Early in his career (as a Pirate), current ace Charlie Morton once gave up ten runs in a single inning. That's a bad, 90.00 ERA for that one "frame."

But you can only lose one game at a time. Suppose a reliever gave up those ten runs in one inning, and then pitched eight one-inning stints with 0 runs. Over those nine innings, his ERA would still be a lousy 10.00, but he would have been responsible for losing at most the one game.

So are there ways that managers adjust for "kitchen sink" pitchers, that "bunch" a disproportionate number of their lost runs into a handful of outings?*

*In the case of Morton, I did my own rough calculation. I calculated that his ERA would be about 3.00, after taking out his worst three outings (which ballooned his ERA to 5.00). So I started him at 0-3 for the three bad (and presumably lost) games. Another pitcher with a 3.00 ERA had a W-L record of 9-4. I "estimated" Morton's record at 9-7 (after adding the three "kitchen-sink" losses), which is a decidedly better record than that of a typical 5.00 ERA pitcher.

1 Answer 1


No, not really. What you're describing would only be useful if it could be shown that this ability to have mostly good performance with little blips of horrible performance was something that was innate, rather than just a normal observation of random chance. It's not, most likely. You also nearly never have pitchers who have quite this extreme of results - starters might sometimes give up this many runs (as Charlie Morton was), but they have huge numbers of innings - 150-200 for a full time starter - and a few "blips" won't matter very much. Relivers, who pitch much fewer innings, would be taken out before they could give up ten runs nearly always.

But either way, you would need to show this was an innate trait of the pitcher, and not just artifacts of chance, for this to be of value. Otherwise you're just describing something random - which everyone is as likely to have as any others. You also have the issue that, perhaps in this 10 run game, the pitcher was injured, or was tipping his pitches, or had some other issue that caused it - stats don't catch that, because they can't. That's why you have to use statistics in combination with other analytical approaches, and not just blindly rely on a few numbers to paint a picture of a player's abilities and likely future performance.

You also use statistics like FIP, instead of ERA, because they take more of the pitcher's performance into account - even with things like streakiness. They don't directly factor it in, mind you, but they more accurately reflect how the pitcher performed relative to possible future performance. Removing the results of random things like high BABIP and mediocre (but not error-prone) fielding, they focus on the things the pitcher can control - and that control is always a factor, streakiness or not.

Finally, you talk about "wins"; that's one statistic that is entirely worthless, and no statistical pitching model worth its salt would even factor at all. You found a 3.0 ERA pitcher with a 9-4 record? Great! May I introduce to you Felix Hernandez, who won a Cy Young with a 2.25 ERA... and a 13-12 record. Pitcher Wins are too much a factor of other things - run support, what team they're playing on, etc. - that they're not predictive at all of future individual performance.

For the most part, pitchers are evaluated on things - like K/BB, K/9, etc. - that wouldn't even really be impacted by these bad games meaningfully; unless he gave up 10 runs all on walks, his rates would likely be mostly be minimally affected by the couple of bad games. There are definitely relievers who do have "high" ERAs from a few bad games - and they're valued because of these other statistics, rather than some sort of made up statistic to evaluate ERA (which, itself, isn't very useful).

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.