The Most Convoluted Statistic: ERA

In many cases, the battle to convince those that rely on traditional metrics for player evaluation is a losing one. Consider, for example, this tweet from Mike Silva.

Trying to figure out why I should take stats with a lower case letter in front seriously

Of course, fans, writers, players, GMs, and all others are certainly welcome to watch, enjoy, and interpret the game in their own way. However, the argument that certain advanced statistics are too convoluted or complicated, whereas traditional stats are best for their simplicity, is simply not an argument grounded in logic.

Consider ERA. The formula behind earned run average is, indeed, quite simple – ERA = 9*ER/IP. Perhaps most appealing, the answer to the question “What is ERA?” has a one sentence answer: the amount of earned runs a pitcher gives up in nine innings.

Still, we are introducing what is a slightly abstract concept here. An earned run is any run for which the pitcher is held accountable. Anybody who has scored a baseball game using a pencil and paper system has likely grappled with the question of whether or not a run is earned. Courtesy of Wikipedia, here’s a handy list of which runs are unearned:

-A batter reaches base on an error (including catcher’s interference), and later scores a run in that inning by any means.
-A baserunner remains on base as the result of an error on a fielder’s choice play that would put the baserunner out except for an error, and subsequently scores.
-A batter reaches first base on a passed ball (but not a wild pitch) and subsequently scores by any means.
-A baserunner scores after the third out would have been made except for an error other than catcher’s interference.
-A batter reaches base on a fielder’s choice which removes a baserunner who has reached base safely on an error or has remained on base as the result of an error, reaching first base on a passed ball on a called or swinging third strike, or remained on base on an error on a fielders’ choice play that should have retired him, and subsequently scores.
-A batter or runner advances one or more bases on an error or passed ball and scores on a play that would otherwise not have provided the opportunity to score.

OK. Simple enough, right? Once we’ve accounted for these largely subjective criteria, including the error, another abstract concept, supposedly we arrive at a number telling us how many runs the pitcher was responsible for. If a ball goes off Franklin Gutierrez‘s glove as he shows great range but just misses, it is not the pitcher’s fault, but if a ball gets under Brad Hawpe‘s glove within any other RF’s range, it’s Ubaldo Jimenez’s responsibility.

Consider me unconvinced. Pure runs allowed does a better job of conveying a pitcher’s true talent, as it avoids this rather arbitrary system of responsibility which leads to unwarranted penalization or not enough penalization of the pitcher. However, for generations we have thought of ERA as the best measurement of a pitcher’s talent.

The real issue is that ERA does not answer the relevant questions about pitching performances. It doesn’t answer either “Who pitched the best?” or “Who will pitch the best?” That is, whose pitching performance was most valuable to his team? Whose performance, based on available data, is likely to be the best going forward? ERA’s weaknesses when it comes to providing these answers aren’t difficult to find, and it’s for these reasons that we use superior statistics such as FIP and tRA, both of which come much closer to achieving this goal.