Systemically Underestimating Uncertainty

One of Nate Silver's big things as far as the science of modeling is concerned is that he thinks people have a tremendous tendency to underestimate uncertainty. People will say, you know, "The Republicans are guaranteed to hold the House," but the truth is that at this point the margin of error on a forecast for the 2012 House results has to be ±50 seats. Heading into last year's election day he said that his model predicted that there was no way to call the result within 5 seats either way and have a better than even chance of being right. Well, I think I'm certainly seeing this phenomenon in how people discuss the Mets' chances in 2011. The consensus seems to be that, if the Mets are lucky, they'll win 80-85 games. That implies that people think they'll win 75-80 in a more realistic scenario. First of all I'd maintain pretty emphatically that I don't think these Mets are worse than last year's Mets, so I think people are being overly pessimistic.

But say I'm wrong. Say I'm being my usual optimistic self there, and we really can't count on Reyes or Beltran or Bay to be healthy, Pagan, Thole, Davis, Pelfrey, Niese, or Dickey to reprise their great 2010 performances, or Emaus, Young, or Capuano to be useful additions. Say the 50th-percentile result for the Mets heading into 2011 is really something like 78-84. I still think that to say that the best-case-scenario 90th-percentile result is 85 games is just plain wrong, along the lines suggested above. Maybe we can't count on any of those things. But isn't it possible that the team will stay healthy? And that we'll get a combination of career years and somewhat disappointing years that averages out to a little bit above everyone's career averages? Because if we do, this team will win 90 games. Maybe that won't happen, but I don't think it's a possibility you can dismiss this early in the season. You just can't know.