Re: Two ways..

From Michael Dalton's BJRnet

Posted by Brh on 4 March 1999, at 3:14 p.m., in response to Re: Two ways.., posted by Gary on 4 March 1999, at 11:24 a.m.

Gary,

Firstly, to be strictly correct, a running count system does not have static indices, thats why its efficiency is lower than a true count system. When fixed indices are used, they are an approximation.

It may help to look at what is called a 'neutral' deck. Consider a 6 deck game with K-O, and lets start with the 'correct' IRC of -4*6=-24 (note this is different from the K-O book). This value of the IRC allows true count calculation.

Assume now that the deck stays neutral throughout, then at each deck level we have an average increase in the RC of +4, since that is the net unbalance per deck:

decks RC
6 -24
5 -20
4 -16
3 -12
2 -8
1 -4 

Now from this neutral position, you can then 'evaluate' the RC at each deck level, in terms of its deviation from neutral. Lets see what an RC of 0 means at each stage of the deck, and evaluate it in terms of 'average' positive density per deck (ie true count)

 deck diff..from..neutral density UTC
6 +24 +4/deck 0
5 +20 +4/deck 0
4 +16 +4/deck 0
3 +12 +4/deck 0
2 +8 +4/deck 0
1 +4 +4/deck 0 

so you see, whenever the RC is zero (or IRC+24), there is on average four extra big cards per deck corresponding closely to a Hi-Lo true count of +4. That is how the pivot works, while the RC is increasing, the number of decks is decreasing, and when you take the difference between the RC and the 'neutral' deck at that point, it is the same.

Instead if we look at say and RC of -12, we have

 deck diff from neutral density UTC
6 +12 +2/deck -2
5 +8 +1.6/deck -2.4
4 +4 +1/deck -3
3 0 +0/deck -4
2 -4 -2/deck -6
1 -8 -8/deck -8

so you see an RC of -12 gives a different 'density' for each deck level. That is just after shuffle and RC of -12 is an advantage deck, with 3 decks remaining it is neutral, and with one remaining it is badly negative.

Now if say a play had a Hi-Lo true count index around zero, then you would have an average RC index of around -12, which would be correct at the 3 deck level, but increasingly bad on either side, which is the best you can do without doing a true count conversion. However, if a play had an index around Hi-Lo TC=+4, then the RC index would be 0 (ie IRC+pivot), and it would be always correct no matter what the deck level.

By the way, SBA knows all about this, and can correctly do unbalanced index generation for both running and true counts.

Hope this helps, Brh.