>> and in fact will probably make it worse depending on how you choose >> the cutpoint. > I'm pretty sure it won't. Otherwise you'd be lowering entropy by doing > a random thing to a random thing.
Doing a random thing to a random thing usually *does* lower entropy when the "random" things are actually deterministic algorithms that may have unexpected correlations. That's why you don't write your own PRNG unless you have a very good understanding of the math. If you are shuffling the deck with, say, numbers from random.org (which uses atmospheric noise), then cutting the deck afterward will have precisely 0 effect, since the (51 * 52!) possible outcomes include 51 copies of each of the 52! orderings, and so the odds of each end up the same. But if you're choosing the cutpoint by getting a value from the same PRNG you used to shuffle, there might very well be a correlation that makes some arrangements more likely than others. -- http://mail.python.org/mailman/listinfo/python-list
