falling true counts

[QFIT]

Yes this is unnecessarily confusing. When we speak of advantage, we speak of the advantage given information currently at hand using a defined strategy. If the TC is high, we have an advantage. And the TC, on average, will remain the same until the shuffle. If you knew the count of the cards after the cut-card (possible with shuffle-tracking) you could calculate your advantage differently.

 

[sagefr0g]

If you have a high rc at the begining of the shoe but it levels off and does not rise further nor really drops the true count would drop as more of the shoe is dealt. Would you not be dropping your bets down as the shoe is depleted of cards because your tc is dropping likewise. If I cut back my smoking from 2packs a day now to say 1 cig a day then die from some other reason other than cancer then smoking didn't kill me did it. blackchipjim

wouldn't the true count raise a bit in that instance?
maybe cut down even further on them cigs bcjim.:whip:

 

[k_c]

i've read words to the effect that lead one to believe that essentially for us to realize the fruits of the advantage of a positive true count that in the process the 'count' falls. by count maybe it's meant RC or TC, i'm not really sure.
it seems to make sense, or at least it's possible if for example your one on one with the dealer and the count is nice and positive,say tc=4 what ever, the dealer has two tens say and you have an ace and ten. well ok the RC went down by four. so here the RC fell.
ok but it's also possible a lot of other things, like for example the dealer has 3,3 and you have ace,10 . here the count doesn't fall.
or say there are two other players at the table. maybe the dealer has 3,3 and you have ace,10 and the other players have 10,6 and 8,7 , what ever.
so the 10,6 guy could get a 5 for 21, the 8,7 guy might get a 6 for 21 and the dealer might get a 2,8,4 or what ever. point is it's not beyond the pale that you have a nice high true count like tc=4 or such and you might realize the fruit of the advantage by getting a snapper, successful double down or what ever and still the count could stay the same or even rise.
i dunno, i just posted this because well maybe me more than anyone else but others as well might think this falling true count idea is so likely to be associated with realizing an advantage sort of thing. i guess my point is that it ain't necessarily so.

This could apply to your "Luck" thread.
A positive TC tends to stay the same, but it doesn't have to.
-You could win and TC goes up. Snapper for you low cards for everyone else and his mother. You are lucky and you are favored to be lucky on the next round.
-You could win and TC stays the same. You are lucky and still have a favorable outlook for the next round.
-You could win and TC goes down. You are lucky you won. TC might not now allow a large bet, but a bird in the hand is worth two in the bush.
-You could push and TC goes up. You are lucky to be in a better situation at no cost.
-You could push and TC remains the same. No harm no foul.
-You could push and TC goes down. Just a cost in time spent.
-You could lose and TC goes up. No problem if your health care covers psychiatric care if it happens again. Not too lucky.
-You could lose and TC stays the same. Requires less psychiatric sessions. Unlucky.
-You could lose and TC goes down. You were unlucky and lost but at least you don't have to worry too much about your mental state unless this has happened many times in succession.

:grin::laugh::grin::laugh::grin::laugh:

 

[iCountNTrack]

well i don't remember reading specifically such an idea as bjcount's paraphrasing indicates, but such implications have been floated about during discussions around here and maybe in some books as well. thing is such wording are probably just used as trying to explain how perhaps even with an advantage or high true count at some point things might not go so well, sort of thing or some other explanation of what ever.

like these lines i read in Blackbelt in Blackjack, page 200.
"...... You don't bet big simply because the count is high, you bet big because the count should come down.
If the count doesn't come down then it must mean that those excess high cards in the deck did not come out. If the count continues to climb, then not only are the excess high cards not being dealt, but a disproportionate number of low cards continue to be dealt, much to you disadvantage.
If the dealer shuffles when the count is +15 to +20, then this means that all those high cards are clumped together in the undealt portion of the shoe. If this happened three shoes in a row, then contrary to what this players count indicated, he never played with an advantage over the house. When the count stays high, your high bets are all for naught. ..."

ok, so i guess my point is, yes this sort of reasoning explains some stuff and is innocent enough in the point it's trying to make. thing is maybe one shouldn't go to far overboard about these points and make the distinction in error that just because the count is continuing to rise that your advantage isn't good sort of thing. it may or may not be good. still worth taking a shot at. lol.
it's not so much that it's been anyone saying some wrong headed statements as it is that some of the statements might could give rise to some wrong headed conclusions, is more the point i was trying to make.:whip:

Snyder creating endless worthless discussions hmmm sounds familiar

 

[sagefr0g]

Snyder creating endless worthless discussions hmmm sounds familiar

well, maybe lol but the the book Blackbelt In Blackjack is a definite classic full of all kinds of blackjack wisdom and Arnold Snyder worthy of his place in Blackjack Hall of Fame, IMHO.

