I got a message saying that this didn' post because I mistyped the address:
Warren said: he underlying theoretial attack is exactly that suggested by Mike Ossipoff for his "bias free Webster" method I reply: Wait a minute--that isn't the name of any of my 4 methods. I have Cycle-Webster, Bias-Free, and Weighted Bias-Free (BFW). BFW takes into account the state-size frequency distribution. Warren continues: , except that the underlying probabilistic model is now an exponential distribtuion not a uniform "distribution" I reply: Ok, you must be referring to Bias-Free, which, of my four methods, is the only one that assumes a uniform distribution. (I use the word in quotes since Ossipoff has in various ways ignored the requirements of probability theory, e.g. in his recent attack on the idea that probability distributions need to be normalizable I reply: A probability distrilbution can take all sorts of forms. For instance, a fair roulette table has a uniform probability distribution, over the numbers that could come up. Yes, some common naturally-occurring distributions involve exp. But the roulette table's uniform probability distribution is still a probabililty distribution. Someone could devise a program so that the screen-printing of a number from a set of numbers has a probability distribution of any shape that he wants to give it. As I said, B/(q+1) can roughly approximate the density function of states over the range of populations. It won't duplicate it exactly. As I said, that's all I claim, and that's all I propose to use it for. You seem to be having a problem about that for some reason. I'm going to send this before this computer loses the Internet. Mike Ossipoff ---- election-methods mailing list - see http://electorama.com/em for list info
