2023 SABR Analytics Conference Research Awards: Voting Open Now!

While looking toward the future with our comprehensive slate of current content, we'd also like to recognize our rich past by drawing upon our extensive online archive of work dating back to 1997. In an effort to highlight the best of what's gone before, we'll be bringing you a weekly blast from BP's past, introducing or re-introducing you to some of the most informative and entertaining authors who have passed through our virtual halls. If you have fond recollections of a BP piece that you'd like to nominate for re-exposure to a wider audience, send us your suggestion.

Nearly seven years after it originally ran on April 8, 2004, Michael Wolverton's well-reasoned anti-earned-run diatribe remains persuasive. Stay tuned for our unveiling of a reworked pitching stat that should make Michael happy…

We at Baseball Prospectus occasionally hear the complaint that we make the game too complicated, with all the numbers and bizarre acronyms we throw around. So today I'm going to do my part to simplify the game. I'm here to suggest that baseball and its fans would be better off without one of its most fundamental, and most complicated, scoring rules. It's time to ditch the "earned" run.

The earned-run rule is widely accepted, or at least tolerated, throughout the baseball world, even in sabermetric circles. There are several reasons for that. For one thing, there's 116 years' worth of tradition behind the rule. I learned the rule because my dad learned the rule because his dad learned the rule, etc. ERA is on the back of every pitcher's baseball card, and it pops up in nearly every baseball-related article or news report you'll see. For another thing, believing in "earned" and "unearned" runs isn't nearly as harmful as, say, believing that RBI are meaningful for evaluating hitting. You have to pick your battles, and in the big scheme of things, this one may not be a battle worth fighting.

Perhaps most importantly, the earned-run rule might have gotten a pass because it's designed to achieve what everyone agrees is a noble goal: separating pitching from fielding. But good intentions aren't enough. The earned-run rule is a lame and counterproductive attempt at solving the pitching/fielding conundrum, one that deserves to be put out of its (and our) misery.

There are many angles from which to attack the idea of "unearned" runs. Here's the one that works best for me:

Pretend for a second that the earned-run rule hadn't been invented over a century ago. In this alternate reality, baseball is played the same as it is now, except there's one fewer column in the box score for pitchers. Pitchers are charged with all the runs they allow, and RA rather than ERA is the de facto standard for evaluating them. Everyone recognizes that good or bad fielding can affect a pitcher's RA, but no one has come up with a good solution to the problem.

Then one day, an enthusiastic columnist writes:

I've done it! I've solved the problem of removing the corrupting influence of fielding on pitchers' RA. We simply pay a sportswriter to sit in the press box, munch Cheetos, and decide which safeties came on plays that should have been made with normal fielding effort. Whenever one of these 'errors' occurs, we reconstruct the inning–not the game, mind you, just the inning–pretending as if the error never happened. Count up the runs that would have scored in this hypothetical reconstructed inning, and you have a revised run total for the pitcher. Things get a lot more complicated for relievers and team totals, and we'll broaden the 'plays that should have been made' definition a little bit, but you get the idea."

For me, shifting the rule's invention into the present day helps illustrate just how absurd it is. If it were proposed today, this idea would be dismissed out of hand, no matter who advanced it. If it had come from the blogosphere, people would use it as an example of the inferiority of the Internet. If it had come from a respected columnist like Peter Gammons or Thomas Boswell, people would wonder if senility was setting in. Absolutely no one would take this idea seriously. So why should it be any different just because the idea originated in 1888 instead of 2004?

It could be construed from our fictitious columnist's quote above that the main problem with the earned-run rule is its reliance on errors. And it's true that errors are one of the problems with the rule. The well-discussed shortcomings of errors as a defensive measurement–their miniscule coverage of defensive plays, their subjective nature, etc.–suggest that ERA isn't doing much to eliminate fielding from pitching numbers.

But focusing on errors is really missing a bigger problem. Even if errors were somehow the perfect measure of fielding performance, the earned-run rule would still be wrong-headed. The main problem with unearned runs isn't errors, it's the notion that the pitcher's job ends whenever an error is made.

