a counters experience should be that 50% of the shoes tend to be more negative than positive and vice versa. That said, however, I just feel like I'm running into disproportionately more negative shoes than positive shoes. By negative I mean that the running count tends to be negative for more than 50% of the shoe on more than 50% of the shoes that I play. Do you know what the frequency distribution actually is in the long run? Do you know of book that addresses this issue?
Frequency Distribution of Shoe Counts
- Thread starter eddiegnz1
- Start date
On a fair deck with a completely randomized shuffle, the frequency distribution should be symmetric and in the long run 50%-50% positive-negative.
However, in reality, there are three issues you have to deal with:
(1) Non-random shuffles (hand, machine, or electronic). A hand shuffle is only as good as the hands that shuffle, and machines have liabilities too. Computerized RNG's aren't completely random either. So it's possible that you are getting non-random shuffles.
(2) Non-fair decks. Eh, let's face it, people can cheat. It's unlikely, but it could happen. To cheat you properly, though, the house would take out high cards and let the count drift upward. That the count drifts downward is actually in your favor - it means all the high cards are coming out.
(3) Recall bias. Given a 50-50 distribution of positive and negative, you might remember the negative shoes more if you're bored waiting for the count to go positive again. The positive sections will speed up in your mind because you're excited, and when you later try to recall, you will think, "wow, I spent a lot of time waiting through those negative counts." The best way to combat this is to keep detailed records - write down exactly where each shoe was halfway through and at the end. And collect a lot of data - hundreds/thousands of shoes.
However, in reality, there are three issues you have to deal with:
(1) Non-random shuffles (hand, machine, or electronic). A hand shuffle is only as good as the hands that shuffle, and machines have liabilities too. Computerized RNG's aren't completely random either. So it's possible that you are getting non-random shuffles.
(2) Non-fair decks. Eh, let's face it, people can cheat. It's unlikely, but it could happen. To cheat you properly, though, the house would take out high cards and let the count drift upward. That the count drifts downward is actually in your favor - it means all the high cards are coming out.
(3) Recall bias. Given a 50-50 distribution of positive and negative, you might remember the negative shoes more if you're bored waiting for the count to go positive again. The positive sections will speed up in your mind because you're excited, and when you later try to recall, you will think, "wow, I spent a lot of time waiting through those negative counts." The best way to combat this is to keep detailed records - write down exactly where each shoe was halfway through and at the end. And collect a lot of data - hundreds/thousands of shoes.
It will most likely be negative more than 50% of the time. Using HiLo and 75% pen the count will be 0 or negative 70% of the time. The other 30% are positive. Most of that 30% when the count will be positive will be for the majority of the shoe. You will find a lot of s hoes that get positive early and stay positive most of the shoe.
maybe the cut card effect? 
:whip:
"There is a reason for this difference related to card counting even for a Basic Strategy player. In this set of circumstances, if you deal to a cut card you may get five, six or seven rounds depending on the cards. This is because the player and dealer can have hands of two cards or many cards. If there are many high cards (e.g. tens) dealt, we will end up with two- and three-card hands. If there are small cards dealt, we will end up with hands with many cards. So, if we are dealing to a fixed point, and we are dealt a lot of large cards, then we will use fewer cards per round and get an extra round or two. The problem is that these extra rounds will be dealt from a deck with fewer high cards because they have been used up."
http://www.blackjackincolor.com/blackjackeffects1.htm
:whip:"There is a reason for this difference related to card counting even for a Basic Strategy player. In this set of circumstances, if you deal to a cut card you may get five, six or seven rounds depending on the cards. This is because the player and dealer can have hands of two cards or many cards. If there are many high cards (e.g. tens) dealt, we will end up with two- and three-card hands. If there are small cards dealt, we will end up with hands with many cards. So, if we are dealing to a fixed point, and we are dealt a lot of large cards, then we will use fewer cards per round and get an extra round or two. The problem is that these extra rounds will be dealt from a deck with fewer high cards because they have been used up."
http://www.blackjackincolor.com/blackjackeffects1.htm
sagefr0g said:
maybe the cut card effect? :whip:
:whip:
Not to mention, one loses alot more hands than one wins because of 3-2 and double and split pay-offs.
Flipping a fair coin might have a symettrical bell-shaped curve or something but I can't see BJ having one.
Didn't we go thru this a while back?
Maybe I just am too used to looking at Don's tables wherein not a single one has a total freq of -counts equalling the total freq of +counts lol.
Even Powersim that calcs TC's to the exact card doesn't have symettrical stuff.
It just seems to be a sad fact of BJ life one spends more time in -counts than + counts lol.
If anyone can show me a sim that says otherwise, I'd love to see it anyway.
Maybe some 10-level count based precisely on EOR's to the exact card or something? I don't know.
It's not like I really care anyway whether they may be symettrical or not. I just care what the freq are for each TC and what the adv is when they do occur.
yeah we were talking about this tc frequency stuff before.
and i think we also said how if you sumproduct the tc freq's and the advantage at a given tc you get the advantage of the game against a flat betting player using indices and basic strategy sort of thing.
i guess if you were just flipping cards out of some pack and not playng that you'd over the long term see symmetry of negative true counts and positive true counts.
but when you start having rules and an inherent advantage for the dealer i guess that some how skews the symmetry of the true count. the mechanics of how that would arise is a mystery to me lol.
and i think we also said how if you sumproduct the tc freq's and the advantage at a given tc you get the advantage of the game against a flat betting player using indices and basic strategy sort of thing.
i guess if you were just flipping cards out of some pack and not playng that you'd over the long term see symmetry of negative true counts and positive true counts.
but when you start having rules and an inherent advantage for the dealer i guess that some how skews the symmetry of the true count. the mechanics of how that would arise is a mystery to me lol.
The curve is NOT symmetrical even if the count is calculated perfectly. See http://www.blackjackincolor.com/truecount8.htm