A high true count is advantageous because at any time t there is a higher probability that the next card dealt would be a high card making:

your double downs and splits more successful

increased the probability of getting a BJ or 20

more probable for a dealer to bust a stiff hand (doesn't mean that the dealer will bust more often, he will bust stiff hands more often, subtle difference)

A fast decrease in Running Count is advantageous but this does not result in a fast decrease in true count because TC= RC/(Cards Left) so numerator and denominator almost cancel each other.


I think it is a good place to also explain the floating advantage concept and make analogy to help visualize it:

The floating advantage is basically the concept that a certain true count is more advantageous as you go deeper in the shoe.

QFIT has nice figure about that:

http://www.blackjackincolor.com/blackjackeffects2.htm

What is the physical reason behind it? well as mentioned above True Count is not really a count it is more a density or from a chemist point of view a concentration. For instance, you have 1 gram of sugar that you mix with 0.25 L of water, your concentration of sugar is 4g/L .

Now the only difference is that sugar solution a homogeneous solution i.e all the sugar is evenly distributed which is not the case for a shoe of cards rich with high cards which is heterogeneous. Meaning it is very probable you will have very sweet areas (rich) and some very bitter (poor areas).

However for a given True count, as the size of the remaining cards decreases the heterogeneity would decrease, and you rich cards will be more
evenly distributed.
__________________
-iCountNTrack-

i got a kick out of the density analogy. i guess the heterogeneity would decrease as long as like Mr. Snyder says the sugar isn't behind the cut card. :laugh:

 

[Martin Gayle]

The True Count has to fall for an AP to realise the advantage. The RC will naturally move as more hands are dealt but this does not neccessarily effect the TC. The RC HAS to move closer to 0 as fewer and fewer cards are left.

If you have 5 cards left in the deck and they are all X's your RC will decrease as you deal those cards but the TC will stay the same.

If d=cards still in play TC= RC/d therefore RC is not dependant on TC.

 

[callipygian]

this is unnecessarily confusing.

I agree it is confusing, but I argue it is not unnecessarily confusing.

There are selected instances when retrospective advantage is useful, for example, in a mathematical analysis of the advantage gained from Wonging.

A player who Wongs out actually benefits in two ways. One, by not playing low counts, he shifts the distribution of counts played towards the high end (prospective advantage). In addition, the shoes in which the count drops may have been played at an advantage (retrospective advantage). While this does not change the playing decisions (which, as you point out, are entirely based on prospective advantage), a calculation of his total EV should include the fact that some of the shoes he Wonged out of were actually positive EV shoes given that the Wonger played the positive end of the shoe and then left.

Example: A 6D game in which the count dropped from TC 0 to -2 over two decks is actually favorable for the player. The player has played the equivalent of two decks at a +4 count.

The contrary is true as well - for shoes in which the count rises from 0 to +2, while the player has a prospective advantage in future bet sizing, the player retrospective advantage has been worse than anticipated because a lot of low cards have come out. In the absence of Wonging or bet scaling, the two effects are symmetrical and cancel each other out. However, as one varies bets and/or Wongs, these effects are non-zero (but small).

 

[QFIT]

The True Count has to fall for an AP to realise the advantage.

Sorry, but this is not true. The TC means that the deck is rich. If you deal more cards and the TC remains the same, this must mean that you have been dealt more high cards than low cards. You must have been dealt rich cards. The running count must drop, not the true count.

 

[QFIT]

I agree it is confusing, but I argue it is not unnecessarily confusing.

I have no problem with additional info. I just don't like temporarily redefining basic terms.

