Two of the most commonly used stats for pitchers are ERA and win/loss record.
Both of them seem, not to put too fine a point on it, to be stupid.
ERA - does not include (unless I’m very confused) walks. Why would the prime stat for measuring a pitcher’s performance not include something that is absolutely directly under a pitcher’s control?
Win/Loss - this has been discussed a lot, but not only does it depend on the team’s offense, it also has all these byzantine and pointless rules about how runners left on base on are still the responsibility of the previous pitcher, yada yada yada.
Seems to me that there’s a really simple thing to do, which is occasionally done but never seems to have caught on as a “primary” stat. We have various numbers that indicate how good a batter is, with batting average being the best known one, and OPS being a more modern one. Why not just use the that (well, the inverse of that) as the pitcher’s stat as well? If .150 is a miserable batting average, then a pitcher who holds opposing batters to an average of .150 is a good pitcher. If .600 is a good OPS, then a pitcher whose batters have an OPS of .600 is a bad pitcher. Simple and easy, right?
Is there some subtletly I’m missing which would make just a single stat like that not work as well as I think it would?
ERA is runs allowed per nine innings pitched, unless the run is unearned. An unearned one is one that would not have scored without the benefit of an error. Walks are not errors. If a pitcher up and walks the first four guys in an inning, the run that scores is earned, and counts to his ERA.
Your confusion here might have to do with the fact that walk aren’t counted when computing a batter’s batting average. It was apparently once felt that this would represent a sort of unearned bonus to a batter’s stats.
It’s now widely believed that the ability of a batter to draw walks is a conspicuous virtue. As these are figured into OPS, that stat is now often considered a better measure than batting average.
Because that’s not what it measures - ERA measures Earned Runs Allowed. If you want a measure of base runners allowed, WHIP (Walks+Hits/IP) is a better tool.
You got a definition, so some context. I’m not really a sabermatrician, but my understanding is that the consensus is, as far as simple directly-calculated stats go, OPS is actually pretty good at measuring hitting ability. Obviously, if you want to add in effects for the different fields (easier to hit home runs in the air of Denver, for instance), and so forth you can get a better measure (and many people who calculate complex stats then normalize them to the league average, to account for changing strike zones, etc.) But for a simple stat, OPS is pretty good.
And, to the OP, ERA really isn’t that bad a simple stat either – as has been mentioned, it does account for walks; it only excludes runs that wouldn’t have scored if there were no errors. ERA is way way way better than Wins, which is probably exceeded in uselessness only by Saves.
But – and I can’t remember whether it was during a broadcast or on ESPN or something-- I have seen discussions or pitchers in the World Series where their opposing OPS was mentioned.
That’s true. It’s not a very precise measure. But it does serve well for a commentator to start with. We can expect that a 17-3 pitcher is better than a 12-8 pitcher. From that point better stats are needed to find out why, the 12-8 pitcher actually could be unlucky due to circumstances but the better stats more likely reveal something that is a real weakness. And even then OPS against pitcher would need more depth to reveal something like poor performance against left handed batters, or poor judgement from a manager who didn’t pick up something like that.
I must nitpick; ERA stands for Earned Run Average.
Earned Runs allowed is a simple counting statistic. Corey Kluber allowed 75 earned runs this year. ERA is the average number of runs he allowed for every 9 innings he pitched, where an “inning pitched” is defined as getting three outs while the pitcher is on the mound (by any means; if a guy is caught stealing that is counted as third of an inning.) Kluber pitched 215 innings, so his ERA is 3.14 (75*9/215).
As Quercus points out, ERA’s a good stat. Actually, it’s a rather excellent statistic for telling you how well a pitcher has pitched. It is not necessarily the case that a guy with a 3.57 ERA actually pitched better than a guy with a 3.72 ERA because some things might not be accounted for in such a small difference, but if a guy has a 2.50 ERA he pitched well, and if a guy has a 5.50 ERA, he didn’t.
I’m not so sure. I’d agree a 3-17 pitcher is probably worse than a 17-3 one, but the difference between 17-3 and 12-8 could easily be that one is pitching for the 2016 Red Sox (878 runs over the regular season) and the 2016 Phillies (610 runs). In fact, eyeballing it, that looks like roughly the difference in wins we’d expect from identical pitchers with the two different levels of run support.
One way to establish how good/bad the won-loss record is is to correlate it with ERA. Since ERA is accepted as having some reasonable value for determining success as a pitcher in a given season, the extent of the correlation between the two would establish the use (or lack thereof) of the W/L record.
I stopped paying attention to W/L as a metric back in the 70s when I would see various pitchers have decent ERAs but lousy won/loss stats, usually because they pitched for the Cubs (or the Angels!). :smack:
I was trying to put that consideration into my post and I guess it got murky. My initial point was that commentators would start with very general and easy to understand stats. Then you need to more details to really make comparisons between pitchers. And you mentioned run support which is the kind of thing to be considered in all stats, it’s really hard to attribute any stat solely to the player unless it’s remarkably good. In these days a batting average over .400 would be extraordinary and tell you a lot about a hitter, otherwise BA is very rough and sometimes misleading indicator. I’m a [del]masochist[/del] Phillies fan and I remember poor Steve Carlton’s 1972 27-10 record, on a decent team he could have been the last 30 game winner. So yeah, my choice of two 20 game pitching W/L losses records wasn’t a great example because of how complex it is to measure performance.