The most glaring example of this problem is that the earned-run rule allows the pitcher to give up runs with impunity after the third out "should have" been recorded. We've all seen these kinds of situations. An error allows Michael Tucker to reach with two outs, and Barry Bonds comes up next to deposit one in McCovey Cove. The result: two runs, none earned. That's nuts. Never mind Tucker's run; you're telling me that Bonds wouldn't have hit his homer leading off the next inning if Tucker had been put out? In this situation, the earned-run rule essentially allows the pitcher to skip Bonds in the lineup, as far as his ERA is concerned.

OK, you might say, why not hypothetically reconstruct the entire game, rather than just the inning? That way, Bonds' homer would count as earned. But of course, that would be even more ridiculous than the current rule. Besides the fact that it's way too complicated to reconstruct an entire game this way, you'd frequently end up with absurd situations where the pitcher is charged with more earned runs than actual runs. It's possible, in hindsight, for errors to save you runs, if you buy into this hypothetical reconstruction philosophy.

No, the right solution to this problem is to recognize this simple baseball truth:

Errors will happen. Good pitchers will minimize the damage caused by them. That is, a good pitcher will allow fewer runners on base before the errors happen (so there aren't runners to score on the errors), and will allow fewer hits and walks after errors happen (so the runners who reached on errors won't score).

This isn't a hypothesis, it's a fact. Preventing unearned runs is a skill that pitchers have, and it usually comes hand-in-hand with the ability to prevent earned runs. I took all the pitchers since 1900 who pitched more than 2,000 innings, and compared their earned run averages (ERA) to their unearned run averages (UERA). (In both cases the values were normalized to the league averages for the years in which they played). A couple of results:

  • The correlation between ERA and UERA was 0.36 — not overwhelming, but pretty strong. That means that pitchers who are good at preventing earned runs are also generally good at preventing unearned runs.
  • Of the top 50 pitchers in career (normalized) ERA, 46 of them were better than average at UERA.

What this means is that when you throw out unearned runs, you're throwing out part of the pitcher's performance. In other words, ERA is understating the run prevention abilities of the best pitchers in the league, and overstating it for the worst. Measures based on ERA, such as Pete Palmer's Pitching Runs from the Total Baseball books, are going to underestimate the seasonal and career values of baseball's all-time greats.

We noted above that the correlation between ERA and UERA wasn't perfect, so let's look at some of the exceptions. Here are a few of the pitchers who had average-to-excellent ERAs combined with terrible UERAs, along with their rankings (out of 317 total pitchers) in both categories. You may notice from the list of names that they also have something else in common.

Rank out of 317
Pitcher                ERA    UERA
Hoyt Wilhelm             4     295
Dutch Leonard           36     307
Tom Candiotti          115     310
Phil Niekro            125     297
Wilbur Wood            126     302
Charlie Hough          174     301

Coincidence? Not bloody likely. Knuckleballers are a huge exception to the "good pitchers prevent unearned runs" rule. Despite being good pitchers overall, these knuckleballers were all among the worst all-time at surrendering unearned runs. Joe Niekro was the only long-time knuckleballer I could think of who didn't fit the pattern, finishing about average in both categories. (And Joe threw the knuckleball far less often than any of these other guys.) The reason is obvious: Knuckleballers produce a lot of passed balls, passed balls count the same as errors in the unearned-run rule, hence a bunch of unearned runs.

So now the situation is even worse than what we saw above. Not only does the unearned-run rule generally undervalue good pitchers, it's also systematically biased in favor of a particular type of pitcher. Phil Niekro, for example, looks quite different when you evaluate his career based on RA rather than ERA. He's a Hall Of Famer either way, but he's not the Inner Circle guy that Total Baseball makes him out to be if you take his unearned runs into account.

