Why Earned Runs Are Silly

ERA - pretty much the most popular stat to measure the effectiveness of a pitcher. It keeps track of the amount of earned runs given up over nine innings. As you can (hopefully) tell by the name, this stat does not take into account unearned runs, which is either a) when a run scores directly as a result of an error, or b) when a runner who reached base as a result of an error comes around to score. The rulebook definition is a little more complicated, but that's the general idea.

One such run was scored on David Price in the fourth inning of tonight's Game 5 of the ALDS. Nelson Cruz, on second as a result of a double, attempts to steal third base. Rays catcher Kelly Shoppach's throw is an erroneous one (I don't know exactly what happened, cuz I'm stuck on my computer with just mlb.com gamecast dots), and Cruz ends up scoring. As a result of an error, so that's an unearned run, right?

It was, but only until the end of the current at-bat, when Ian Kinsler knocks a single. As soon as that happened, the run in the box score changed from unearned to earned. Why is that, you ask? It's because of some screwy rule and a judgment call that allows the official scorer to change the status of a run in retrospect.

If, in the official scorer's estimation, a play following an unearned run would have resulted in that run scoring, then that unearned run becomes an earned run. So in this situation, we basically pretend that Cruz scored on Kinsler's clean hit rather than Shoppach's messy error. Funny thing is that Shoppach still gets charged with that error, in addition to Price getting charged with an earned run.

But here's the thing: Kinsler singled on a hanging curveball on the outside part of the plate, a pitch that was selected and thrown after Cruz had scored on that error. Who's to say that Price would have thrown that pitch if he still had a runner sitting on second (or on third, if Cruz had successfully stolen the base)? Any decision in baseball (as in life) is affected both by the circumstances in which that decision is made and by what events immediately precede that decision. When we get into pretending that things had turned out differently and basing our statistics on retroactive what-ifs, you confuse player evaluation, even if it's just to a trivial extent.