One of the things I do on a regular basis is check the weather forecast.  Will it be warm enough to leave my windows open today?  Do I need an umbrella?  Will the sky fall on my head?  Etc.  But checking the forecast means I'm relying on a prediction of future events, and unless you're Steve Jobs, your predictions are occasionally wrong.  It's ok to be wrong, but three pieces of information would help me understand things better: 
  1. Who was wrong?  Was it a specific meteorologist, or a certain atmospheric model?
  2. How often?  If Mike the Meteorologist is right 30% of the time, and Bob the Meteorologist is right 60% of the time, I'd probably put more faith in Bob.  Even though Bob has an annoying voice and can't hold his liquor.
  3. How wrong?  If the predicted temperature was 75°F but it's actually 70°F, that's better than a predicted temperature of 35°F versus an actual temperature of 30°F.  Yes, they're both 5° off, but the first prediction is 7% off, while the second is 17% off.  That's a fairly significant difference.
Ideally, you could combine all those assessments into a single number, so any prediction of future weather events would come with a percentage probability of it actually happening.  That way, when they say it's supposed to snow next week but it turns out to be sunny and hot, you can simply look at the probability and realize something like 5% isn't a very confident number. #nature