True Value of hi-lo TC points...(kinda long)
The following is a re-post from BJMath.com. I believe this pertains to the question your asking, and may be of interest.
Of course the comments made by Igor are accurate, but I believe they fail to take the value (& therefore the effect on advantage) of the high cards into consideration. I am open to comment and cririque on this theory.
This is some further research into a matter I posted on below under the handle "curious" (I was offline and on a relatives PC for a while). I believe the findings may be of interest, and am open to comment & critique.
>EORs in a balanced count sum to precisely zero.
Of course, you are correct here.
>There is no such effect as the one that is being mentioned. Anything that doesn't sum to zero is due to rounding error.
My point was not that the EoR's don't sum to zero. I stated that the EoR's OF THE CARDS WE "TAG" don't sum to zero.
Let's use a hi-lo one deck example for simplicity:
-Effects of Removal-
2-6 =.504(mean)
7 = 0.28
8 = 0.00
9 = -.18
T-A =.53(mean)
For our example, lets assume we are half-way through the deck with exactly 26 cards remaining and a Running (and therefore True) Count of zero. Let's assume the following about the composition of the played cards:
-Cards removed @ 1/2 deck w/ hi-lo TC=0-
7 = (4/52 = 2/26) = 2
8 = (4/52 = 2/26) = 2
9 = (4/52 = 2/26) = 2
2-6 = (26-2+2+2)/2 = 10
T-A = (26-2+2+2)/2 = 10
So the effects of these cards removed would be:
-Effects of Removal-
7 = 2*0.28 = .56
8 = 2*00.0 = 0.0
9 = 2*-.18 =-.36
2-6 = 10*.504 = 5.04
T-A = 10*-.53 = -5.3
So the sum of these EoR's is -.06, as I stated in my follow-up. But, I believe we need to divide this by decks remaining, so the result is -.12 in this case. Therefore, on average, at the half-deck level in SD w/ hi-lo, our advantage at a TC of zero will be approximately .12 less than we presume. Obviously not very notable, but in fairness, the the effect would be magnified deeper into the deck as I originally stated, though it still would rarely approach even -.2 in the deepest dealt games.
But I believe my original statement about the effects this will have on the Floating Advantage may be of merit here. Obviously, as stated this effect is not substantial on our overall edge, but would be, on a smaller sample like the the FA.
Now let's examine the effect on higher counts. Let's take the same example of hi-lo, single deck at 1/2 deck, but with a Running Count of +3 this time. Let's assume the following about the composition of played cards:
-Cards removed @ 1/2 deck w/ hi-lo TC-
7 = 1
8 = 1
9 = 1
2-6 = 13
T-A = 10
So the effects of these cards removed would be:
-Effects of Removal-
7 = 1*0.28 = 0.28
8 = 1*0.00 = 0.00
9 = 1*-.18 = -.18
2-6 = 13*0.504 = 6.552
T-A = 10*-.530 = -5.30
So the sum of these effects are +1.352%, but again we must divide by decks remaining to get +2.704. So the commonly used .5% per TC would have us assuming an advantage of .5*(3/.5)= 3%. Note that we are count now errs by .3%, and if you use the .56% number I've seen touted in some books, you will err by .66%. All of the sudden the effects aren't so minimal anymore, huh? Of course the effect would only be even this large at a TC of +6 or more, which we will only see a little over 2% of the time, depending on penetration, but it seems more possible that the number could definitely reach -.2 than I originally presumed.
ANS