Yesterday, I was watching the Washington Nationals vs Chicago Cubs baseball game. During the broadcast, an interesting conversation about sabermetrics (that is, the empirical study of baseball statistics) came about. One of the commentators mentioned the phenomenon of shifting (adjusting the defensive positions of players in order to greatly increase the opportunity of making an out) and, in a related fashion, the virtual disappearance of sacrifice bunting. The conversation lead into a broader discussion of sabermetrics and how effective it really is. One of the commentators pointed out that same sabermetric-heavy teams, like the Oakland A’s or the Red Sox under Ben Charrington, had very little success whereas other teams/managers that rely less on measurements have enjoyed more success.
Before I continue on, I’d like to mention that I have not verified the claims the commentator made. I have no idea if the examples he cited are norms or exceptions to the rules.
There does strike me to some truth to what the commentators were saying. Ben Charrington was the Red Sox general manager from late 2011-2015. He is very much a Sebermetrics guy. During the course of his tenure, the team saw 1 Word Series (2013) and 3 last-place finishes (2012, 2014, 2015). Numerically, all of the teams he put together were, on paper, world class teams. But, in each of the losing seasons, chemistry issues, management issues, and (as in the case of the 2014 team) copious amounts of bad luck came into play.
Baseball is very much a game of details. Some of the best managers (like Don Zimmer, Joe Maddon, Terry Franconia, and Dusty Baker) rely on statistics, but also very much on their “gut,” or what Hayek might have referred to as “local knowledge.” Dusty Baker (who currently manages the Washington Nationals) knows his players better than any statistician. He knows their mood on any given day at the park. He can read their body language. The statistician may argue that (making up numbers here) Bruce Harper hits .400 in day games after a night game when a left-handed pitcher is on the mound, but if Baker sees Harper is distracted, he may opt to replace him in the game.
This local knowledge is important to fielding a championship team and winning games. Statistics are important too; none of this is to discourage sabermetrics (I personally am a huge fan. I used sabermetric analysis in my undergraduate thesis). But men are not machines. Without application of local knowledge, these mathematical models will fail. This is true of baseball and of the economy.