Most people believe that Galileo's big contribution to science was noting that the Earth revolved around the sun. That's not really true. Astronomers from various cultures had worked that out long before and, even in Europe at the time, the scientific community wasn't exactly shocked by the idea (Kepler was expounding an even better model at the same time).
However, Galileo is rightly considered the "father of science" because he was the one that really moved science away from trying to explain the universe and instead settle for describing it. It took a while for that concept to really take hold, but once it did, Europe immediately vaulted to the forefront of innovation and, as innovators tend to do, took over the world.
Statistics teachers seem to have missed the memo.
This is particularly ironic since if there was ever a field that focused on description over explanation, it would be statistics. Still, stats textbooks are littered with these problems that supposedly debunk commonly held beliefs simply because they can provide a descriptive model that does not require them. Note, I didn't say they can provide a descriptive model that disproves them. Even statisticians are generally careful not to go that far. They just figure that if the random model covers it, the phenomenon probably doesn't exist. QED.
Well, no.
What the hell is "random" anyway. Try pinning a statistician down on that one. It can be amusing. At Cornell, I took a course called "Frequentist Theories of Statistics". When I walked into class on the first day, there were seven other students in the room. Four of them were faculty. One of them had been awarded the Nobel Prize. Really! This is a deep question and even the world's best minds struggle with it. I got an A in that class, not because I had any clue but, as the professor told me when I defended my term paper, I "wasn't afraid to face the subject head on." He then added, "That's pretty impressive for a guy coming from a school with a T in the name."
Then he farted. Really loud. He was super smart about math but rather ignorant as to social graces.
I went on to explain that my choice of a tech school over a university for undergrad was based solely on the fact that my dyslexia would kill me in a traditional literature course; I was quite willing to ponder philosophical questions as long as I didn't have to read 100,000 words a week to do so.
So, what's my point?
We had a problem on the homework for Bayesian Stats that talked about the "hot hand" in basketball; the idea that a player would get on a roll where they seem to be sinking everything. They then presented some stats and we had to show whether there was any evidence of non-random activity. Obviously, there wasn't, or I wouldn't be writing this.
I've seen this problem in various forms in at least half the statistics classes I've ever taken (and I've taken quite a few). I get that they are just using it as a convenient metaphor but, as someone who knows more than a bit about both stats and sports I can assure you that it's the statisticians that are full of shit.
First off, the set of sports data is so vast that you can easily find a slice of it to prove or disprove just about anything. Even without nefarious intent, one can get sucked in by something that appears to be a trend when it's really just noise.
More importantly, it violates Galileo's specification of the scientific method. Experiments only prove things by showing that the data is completely incongruous with competing theories; leaving only the theory to be accepted. Simply showing that the same thing can happen under multiple competing theories (in this case, random versus real variation in success) proves absolutely nothing.
More importantly still, description does not trump truth. Every elite athlete in the world will attest that there are times when your game simply gets elevated. Where everything suddenly seems easy. The standard term is "in the zone." People who have never experienced this have no clue. People who have will defend it with the vehemence of a TV Evangelist. It is a TRUTH. Not a fact.
OK, I've been up way to long this weekend working. But it was fun to rant.
No comments:
Post a Comment