Feel free to do the math out. It has nothing to do with "eliminating" combinations. Read more

We don't mind actual answers! Read more

It's not saying that they'd get the first season's results right, because you're right, that wouldn't make any sense. If the odds of getting any bracket wrong in a single attempt are (128 billion - 1)/(128 billion), then if there were 91 billion attempts (13 years * 7 billion people), the odds that every single one… Read more

I think you can have it both ways. This comic (still funny!) makes a much stronger point about data interpretation using a much more realistic example: http://xkcd.com/882/ Read more

Yeah they're having some trouble. Read more

I disagree—making a "reasonably informed" bracket doesn't mean picking the higher seed every time. 18,583 out of about 3 million Yahoo! brackets had FCGU in the Sweet 16 last year. I doubt that all of those picks were join coin-flips or equivalents. Read more

Randall's point is well taken is well taken, but I think he could have found a more true-to-life example than a detector that rolls two dice and lies if they both come up six. Read more

CHALLENGE ENTRY BRACKETS WILL ONLY BE ACCEPTED IN THE CHALLENGE FROM THE FIRST FIFTEEN MILLION (15,000,000) ENTRANTS WHO COMPLETE STEPS 3(A)-(C) BELOW (THE "REGISTRATION PROCESS"), AS DETERMINED BY YAHOO! INC. ("YAHOO"). Read more

9.2 quintillion was widely reported in previous years as well, because it's simply 2^63. That it's in the the official contest rules this year doesn't mean much to me. It's really the only figure they can report, because anything lower requires some probabilistic estimates that aren't typically used in these sort of… Read more

It's a very hypothetical example, to illustrate the enormous gap between 1 in 128 billion and 1 in 9.2 quintillion. You could also point out that many of the 7 billion people on Earth are babies. Read more

That means that, in previous years, Southern Miss's profile would have given them a 97 percent chance of making it. That doesn't necessarily mean that this was a strong bubble year, but it might. Read more

No, just that the brackets are independent. Read more

I usually like XKCD, but this scenario is so ridiculously contrived that I think it kind of ruins his point. Non-Bayesian statisticians still use common sense. Read more

No, but remember that that number also assumes that everyone on the planet fills out a bracket. Read more

We talked about this in another thread. It's more like they've gone 13ish for 15ish over the last three seasons in terms of actually tricky bubble teams, which is pretty good. Read more

I thought that was weird too, but it looks like you don't have to pick them to win the billion dollars. Read more

Very fair point, this is an issue with election results as well. I would say that you're underestimating the uncertainty at the edge of the bubble a little bit. Read more

It had Drexel getting a bid in 2012 instead of Iona. Read more

I was working in urban planning, but kept a shitty little side-blog where I made maps about sports. I picked a fight with a town in Montana, made it onto Deadspin, got a part-time internship which turned into a part-time fellowship, and then got hired full-time and quit my other job. Read more

TICK TOCK Read more