View Single Post
Old
11-21-2012, 01:26 AM
  #748
TheMoreYouKnow
Registered User
 
TheMoreYouKnow's Avatar
 
Join Date: May 2007
Location: Eire
Posts: 8,484
vCash: 500
Quote:
Originally Posted by jaster View Post
Sabremetrics guys don't have a leg to stand on if you think the leg they are trying to stand on is that the MVP has to go to the sabremetric stats leader. But I don't see anyone saying that. People are just making an argument, and using advanced metrics to support it. Also, I disagree that advanced metrics are arbitrary. In fact, they are the opposite of arbitrary.
Well, you would be correct if the approach was "Well I think Trout should be voted MVP because I believe these stats show he was the best player" but all I hear is "Trout *is* the MVP because these numbers say so."

But of course they are arbitrary, I love statistics myself, I use them professionally, I learned about them to a considerable extent in college. I'm one of those guys who puts "SPSS" on the skills part of his CV. But when I try to get a certain statistical output, I need to make a decision about the input, the computer won't do *that* for me. It'll calculate it for me but it won't decide what is worth what and why that is. That's all me. Bill James needs a book to explain what goes into "win shares" but are you telling me "win shares" came down from God or were the result of James thinking about baseball?

That doesn't mean they are worthless, but that means that at the end of the day James, like anyone else using statistics in an analytical way, had to make a judgment call somewhere in that monstrosity of a calculation. And that judgment call *is* somewhat arbitrary because you go beyond empiricism and into the realm of your own assumptions and guesses. That's fine, that's what social science is, and that can give us valuable insights into many things. But the reality is also that it is somewhat less verified than say a proven chemical reaction we can test and test and test with the same outcome again and again.

That is too often forgotten by the hardcore "stathead" camp who tend to treat sabremetrics as gospel, when in reality the projections made by the stat gurus have been shown as simply wrong by real results quite often. I used to buy Baseball Prospectus every year and really I don't think they were *so* good at predicting either player success or team success, that I would consider an advanced baseball metric as anything *close* to a fact. I think there's value to sabremetrics but too many people get too dogmatic about it.

And as a result people get so angry about it, like you could sense real anger that Cabrera got the MVP when at the end of the day it's just trivia and little else. Social science, its methods and its theories are tempting because they seem to bring order into chaos, give answers to the great questions of the time where otherwise you just have opinions. But it's ultimately an instrument that has its flaws and it's always dangerous when devotees of a particular theory decide to ignore such flaws and declare their creed "fact" when really it's just that, a view, an opinion as well at the end of the day.

Baseball is a game, it's far too trivial really to get into these wars over it, but still so many of them do. Yet baseball is actually the perfect microcosm for showing all the limitations of social science and thus statistics and their analysis. The interaction of everything that happens in a baseball game both within and between players, their tools, the environment ranging from the stadium to the weather, it's all in its entirety far too complex to be grasped by anything devised by Bill James, just like the world at large has so far eluded any attempts at comprehensive explanation by sociologists, economists and so forth.

TheMoreYouKnow is online now