Friday, February 01, 2019
Meta-fail-to-see
Early today, just after writing an item about the frustration of watching people who ALMOST reach a conclusion but fail to see the last step, I posted an attempted debunk of a comment about Groundhog Day.
A stock talker said "Punxie is only 39% accurate. That's worse than flipping a coin."
I had the bright idea that this implied the rule was backwards; that the prediction of the OPPOSITE outcome was 61%. So Punxie is a good predictor but the rule is backwards.
Later I took another look and realized my total failure, and deleted the original.
I should have left it there, because it was actually an even better example of fail-to-see.
I've never been good at the math of probability. I can do the arithmetic of permutations and combinations, but the connection to real problems has always been JUST BEYOND MY GRASP. Sometimes I see it, sometimes I don't. This was true when I was young, and probably more true now that I'm old and fading.
Maybe I should have a little more sympathy for the other fail-to-see. When a point is just beyond your grasp, you don't KNOW that it's just beyond your grasp, because it's just beyond your grasp.
[Why is probability hard? I'm strictly an experiential learner. I master pieces of math by using them for work. Algebra and some calculus for electronics, trig for graphics, symbolic logic for programming. Probability is the math of gambling. I've never made a bet on anything. Not cards or sports or horses or stocks or lotteries. So I've never needed to master the math of gambling.]
Labels: coot-proofing, Experiential education
¶ 3:40 PM