The bottom line is that there's little evidence that unearned runs have much to do with fielding, and there's plenty of evidence that they have a lot to do with pitching. ERA has a long tradition as a pitching metric, but if we replaced it with simple RA, we'd be getting a more accurate measure of a pitcher's contribution. And if the average baseball fan would free up the brain cells he's using to store the complicated earned-run rule, and use those cells to learn about other elements of the game, baseball would be the better for it.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
well done.
I've always loved the fact that a pitcher's *own* errors can lead to unearned runs. Why? It's still his fault either way!
7 years later, and this article is nearly just as applicable. A lot of people have come to the realization that using errors and fielding percentage to determine a fielder's ability is very flawed, but still have no issues using ERA. Maybe 7 years from now, people will finally start using runs against.
I disagree totally. There is no way to evaluate evenly a pitcher for the Pirates or Marlins, who have very poor defenses, with a pitcher from a great defensive team. Every pitcher has enough pride to not want the unearned runs scoring knowing they affect his WHIP and possibly his W-L record. With the pitcher not worrying about a base runner and not pitching from the stretch, Bonds may not have hit a homer leading off the next inning, to comment on your example. A "good" pitcher can have two successive errors leading to an unearned run. Leave it the way it is.
You can make the same team arguments for runs and rbis. Baseball is a team game, thus the team will have an impact on the stats. I think this is a great article and makes complete sense. While there is no perfect metric, RA is a better way the evaluate a pitchers performance. I realize fantasy baseball is stupid, but I play it anyway. I am thrilled, but feel like I have cheated, when one of my pitchers gives up unearned runs. The important fact is that a better pitcher is going to minimize the impact of errors, thus in the end having a lower and more meaningful RA.
So should we grant RBI to players who hit a HR just after a teammate gets picked off? After all, the hitter didn't deserve to get deprived of the RBI. All baseball stats are flawed. Not sure I follow the level of agitation with ERA. If "stat reform" is the platform, I think wins, losses and saves deserve attention before ERA.
It seems clear that RA is a more accurate metric than ERA. As stated, the main reason ERA has been tolerated is that it is a traditional statistic. I also think there is a belief that not enough is gained by going from ERA to RA that we should upset the traditionalists on that one. Thus, we have focused our energies on eliminating pitcher w/l record and looking for fielding and sequencing independent alternatives to ERA.

I'd personally like for us to concentrate more on RA, but I'm wondering how much going from ERA to RA would change the game's history? Is it going to lead us to vastly different conclusions about pitchers or is it just another metric which will please the stat crowd but confuse the traditionalists? There may be some interesting research there.
It's not clear at all to me that RA is a more accurate metric than ERA. A pitchers true skill level is best known if you factor out bad defense. What's more obvious than that? There is absolutely nothing to gain by going from ERA to RA. The focus should be on greatly reducing the significance of W-L records, which is so dependant on run support and out of the pitchers control. Just ask King Felix.
So why not simply take RA and adjust to Park Factors and the difference in his own defense and that of the league average?
Another important factor not discussed nearly enough is the Inherited runner/Inherited runner scoring effect on the starter/reliever. A rather unlucky Dodger Chad Billingsly (19 LOB and 10 scoring) vs. a very lucky Marlins Anibal Sanchez (25 LOB and 3 scoring) has a profound effect on ERA or RA. These are not all "runner on third no outs compared to runner on first two outs" scenarios of inherited runners. A much more accurate metric than ERA to RA would be some way of mitigating the inherited run damage that a crummy reliever allows.
How about Total Bases allowed? A simple counting stat measuring how much base-runner advancement happened under a particular pitcher. For example, a walk followed by a double would be five bases if the lead runner holds up at 3B, but 6 bases allowed if he scores. This could then be divided by 9 (for IP) to be read somewhat like WHIP.
Yes, getting a lot of help (or, alternatively, a lot of kerosene) from the firemen in the bullpen can have an impact on a pitcher's stats.

BP's Support Neutral stats for starters takes that into account: the "support neutral" refers to averaging out both the offensive support and the bullpen support.

In any event, claiming that ERA is preferable to RA because it helps smooth out differences in defensive support when comparing pitchers on different teams is just nonsense: since errors is a very unreliable stat for evaluating fielding performance, it is not at all clear that tracking "unearned runs" separately from "earned runs" tells us anything useful.
A pitcher's skill level is best determined by factoring out defense, but ERA doesn't do a good job of that. I don't think ERA really provides anything that RA doesn't except make things more complicated. Until, we figure out how to measure defense better, I prefer RA.
So, basically what Wolverton was saying here is that knuckleballers are over-rated?
Francisco Liriano had an amazing zero unearned runs last year. Likely a result of statuesque outfielders and offical scorering that bought into "the twins are the best fielding team in baseball!" crap.