When questioned about his Oakland A's lack of postseason success, Billy Beane famously remarked, "My shit doesn't work in the playoffs,"
There's a couple messages one could take from this statement:
- The baseball playoffs are essentially a crapshoot, and it's silly to draw any grand conclusions from a team's performance in a 5 or 7 game series. The 162 game regular schedule is a much truer test of a team's makeup. Beane's job was to get the team in position to be in the playoffs, and he did that.
- The A's were constructed for the long grind of the regular season, not the radically different postseason. The things that helped the A's be successful in the regular season worked against them in the playoffs.
- The way the A's were constructed was missing a critical element that was exposed in the pressure of postseason play.
Beane's defenders tended toward the first explanation, his critics tended to the third. I'm probably closer to the second.
I think one thing working against the A's is that the data Beane was working from was not fully mature, and only captured things that were easy to measure, such as the three true outcomes. There weren't good statistics for baserunning and defense, so they were put into the pile of things that don't matter as much as traditionalists think, like clutch hitting and clubhouse chemistry.
Now, of course, we have better data about these things, and analytics-strong teams include great defensive players and aggressively shift to leverage them. But Beane was still working with the 1.0 version of analytics, which was missing this, resulting in distortions. The Moneyball A's were built around hulking sluggers who walked a lot but weren't so great at running of fielding. This left some gaps which their postseason opponents took advantage of.
This popped (back) into my mind reading Sonny Bunch's comparison of the Trump campaign to the Moneyball A's. Bunch thinks that Trump was the "Moneyball" candidate, embracing unorthodox strategies that gave him an advantage over the more conventional Clinton campaign.
I think the opposite -- I think the Clinton campaign was more of "Moneyball" team, and fell short for similar reasons to why the Moneyball A's never won a World Series.
The Clinton campaign's decisions were very driven by the polls, and by the analysis thereof. Decisions like not campaigning in states like Wisconsin and Michigan which unexpectedly turned the election to Clinton were driven by data-savvy people in Brooklyn over the objections of the people on the ground.
This resulted in a distorted campaign, and an anomalous result.
And I think the lesson in not necessarily to abandon the analytic approach, but to get better at it, and until you are better, complement it with some old-fashioned ground work.