See my other post...
It answers your question. Very little work has been done on squared skew distributions, but latency is ALWAYS there.
It answers your question. Very little work has been done on squared skew distributions, but latency is ALWAYS there.
If a player has slightly decreased the casino advantage then they have also slightly decreased variance, no?
Don demonstrated this is not a valid assumption to make, that decreased casino advantage does NOT mean decreased variance. Variance is how widely something varies. For instance, a video poker machine with 100% fullpay has no casino advantage (with optimum play) but has very large variance, due to large royal flush payouts and player losses while waiting for a royal. Don's example was the double after splits rule in bj, which decreases casino advantage while increasing variance.
See? Variance and Advantage are completely separate things; they do not depend on each other.
There's no reason to think a CSM decreases variance--there could just as easily be chunks of high cards or low cards, it's just that we have no idea when they are coming. But then a non-counting BS player doesn't know when they are coming in a shoe, either.
JV
high count - we still gain, whether or not it is not as much as it seems
medium count - we still gain
low count - we lose a little
No, here's about how it works for a bs player. These numbers are just guesses, for demonstration purposes.
EARLY in the shoe:
high counts (25% of the time): advantage 1.5%
medium counts (50% of the time): advantage -0.5%
low counts (25% of the time): advantage -2.5%
overall advantage: (25(1.5%) + 50(-0.5%) + 25(-2.5%))/100 = -0.5%
LATE in the shoe:
high counts: advantage is less than normal, say 1.0%
medium counts: advantage is more than normal, say 0.0%
low counts: advantage is less than normal, say -3.0%
overall advantage: (25(1.0%) + 50(0.0%) + 25(-3.0%))/100 = -0.5%
Overall advantage works out the same either way.
JV
The CSM's I've seen take the cards played and insert them into the machine. This means these cards cannot be dealt in the next immediate hand. Therefore, if I'm correct, this clump of cards just played influence the probability of the immediate next hand. Whether this is significant depends, of course, on a number of factors; eg. number of cards dealt; value of cards dealt; number of decks in play; etc.Am I off base?
Of course it's so easy to change scenarios. I can easily create a scenario in which a machince that pays out 100% will have a much higher variance than another exact replica that pays out a 100%.
The difference is one machince pays a huge jackpot every 10 years where the other pays a small jackpot every 24 hours, get the picture?
Ok, now lets stick to apples and apples where the same machinces pay out every 24 hours but one machine is set at 99% and the other 100%.
All bets are the same amount.
Which machine do you think will have the higher variance?
Are there any patterns in the way cards return to play - could the machines be sequenced?
The Bishop made some comment to that effect recently. If the machine was riffling perfectly then the card order would be completely deterministic and the advantage player would clean up-it would be like playing a cooler. To be honest it never occurred to me that the designers could be that stupid and it is not something I have ever looked for.
Your other questions are mostly covered by the csm article on my homepage. That article does not give figures for the various different machines-that would be very difficult, since many brands exist, which are constantly being updated, and even have settings to alter their behaviour. It does however, give a general heuristic for determining the variables you are after.
Increasingly I suspect, as you clock these machines, you will discover that they cannot be beaten. But at the current time, only the King comes close to invincibility with regard to the exploitation of card latency, and some have suggested convincingly that other approaches may be successful with that device.
If the makers found out, they might be able to change the inner workings without any outside clues.
You might sit at a game thinking you know what's up all the while the machine is working differently.
With statistical analysis you would, effectively, know the workings of the machine. I don't think people really get how powerful the information inside a stats textbook is. You would also be able to adjust to changes in the fundamental workings of the device (roulette bias players have learnt to deal with this sort of thing for over a century).
You can't ask for better confirmation than that.
I would suggest, for purposes of simulation, that a usable model for CSM's might be a random insertion model. In this model, we start with a randomly shuffled shoe, and after every round, the used cards are inserted back into the shoe at random locations. This would be easy to simulate by computer, and, I suspect, might be close enough to the actual workings of these beasts to be of some value.
I don't think this would differ substantially from the "random shuffle every round" model for Basic Strategy players, but it there would be some latency effects that might be eploitable for advantage players. One of my concerns about CSM's is that, assuming a random insertion model, is that is the casino were to load the machine with more 10's and aces at the bottom, it might take quite awhile for the 10's and aces to get into play. This might give the casino a considerable extra edge for many rounds, and would be very hard to detect. It might even be possible to devise a shuffle procedure with new decks that would accomplish this "preferential loading" without being immediately obvious to the dealer performing the shuffle or to regulators viewing tapes.
In the time I have been reading this board, I have seen many posts by people who seem to believe that playing the last deck of a six deck shoe is just like playing a single deck.
Besides, if this were true, you wouldn't have to wait the last deck. If you have the cut, just cut the bottom deck to the top and leave after that's been dealt. :-)
he performed exhaustive mathematical studies, using his considerable abilities, to come up with his illuminating comment, which has helped us all to better understand the intricacies, of the game. :-)
Don
Why do you say that BS wins slightly more with CSM's?
