Statistics (Or as we said in Pennsylvania, "Suh-tistics")
November 5th, 2004

Those of you watching the NFL coverage over the weekend may have heard the normally brilliant Chris Berman repeating some gibberish about the Redskins predicting the outcome of the presidential election with astounding accuracy. You can read more about this now-defunct nugget of trivia at Snopes.

Chris, I ask you to please consider the following facts before repeating such nonsense in the future:

  1. In most election years, the outcome of the race can be predicted with better than 50% accuracy (ie, better than a coin flip). In really close years like this one, it’s never worse than 50/50, Ross Perot notwithstanding.
  2. The Redskins, or any other sports team winning any particular game over long stretches is probably around 50% too. Other predictors, like the Nickelodeon mock election, have a better chance since they’re actually trying to be predictors.
  3. There are at least 500 bazillion things you could point to as predictors of the outcome, depending on how far you want to stretch.

Given those odds, it would be really freaking amazing if a few of those predictors did not predict the outcome for long stretches of time. I tried, in a very passionate yet fruitless attempt, to explain this to my boss. “No raise for this guy,” is what I’m sure he was thinking.

One other stat I’d like to point out is the odds that nobody will listen to me and repeat another inane predictor in 2008. I think that this may be infinitesmially close to 100%.

As for me, I voted for Roethlisberger.