Sunday, January 22, 2017

Regression To The Mean

Seems like a wet year. Maybe it is, I haven’t been able to find a really good site for monthly totals for the past few years. But part of why it seems so wet is that the last few years have been so dry. Really dry. Exceptional drought dry. So now, even if we get an average year it will seem pretty wet.



 And even with the trend seeming to be towards a hotter and drier climate for our area, we will get periods when it’s not so hot, nor so dry. It’s known as regression towards the mean.

 Regression to the mean is the tendency for exceptional measurements to be followed by measurements that are more average (the mean). An exceptional racehorse may have descendants that are very good, but probably not exceptional like the parent. This just happens because more possible values are on the side towards average rather than away. That’s kind of what exceptional means, those values are the exception. Not the Rule.

 But this real statistical effect can lead us to some faulty thinking. If you toss a coin and get heads several times in a row, we expect that the chances are higher that it will be followed by a tail. But this is just the Gambler’s Fallacy. We expect a trend of some type to be followed by the opposite to kind of even things out. But that’s not true.

 Sure, in the long run, things can even out, back towards average, but you shouldn’t expect to see the universe actively working to put things right. Probability just doesn’t work that way. So just because this year may turn out to be average for rain, or even wet, doesn’t mean we can count on next year being the same.

 So let’s enjoy, and take advantage of, the rain we’re getting, but keep in mind that even in an average year Tehachapi’s just barely outside the definition of a desert. So let’s keep planing for dry.

 The future may not be ours to see (Que sera, sera) but that doesn’t mean we can’t make some plans.

No comments:

Post a Comment