 

[bj bob]

Adding clarification / confusion

Let me first begin with an example of an interesting observation of the relationship between the true count and the running count. Take a fresh 6 deck shoe and let's say that 27 low cards are dealt, 15 high cards and 10 neutral cards. We now know that the sum of total cards dealt at this point equals 52 (1 deck) , the RC= +12 and the TC= +2.4. Now let's suppose that the RC stabilizes form this point until the last 1.25 decks. In this scenario we see that there is absolutely no change in the RC; however, as the shoe is dealt down and the deck "mile markers" are passed, the TC reaches +3(@4D) while the RC remains stagnant. In fact, as the 3D marker is reached the TC moves up to +4, +6 at 2D and finally +12 as we approach the cut card. Therefore, we have just witnessed an exponential movement in the TC without the RC changing since the first deck.
Another curious observation regarding the true count I have yet to mathematically evaluate is that which has to do with "deck saturation". Let me explain.
Let's take the last remaining deck of a shoe. The value , by calculation is known to be +2 TC and it's composition is 27H and 25L. The TC of another remaining deck is identical, i.e. +2, however it's composition is made up of 18 7's, 18 8's, 9 H and 7 L. Even though the TC's of these two decks are equal, something tells me that my expected value would be significantly greater playing one rather than the other, since one deck is "saturated' with neutral cards while the other is not. Any insight here?

 

[k_c]

Let me first begin with an example of an interesting observation of the relationship between the true count and the running count. Take a fresh 6 deck shoe and let's say that 27 low cards are dealt, 15 high cards and 10 neutral cards. We now know that the sum of total cards dealt at this point equals 52 (1 deck) , the RC= +12 and the TC= +2.4. Now let's suppose that the RC stabilizes form this point until the last 1.25 decks. In this scenario we see that there is absolutely no change in the RC; however, as the shoe is dealt down and the deck "mile markers" are passed, the TC reaches +3(@4D) while the RC remains stagnant. In fact, as the 3D marker is reached the TC moves up to +4, +6 at 2D and finally +12 as we approach the cut card. Therefore, we have just witnessed an exponential movement in the TC without the RC changing since the first deck.
Another curious observation regarding the true count I have yet to mathematically evaluate is that which has to do with "deck saturation". Let me explain.
Let's take the last remaining deck of a shoe. The value , by calculation is known to be +2 TC and it's composition is 27H and 25L. The TC of another remaining deck is identical, i.e. +2, however it's composition is made up of 18 7's, 18 8's, 9 H and 7 L. Even though the TC's of these two decks are equal, something tells me that my expected value would be significantly greater playing one rather than the other, since one deck is "saturated' with neutral cards while the other is not. Any insight here?

There are a lot of possible subsets for a 6 deck shoe where 5 decks have been dealt and the running count is +2 (and therefore TC = +2.) All of the subsets are not equally probable, however. The image shows the probability of drawing each rank calculated from a weighted average of all of the possible 52 card subsets with a HiLo running count of +2 dealt from a 6 deck shoe. i.e. the more probable a subset, the more influence it has on the final probabilities.

 

[iCountNTrack]

I agree it is confusing, but I argue it is not unnecessarily confusing.

There are selected instances when retrospective advantage is useful, for example, in a mathematical analysis of the advantage gained from Wonging.

A player who Wongs out actually benefits in two ways. One, by not playing low counts, he shifts the distribution of counts played towards the high end (prospective advantage). In addition, the shoes in which the count drops may have been played at an advantage (retrospective advantage). While this does not change the playing decisions (which, as you point out, are entirely based on prospective advantage), a calculation of his total EV should include the fact that some of the shoes he Wonged out of were actually positive EV shoes given that the Wonger played the positive end of the shoe and then left.

Example: A 6D game in which the count dropped from TC 0 to -2 over two decks is actually favorable for the player. The player has played the equivalent of two decks at a +4 count.

The contrary is true as well - for shoes in which the count rises from 0 to +2, while the player has a prospective advantage in future bet sizing, the player retrospective advantage has been worse than anticipated because a lot of low cards have come out. In the absence of Wonging or bet scaling, the two effects are symmetrical and cancel each other out. However, as one varies bets and/or Wongs, these effects are non-zero (but small).

A situation is NOT advantageous unless we know it is advantageous from information we have. Keyword is information we have, so in the example you gave only we only know that TC=0 with only that information in hand, we identify it as non-advantageous situation. If we were shuffle-tracking and had some ADDITIONAL info that the the coming slug is rich in high cards, we know then that the situation is advantageous. That what I always tell people is that a good shuffle tracking strategy will give more information that will enable you identify more advantageous situations.

