Readers who paid close attention to our probs and stats discussion on Monday will be better prepared to discuss the conclusions of Dr. Bob Rimmer from mathstate.com:
Data scaled by 49 days. Last data point on Tue 16 Sep 2008.
Probability of McCain win in November 0.649889
Probability of Obama win in November 0.447952Here are similar calculations based on a sample of only the last 90 days.
Probability of McCain win in November 0.874503
Probability of Obama win in November 0.272771
Arthur DeVany considers this turnaround in a futures-driven market analysis to be a fat-tailed “Black Swan” event generated by the Sarah Palin pick (H/T to Jonathan at Chicago Boyz for the link).
For those dismayed by these results, take cheer: There are still 48 days and four debates before voters go to the polls. Practically an eternity.
Chart-analysts and astrologists should be granted a free wall-to-wall counseling as part of physical therapy.
I only majored in Math to get a scholarship from the Navy. I never enjoyed it-Statistics was a touch and go battle all the way to the end of the semester. I took my C- like a man. And was grateful for it.
I remember a quote in a college stats book many years ago – “Statistics is like a bikini, what it reveals is interesting, what it conceals is vital.”
any way you slice it, it’s always a 50/50 proposition…. either he will, or he won’t.
skippy: you got a C-? so you would have been a wee bit closer to the top than i.
If you got a D in anything you went on double secret probation with Naval Science Department. After junior year they used to re-inforce that point with pointed reminders of what chipping paint and working in deck department on a ship was going to be like if you flunked out.
Skippy, don’t forget the dreaded cleaning out the sanitary lines in the female heads
Excuse me for being dense, but am I right in assuming that a probability of 1.00 means it’s gonna happen, like in the sun coming up tomorrow? ’cause if I am right, then why do the two possibilities added together equal more than 1?
I understand math, but only up to the point where it describes something I can touch or see.
Oops, never mind. Read a little closer on the link, which still looks slightly like Greek Voodoo, and found this.
“There are some problems with the prediction market price model, first, prices have an practical upper limit of 100. Second we see that the parameters of our two distributions are quite different and the probabilities don’t come close to adding to 1. We present a more sophistocated approach to analysis on our Presidential Prediction page. ”
Of course, on the Presidiential Prediciton page, McCain still squeeks out a win. Too bad statistics couldn’t be frozen when you find the answer you want.
Pogue, great line, thanks.
That’s a keeper.
Re:”Too bad statistics couldn’t be frozen when you find the answer you want.”
Wish that was true of the stock market.
I still remember that we have to factor in the good ole electoral college.
Intrade.com has a map of each of the prediction markets for the states. The current map has it 269-269.
I’m confused by all of this. So…… what’s new. I’m still handsome
I’m pretty sure it was Mark Twain who said “There are lies, damned lies, and statistics.”
There are a number of prediction sites that are out there, and it’s too close to call. Most of them have McCain leading, with a number of undecideds.
To me, this is reminiscent of 1980. The race was tight until Reagan and Carter debated, and people saw that Reagan wasn’t that bad. I realize that the comparison isn’t completely accurate, but my take on it is that people are waiting to see how the two of them do in the debates.
For what it’s worth, Reagan and Carter only debated once.