Osama, he dead.

Chuck, the notion I’ve been exploring — I’m more than a bit of a Bayesian — is that we could better look at this as samples of a normal distribution where the prior probability is normal with the mean being exactly 50/50 Bush/Kerry, and with a very narrow variance (because we’re actually predicting a fixed number), and then apply each poll’s mean and variance to improve the posterior distribution.

That would be a Kalman filter, no? There is the question of initializing the filter, in this case the variance of the initial 50/50 split. The other problem is modeling the dynamics for the time update. For that I would probably let the percentages remain constant and throw in some process noise to allow drift; in practice that would act a like a moving average.

Yeah, I like the Bayesian approach when the uncertainties involve lack of knowledge on the part of the observer. For physics I prefer the classical interpretation: entropy is a property of the object, not the observer. Jaynes got himself into all sorts of trouble pushing the latter point of view.

Lindenen, I think the thing is that UBL hasn’t been proven dead. If, as I think John© is suggesting, UBL is a monomolecular film of oxidizing lipids on some newish rubble, it’d be very difficult to prove him dead.

But the longer we go without seeing pictures of UBL holding a recent copy of the LA Times, the more we have to think he’s probably dead.

John, I agree with you about the amrkets, and — now that I’ve seen some comparisons of the different market’s rules, I agree the IEM is probably best of the bunch. My point was that the RCP is getting a lot of attention now, methodological problems or no. (And, to be fair to RCP, at least they’re making some attempt, even though flawed, to make a better inference than the usual MSM one.)

Chuck, the notion I’ve been exploring — I’m more than a bit of a Bayesian — is that we could better look at this as samples of a normal distribution where the prior probability is normal with the mean being exactly 50/50 Bush/Kerry, and with a very narrow variance (because we’re actually predicting a fixed number), and then apply each poll’s mean and variance to improve the posterior distribution. Just noodling around, it looks like the result is, as I said, something between 3.5 and 5 percent in Bush’s favor.

Like you, I don’t give much credit to voter registration drives – I don’t believe they will add particularly meaningful numbers to actual turnout.

People who couldn’t be bothered to register to vote until a volunteer got in their face while they were shopping aren’t likely to do anything different than they’ve always done which is, of course, fail to vote. They have always been, and remain, typically unaware of the location of their polling place, will find some inconvenience that prevents them from getting there, will see a line and vow to return later rather than wait, etc. Many of the legions of college kids signed up failed to submit their absentee application or will fail to send off the actual ballot.

The only places where GOTV movements will have any impact is precisely where their impact is most diluted – the deepest Blue sections of the Bluest states. They may help pull the popular vote in the desired direction to a small degree, but they won’t effect the EC in any real way. All JMO, of course.

During normal times, it is absurd for pollsters to delude themselves about their results. They ultimately earn money only for being accurate. Being wrong hinders oneís marketing ability to prospect for new customers. Sadly, though, we do not live in normal times. The anti-Bush hysteria has reached the point that I strongly suspect that a number of liberal pollsters are going crazy.

Someone (and I forget who) recently suggested that a number of these professional pollsters who support John Kerry may be premising their results on the huge sign up of newly registered Democratic voters. If so, there a serious flaw in this line of reasoning. The reality is that a good size number of these recently registered citizens—may not bother to show up on election day! Only those who passionately adore their candidate will almost certainly vote. The more indifferent often find some excuse not to take the time. How many Democrat inclined voters consider Kerry to be their secular savior? My guess is that the figure hovers around a mere 40%.

Alas, that is because we will never likely find the body. It probably was destroyed. There may not even be an DNA evidence left. In other words, the mystery may never be completely solved. But we are still able to employ our common sense. It would be foolish for Osama bin Ladin to remain silent. His marketing value would be greatly increased by making occasional video appearances. Try imagining the impact of Osama holding up a newspaper dated a week ago.

But what happens is what happens and come Nov 2 we will know. At least I hope we do.

And as for Andrew Sullivan why doesn’t he just pack up and go over to Afhganistan and do a little first hand reporting if he thinks he is such a hot shot on the subject? Christ, people like Kerry had a decade to get this son of a bitch and could not bother to show up to work and now some snarky little dick like Sully thinks he is an expert.

disgusting.