 

[sagefr0g]

quote from bjbob
Let's take the last remaining deck of a shoe. The value , by calculation is known to be +2 TC and it's composition is 27H and 25L. The TC of another remaining deck is identical, i.e. +2, however it's composition is made up of 18 7's, 18 8's, 9 H and 7 L. Even though the TC's of these two decks are equal, something tells me that my expected value would be significantly greater playing one rather than the other, since one deck is "saturated' with neutral cards while the other is not. Any insight here?

There are a lot of possible subsets for a 6 deck shoe where 5 decks have been dealt and the running count is +2 (and therefore TC = +2.) All of the subsets are not equally probable, however. The image shows the probability of drawing each rank calculated from a weighted average of all of the possible 52 card subsets with a HiLo running count of +2 dealt from a 6 deck shoe. i.e. the more probable a subset, the more influence it has on the final probabilities.

i tryed to set up bjbob's scenario with tdca.
set up 90%pen, 6deck, s17daslsr
results seem strange?:whip:

 

[Canceler]

Yes, and no...

results seem strange?:whip:

From my previous playing around with k_c's program I knew that the "neutral" cards had quite an effect on the EV. You've just shown that rather dramatically, as well as proving bjbob's point!

 

[k_c]

i tryed to set up bjbob's scenario with tdca.
set up 90%pen, 6deck, s17daslsr
results seem strange?:whip:

A few points:
1) The compositions that were input are perfectly valid compositions but they are far from a reasonable representation of what the average composition for a +2 count with 52 cards remaining is. The point of my other post was that not all compositions are equally probable. Yes extreme compositions are taken into account in the image of my other post but if their likelihood of occurrence is very small they lend very little weight to the final average comp. A far more likely composition is input into the first image attached to this post.

2) tdca presently uses total dependent basic strategy to compute overall expected values. I input the same comp you did with the large number of 7s and 8s into cdca and computed using best possible strategy. Overall EV using td basic strategy was about -20% for this comp whereas using best strategy overall EV is about +20%. This is shown in the second image attached to this post. It would take super human player to recognize this comp let alone play it perfectly, though, in the unlikely event that this comp is encountered in the first place.

3) Just multilpy the probability of each rank in the image from my other post by 52 (because there are 52 cards remaining) to get the approximate comp of a +2 HiLo RC dealt from 6 decks with 52 cards remaining. From this you can get a fairly reasonable representation of of what the comp would be for this HiLo count dealt from 6 decks at this pen.
.07295*52 = 3.7934 = (average number each of 2,3,4,5,6) [~19 total]
.07736*52 = 4.02272 = (average number each of 7,8,9) [~12 total]
.32255*52 = 16.7726 = (average number of ten value cards) [~17 total]
.08064*52 = 4.19328 = (average number of aces) [~4 total]

4) I wrote a program to compute average comp for a given HiLo or KO count. A reasonable alternative is to just assume that each 0 card (7,8,9 for HiLo) accounts for 1/13 of the total cards. That will leave 10/13 of the total cards that may be high or low. A zero count will have 5/13 high cards and 5/13 low cards. The high and low cards should be distributed as evenly as possible. The percentage of high or low cards can be changed to create a plus or minus count but the sum of the percentages of high and low cards present should be a constant 10/13. If there are too few cards present it becomes difficult or impossible to do this.

 

[sagefr0g]

k_c, what is the hi/lo composition enumerator program in this link: http://www.blackjackinfo.com/bb/showpost.php?p=118871&postcount=31 ?
i didn't see it available on your site.

 

[k_c]

k_c, what is the hi/lo composition enumerator program in this link: http://www.blackjackinfo.com/bb/showpost.php?p=118871&postcount=31 ?
i didn't see it available on your site.

I haven't made it available. It's a program I wrote that computes the probability of drawing any given rank for any given HiLo or KO running count at any given penetration for 1 to 8 decks. It considers all of the possible subsets that result in the given RC/pen. The probability of drawing any given rank is different for each subset and each subset has a different probability of occurring so each subset is weighted according to its probability to get the overall probability of drawing each rank. Additionally it can optionally consider any number of specific removals in the calculation.

Its ultimate application could be to compute EVs based solely on a HiLo or KO count but it would be way too slow to be practical. I've used it to compute insurance indices for HiLo and true counted KO though. In order to determine whether or not insurance should be taken all that is needed is to know that probability of drawing a 10 is greater than 1/3 given the RC/pen/optional removals.