Tuesday, September 4, 2012


Just a quick shot for now, as we are on the road for the next few days...but we'll get this idea started and supplement as possible during our period of scrambling around.

All the discussion of "playing over one's head" engendered by the Baltimore Orioles got us to thinking about what is the best way to represent team performance in a chart.

It's particularly important when doing so to capture some kind of running totals for this, because the use of "season to date" information is sorely lacking in context. (Many folks have pointed out that the O's roster is much different in September than it was in May, which is true...we need a way to capture some kind of "in medias res" for teams as their performance fluctuates during a season.)

So we'll start with this one. We'll take a succession of 30-game snapshots, and they'll be organized around the concept of run differential (a favorite of those folk who are in a bend-or-be-bent relationship with the mainstream media). It's a good place to start, though like anything else, it's not the final word.

We'll just do the AL East, given that it continues to be the lightning rod for all things that matter in the little world of neo-sabermetrics. Warning: the chart is a little busy, but we've made it as big as possible in order to make it possible to follow the lines:

As we said: a little busy, but you can see that the Orioles have been running on baseball's version of deficit spending for a good portion of the season. Following the Red Sox line, we can see how their run differential has gone down the drain in the last forty games or so. The Yankees have been pretty steady all year, though they are giving ground of late. Right now the Rays are playing better than anyone else in the AL East for the entirety of the 2012 season. And the less said about the injury-riddled Blue Jays, the better.

But this doesn't encompass the actual won-loss records of the teams, so it's one of those gauzy "should have been" characterizations. There must be another way to synthesize this with the actual results, and that's what we'll start to tackle next time.