It’s a Lock
November 4, 2006 Leave a comment
The lead article on the ESPN football page is teased right now with the following paragraph:
Our experts don’t agree very often. They don’t even agree to disagree. So when they unanimously select five games on the Week 9 schedule, consider those picks as good as W’s.
Well, that sounded like bullsh-t to me, so I went back and looked at the first eight weeks of ESPN “expert” predictions. I counted 47 games on which the ESPN “experts” all agreed, or just under six per game, which means that five unanimous predictions are not at all unusually great, as the lead-in implies; in fact, they’re a bit less than usual.
I then counted how many times the experts got the games right, which turns out to be 33. That’s a 70% success rate, which means there is only a 17% chance that those games are “as good as W’s.” In fact, the “experts” are almost sure to be wrong on one or two of these games.
Just more proof that even “experts” can’t accurately predict the outcome of one game, even if it’s “a lock.”
Update: Experts went 2-3 that week.