The Brewers’ last 4 games have all been decided by one run, and they’ve won exactly half of them, which is exactly what you would expect. Most research that has been done on the subject has suggested that baseball teams generally have no particular skill at winning one run games. This seems logical, but it does NOT reflect how we experience sports. Had Jerry Hairston, Jr. not made a spectular diving play in the bottom of the 8th on Thursday, had Alex Gonzalez been able to bloop one over the infield in the bottom of the 9th last night, we would be looking at a “red hot” Brewers team, ready to take over the NL Central by storm. On the other hand, with some different bounces on Tuesday and Wednesday, we would be looking at a “struggling” Brewers team, that obviously needs to overcome some issues. The standard sabermetric article would at this point acknowledge that it’s the same team, no matter where the bounces go. I think it’s worth discussing, however, what we mean when we say it’s the same team.
Speaking very broadly, I tend to think of a team’s performance, both over a single game and over an entire season, as consisting of two components: the “true talent” component and the “random” component. In theory, these components are pretty straightforward to separate: if the team’s lineup could take an infinite number of at-bats, and the team’s pitchers could face an infinite number of hitters, what would their final stats be? Taking things out to infinity should make all randomness even out, and just leave us with true talent. The random component is then any difference between the true talent we expect and the actual results we see. In practice, of course, we never get to see the results of this experiment. Instead, we try to find ways to identify it in a smaller sample, looking at things like BABIP and FIP and Pythagorean records and so forth. Considering how long the game has been around, these things have only been figured out recently, and they’re only still gaining acceptance. That’s because trying to separate true talent from randomness is not trivial. One of the most important features of how our brains process this game is that true talent and randomness don’t really look all that different.
Suppose we have two 90-win playoff teams. If one of these teams has excellent true talent, and average randomness, and the other team has average true talent, but excellent randomness, that doesn’t change the fact that they’re both 90-win teams. We hypothesize that if we were to replay the season, the team with the better true talent will win more games, but we aren’t replaying the season. Each of these teams will play, at most, another 20 games, and randomness doesn’t shy away from 20 games. In fact, I would argue that the past two World Series winners, the Cardinals and the Giants, are teams that were stronger in the “random” department than in the “true talent” department. I’m claiming that if we were to replay the 2010 and 2011 baseball seasons, we’d see different champions. However, like I said before, those seasons aren’t being replayed, and the good draws of randomness are just as much of a part of the 2011 Cardinals and 2010 Giants as anything else!
This is where we come to the key question: when a team draws randomness well, is that team lucky or hot? Is that team just having things go their way, or is the team making things go their way? The disappointing answer is that we don’t know, because these are basically indistinguishable to us. The key word is “sustainability”. We know, from historical data, that being “hot” is not something a team can sustain. Everyone tends to return to their true talent; everyone tends to win half of their one-run games. However, we can posit that a team was hot at a given point in time, and would remain so if those games were replayed. If the 2010 Giants got “lucky” in the postseason, we would not expect them to win very often if we replayed that postseason repeatedly. But if they got “hot”, we would expect to seem them win frequently if we could engage in this wonderful thought experiment. Unfortunately, we only get to see the one realization, the actual realization of this postseason, and so being hot and being lucky look the same to us.
On Tuesday night, the Brewers were down 3-4 and sent their #5 hitter up to the plate to start the bottom of the 9th, and ended up winning the game 5-4. On Friday night, they found themselves in the exact same position, and ended up losing the game 3-4. Maybe they were hot on Tuesday night and cold on Friday night. Maybe they were lucky on Tuesday night and unlucky on Friday night. From our perspective, that’s basically just two different ways of saying the same thing.