Chance of rain 'always' correct

Cecil's Columns/Staff Reports
1

Cecil,

How careless to have opened yourself up with such absolutes as “On the question of how reliable the figures are, the amazing truth is that they are absolutely 100 percent reliable all the time”. Not to be nitpicky, but I’m a nitpick. The historical data will be accurate, in reference to history. But as far as reliability goes for “today”, always is definitely wrong. For any day under the current conditions where it either has rained 0% of the time, or 100% of the time, you could see a first. The first time you either get rain, or don’t get rain, under “today’s” historic conditions, the prediction will be incorrect, hence not always reliable.

Dave

http://www.straightdope.com/classics/a2_115b.html

2

I think you missed Cecil’s point. Even if the figure was reported as 0% and then it rained, the figure was still correct and “100% reliable” and everything Cecil said, because as he pointed out, it’s only a measurement of past conditions, not a prediction of future conditions. Cecil didn’t say the figures always predict the weather correctly, just that the figures are correct as is. Not very useful, practically speaking, since it’s basically just a semantic game, but still.

3

Reliable and valid aren’t the same thing.