Quote of the Day

[jpmassar]

The odds against that, if the model had been specified correctly, were roughly one tredecillion (a 1 followed by 42 zeroes) to one.

-- Nate Silver, debunking yet another statistician doing perverted things with models.