A passage from a book that I wanted to preserve for the record here, because I think it's valuable.
Bill James. If you don't know James, he's the one who first really started promulgating the idea of seeking answers to things within sports data (baseball in his case). Suffice it so say - I would not be doing much of any of my professional items right now, the PhD in mathematics, the career as a health actuary, the goaltender history work - without Bill James leading me there from a very young age.
Anyhow, I was re-reading the 1982 Baseball Abstract. I found this passage insightful; I thought others seeking to do the sorts of things in this chapter might as well.
"In the same way that a monkey wrench or a hammer or a screwdriver or any other tool can be used to tear things apart or to render them inoperable, [sabermetric tools], if used carelessly, will do more harm than good, will lead to false conclusions. But they give you a chance. A monkey wrench is not a guarantee that you can fix anything; it is merely a way to get to the problem. Without it you have no chance. These tools enable you to take baseball games or teams apart.
No sabermetrician has ever discovered anything of interest by compiling large stacks of numbers and shaking them vigorously to see what happens to fall out. One thing that interviewers like to ask me sometimes is 'What is the most amazing thing that you have ever found in all these statistics?' I never know what to say. I don't 'find' things in that way. What would a mechanic say if you asked him what was the most amazing thing that he had ever seen in an engine? It seems safe to assume that, whatever it was, it didn't belong there. Anyway, there are sabermetricians who do begin with the numbers, certainly; there are people who 'analyze' the game of baseball by correlating anything with everything, who spend their time frantically rushing from one column to another looking for connections and running regression analyses till hell wouldn't have it. This is the sabermetric equivalent of kicking a television set. If it is done with intelligence, it becomes the equivalent of kicking the television set vigorously. If it is done with persistence, it becomes the equivalent of kicking the television set repeatedly. If your TV goes on the blink at a crucial moment, you may derive a certain amount of gratification from kicking it; if baseball mystifies you, you may derive some satisfaction out of correlating things willy-nilly, running regression analyses and making up more and more ways to rate the hitters. But it is not going to fix the television set. Unless you get unusually lucky."
