During today’s Mets-Rockies game, Mets play-by-play man Gary Cohen said something that I found both confusing and – because I’m thin-skinned about this stuff – obnoxious. To very closely paraphrase:
“Nowadays, there’s so much more of an emphasis being placed on statistics and statistical analysis when putting together a team. You see more and more front offices embracing that way of doing things. Which is fine; there’s a certain place in the game for that. But the more you actually watch the game, the more you realize that statistics don’t win games.”
(color commentators agree, discussion on the importance of intangibles ensues)
I’ll tackle the confusing part first. Namely, I’m not sure I understand what point Cohen is trying to make. I would guess that he’s trying to say that you cannot rely solely on statistical analysis when evaluating individual or team performance, that there’s more to constructing a winner than high batting averages and low ERAs. This is, of course, true. Exclusively statistical analysis would suggest that the New York Yankees have a real prospect on their hands in Shelley Duncan, but anyone who has seen Duncan play in the majors knows that he is – at best – a bench player on a second-tier team. It took me forever to be able to admit this, but it really and truly takes the careful combination of objective (statistical) and subjective (scouting) evaluation to identify major league talent and assemble it effectively.
I’m almost positive Cohen was trying to endorse this balance. The problem, however, is that he actually said nothing like that. He said that “statistics don’t win games,” which is about as wrong as you can get. Baseball teams win and lose games based on the number of runs they score and allow. Runs themselves are a statistic, which should automatically disprove Cohen’s theory, but I’ll continue. Teams like Cohen’s Mets score runs (well, not these Mets) by hitting singles, doubles, triples, and home runs. They prevent runs by accumulating strikeouts, avoiding walks, and inducing put-outs. These, too, are statistics. Statistics represent events that determine the outcome of a game. Sure, whether or not David Wright thinks Angel Pagan (great name or greatest name?) is a raging jerk might affect Wright’s performance, but it remains inevitable that his play – as documented by statistics – will affect the level of his team’s success. As I hope you can see, Cohen’s thesis statement is totally incorrect.
In addition to the content, I also found Cohen’s tone more than a little obnoxious. More specifically, his condescending “the more you actually watch the game” rubbed me the wrong way. As many of you may know, a common stereotype amongst the old-school baseball contingent is that those advocating statistical analysis don’t actually watch the games themselves. Instead, it is usually implied and often said that we watch games through the box score, or perhaps in some Matrix-like alternate reality. I’m very (overly?) sensitive to this implication, but I can’t help how I feel. So, to Mr. Cohen and anyone else who thumbs their nose at advocates of objective analysis, I say this: For every stat geek that evaluates players based on nothing but VORP and SNLVAR, there’s a baseball romantic that judges exclusively on a player’s hustle and the look he’s got in his eye. The ultimate goal is to meet in the middle. Until we get there, however, I’d appreciate it if the subtle derision of the statistically-inclined community for its entirely valid (and often accurate) approach to evaluating baseball came to an end.