Is “wet bias” a bad thing?

“Bias” tends to have negative connotations. It’s the “wrong” answer.

The problem here is a translation issue. It’s going from the world of One Answers (mathematics) to the world of Many Answers (life).

Weather is a fascinating demonstration. Nate Silver writes in the 2020 edition of The Signal and the Noise, “The further you get from the government’s original data and the more consumer facing the forecast, the worse this bias becomes.”

Relatedly:

(John Gruber) “I staunchly believe that Fahrenheit is the better scale for weather because it’s based on the human condition. Who gives a crap about what the boiling point of water is, it’s the most ridiculous thing I’ve ever heard in my life.”

(Ben Thompson) “The other thing is that Celsius is not precise enough. In the car it adjusts it by point-five because a single degree of celsius is too much for the car. Fahrenheit is more finely grained in a positive way.”

This is why we have a wet bias. We design weather for people.

Silver again, “It’s deliberate and it has to do with economic incentives. People notice one kind of mistake, the failure to predict rain, more than another kind, false alarms. If it rains when it’s not supposed to they curse the weatherman for ruining their picnic. Whereas an unexpectedly sunny day is taken as a serendipitous bonus.”

One change in my thinking over the yeas has been to reframe ‘bias’ as ‘tendency’ and then consider what’s happening. Humans are only illogical in the game of optimization, which matters in the world of calculations rather than considerations.

Wet bias may be inaccurate but that doesn’t make it wrong.