Increasingly, I see ERA as a statistic that is more misleading than enlightening when it comes to evaluating or comparing pitchers.
The point of ERA is to try to adjust for the pitcher who suffers from a poor defense. But it doesn't really do a good job of that. First of all, a good defense does not merely produce fewer errors; it turns possible hits into outs, turns double plays out of grounders, gets runners stealing, etc. Secondly, while ERA accounts for unearned runs, it doesn't account for the toll a long inning takes on a pitcher in that game, or the fact that his total innings pitched (the denominator in the fraction of the formula) will go down if his defense is bad.
Worse, ERA overlooks one of the best characteristics of a top pitcher: The ability to hunker down and pitch around errors. Just because your shortstop boots one ball doesn't mean you shouldn't be evaluated negatively if you then give up three singles and a homer.
Also, it's a fact that pitchers who walk lots of guys suffer more unearned runs. That's because they put their defense back on their heels, then put them in pressure cooker situations whenever balls go into play. In other words, sometimes a pitcher with poor control CREATES fielding errors that lead to unearned runs. So the unearned runs really are the pitcher's fault all along.
This time of year, in high school especially, we see the other failing of ERA: it can be manipulated. If a coach wants his player to get All District honors or whatever, he can cook up a low ERA simply by adjusting his scorebook. A 2-out hit gets re-scored as an E, and viola! All the runs are now unearned. I'm amazed at how you can find all these pitchers with multiple losses in their record, but my, their ERA's sure do sparkle.
A modest proposal: Let's evaluate pitchers on RUNS PER GAME, not "earned" runs per game. True, those whose defenses are genuinely substandard will suffer some, but I submit that RPG would be a more reliable measure of performance than ERA.
Original Post