On Sat, 7 Feb 2004 Michael Perelman wrote: > The electability phenomenon has the potential to allow small bits of > information to cascade into huge consequences. It also means that when > people learn about the problems with Kerry, they will be more easily > discouraged.
No on both counts. You seem to be making Kinsleys' mistake, i.e. mistaking this for a stock price system, where (a) the "voters" are competitors and (b) the price can change at any moment -- and people have an interest in it changing. Here we have the convergent system with cooperative actors. Everyone is trying to agree on one candidate, and everyone knows the sooner he is picked the better, and everyone knows there can only be one. The cascade functions to rush the crowd to one point and to make it more impossible to to change from that point with every passing day, because if winning is the only value to be maximized, the rate at which any other choice becomes senseless accelerates for every rational actor. New information will make no difference. The converse of only caring about winning is we don't fucking care who he is. Info is irrelevant. In addition, the chances are low of there being any new info on someone who's been in public life for 35 years. And was in a very public bare-knuckles drag-out election only four years ago. And is a corpse. Michael
