Thought I?d post some simple shuffles to show some of the effects of dealer sloppiness in shuffling. First, three shuffle maps. The shuffle is a very simple one pass, riffle & restack with four grabs. First we show the mapping of the cards given a robot dealer. That is perfect grabs and riffles. The left stack of cards is pre-shuffle and the right post-shuffle. I have painted the cards in four colors:



It?s easy to see how the very basic concepts of shuffle tracking work in the above example. If you have knowledge of the count in a zone before the shuffle, you have knowledge of a zone after the shuffle. However, dealers are not robots. There is much variation in grabs, splits, riffles, etc. Below is the same shuffle with a human dealer.



Clearly the information is less perfect. And clearly it is still of value. Now let?s look at a dealer on drugs:



With such a simple shuffle, there is still plenty of information available after the shuffle. Let?s look at all three:



We can see that the tracked cards are more spread out the sloppier the dealer. But still not scattered randomly. We can get a better view by isolating one section of cards:



Now this is a simple shuffle. Let?s try a much more complex shuffle:



Real mess huh? Yes, but that doesn?t make it un-trackable. You must narrow the zones considerably. And you can see where some groups of cards are not spread too wildly. This is easier to see with colors isolated.

Why have I posted these? Well, a post on BJFO this week claims that CVShuffle cannot create these maps. That it only creates maps of robot dealers. Clearly this is incorrect. (One of dozens of false claims about software on that site.) Also because they?re pretty. But I also wanted to point out two other misleading posts in the same thread. One claims that only single-rifles can be tracked. In fact, a quadruple riffle would make little difference in these maps. The other post suggests that the gain in tracking is so large that risk can be ignored and there is no need for simulation. This is a very dangerous belief. In the case of the first, very simple shuffle, there is a very large gain in SCORE from tracking. However, you are not going to find this shuffle with good rules. So, you are doubling a lousy SCORE. That does not give you a high enough SCORE to ignore risk. (Is there a high enough SCORE to ignore risk?) In the case of the complex shuffle; yes you can find it with good rules. But, you will not be able to realize a gain on every shuffle and your overall gain in SCORE is again too low to not understand your risk. Unless you have an infinite bankroll, it never makes sense to ignore risk.