What is ERA in Baseball and Pitching Stats Explained
One of the most used statistics in baseball is ERA or earned run average. It is often used (along with WHIP) as a quick look to evaluate how good a pitcher is. But does it tell the whole picture?
In this article we look at ERA in baseball and help to understand how it fits into the complex world of baseball stats. What is ERA in baseball? Then, we’ll take at other pitching stats and ERA variants to describe the pros and cons of each. Finally, we’ll comment on why it is always the simplest stats that stick with us the most.
What is ERA in Baseball
ERA or Earned Run Average is one way to measure how good a pitcher is at his job. In baseball, the name of the game is scoring runs. For pitchers, though, your goal is to prevent the other team from scoring runs. The less your opponents score, the better you are at your job!
ERA measures a pitcher’s skills by measuring how many runs he lets the other team score. The earned run average stat doesn’t just count the number of runs given up, though. It does two things to make the stat better and more representative of how good a pitcher actually is.
- Normalizes the runs by the number of innings pitched
- Uses the concept of an earned run to partially remove the impact of the defenses behind the pitcher
Normalizing by Innings Pitched
Not all pitchers pitch the same number of innings. Giving up 2 runs in 1 inning is way worse than 3 runs in 9 innings! If we take the number of runs and divide by the number of innings, we get a better measure of pitcher quality.
Instead of presenting runs per inning, earned run average presents “runs per 9 innings” which is easily interpretable as “runs per game”. This can be done by first figuring out the number of games a pitcher has pitched for.
If they have thrown for 45 innings, they’ve pitched 5 games (equivalently). If they have thrown for 3 innings, they have thrown for 1/3 of a game. This number of games is what we divide the runs by to normalize it.
What is better: 6 runs in 27 innings or 1 run in 3?
For example, 6 runs in 27 innings would be an ERA of 2 (very good!). This comes from 27 innings converting to 3 full games. Then, dividing the 6 runs by 3 games is an ERA of 2.
Another example, 1 run in 3 innings is an ERA of 3! The 3 innings is 1/3 of a game. Allowing 1 run in 1/3 of a game is an earned run average of 3 which is computed by dividing 1 by 1/3.
This normalization helps us compare pitchers who have pitched for differing numbers of innings.
What is an Earned Run
The other adjustment made by the ERA stat is using the concept of “earned” runs instead of just plain old runs. The idea is that some runs are a pitcher’s fault, but some aren’t. If the defense made a mistake, that isn’t the pitcher’s fault. The pitcher shouldn’t be penalized because of it.
The main way this happens is errors. Sometimes a defensive player makes an error and doesn’t get a player out he should have. Even worse, sometimes the batter who should have been out ends up scoring a run. That never should have happened in the first place if the defensive player did his job.
Earned runs try to distinguish between runs which were and weren’t the pitcher’s fault. It only counts the runs which were the pitcher’s fault when evaluating the pitcher’s quality. Earned run average only counts earned runs, not all runs.
Earned Run Average Formula
In the last few sections, we described how ERA was defined and a little bit about why it was defined that way. It can also be helpful to see the ERA formulas written out. If the previous section was “what is ERA, informally” this is “what is ERA in baseball, explicitly”.
Earned run average combines two numbers. The first is the number of innings pitched. We’ll write this as IP . The second is the number of earned runs which we’ll denote ER .
Then, earned run average is given by the formula ERA = 9\times \frac{ER}{IP} .
It is clear by looking at this formula that ERA is simply 9 times the per-inning average number of earned runs. This is the expected number of runs a pitcher would allow if they pitched a full game with great defense.
The History of Earned Run Average
ERA has been around for forever. Early in the days of baseball, pitchers usually threw for the whole game. This is before modern medicine and, distinctly more importantly, before guys were routinely cracking 100 MPH with insane spin rates.
When only one guy pitched for 9 innings, it was easy to see how well a pitcher did. To do this, you would just look at his runs or his runs per game. But when relief pitching became popular, it became more difficult to rank the quality of various pitchers.
The stat was invented by Henry Chadwick, a man known lovingly as “the father of baseball”. Chadwick was a reporter, Both a journalist and a statistician (and, as you know, we love statisticians here), he contributed monumentally to Baseball’s popularity in a time when information was harder to come by.
One of the ways Chadwick did this was by inventing the Earned Run Average. The National League formally adopted the stat in 1912 and hasn’t looked back since.
Where ERA Falls Short
Though ERA takes a few mitigating factors into account, it is still not a perfect measure of how good a pitcher is. Normalizing the raw data by innings pitched. Normalizing for the defense is even better! But it can still do better.
The problem with ERA is variance. Sometimes the difference between a lot of runs and no runs is just 1 pitch. And even worse, sometimes it is that the batter made a great play and not the fault of the pitcher.
Consider the case where the bases are loaded with 2 outs. If the batter strikes out, the pitchers gets credit for a GOOD inning. No runs allowed! This doesn’t tell the whole story, though. The pitcher was on the precipice of disaster.
On the other hand, if the batter hits an unbelievable pitch for a home run, the pitcher gets the credit for all 4 runs. They had a bad inning, but maybe the four runs is too punishing.
The amount of runs scored depends on a ton of external factors. A lot of luck is involved, too. Better stats would depend less on a high variance stat and capture some better measure of pitcher quality.
Different Pitching Stats
What is ERA if not the simplest pitching stat. There are many more complicated stats that try to do the same thing. One such stat is WHIP, which we’ve discussed at length before. Whip is probably better at capturing overall pitcher quality because counting walks and hits has lower variance than counting earned runs. But it still doesn’t work perfectly because it doesn’t perfectly capture the difference between types of hits.
Let’s talk about two more pitching stats which are alternatives to ERA. The first is DICE and the second is RE24.
DICE is an interesting stat which stands for “defense independent component ERA”. The idea is to completely take the defense out of the equation by only including home runs, walks, and strikeouts in evaluating a pitcher. The problem is that HRs, BBs, and Ks only tell a very small part of a pitcher’s quality. Some guys just don’t give up HRs or don’t get strikeouts. It doesn’t mean they are necessarily good or bad.
A second stat which I’ve written about before is RE24. RE24 takes into account everything that happens to a pitcher and puts the perfect weight on it. I think RE24 is better for batters than it is for pitcher’s, though. The problem with RE24 for pitchers is that (a) it doesn’t take into account the defense at all and (b) you can still escape from a bad situation with a single strikeout without being penalized.
A perfect stat would be one that penalizes a pitcher that allows the bases loaded but then gets a lucky strikeout. Figuring out precisely how to do this is difficult, though.
Where ERA Succeeds
While I’ve discussed a bit about what I don’t like about ERA as a stat, it does have one really good attribute. The reason ERA is so good and is used so often is that it is extremely simple and easy to interpret.
All stats exist on a spectrum from simple/easy to interpret and complicated/hard to interpret. Each point on the spectrum has its own utility. The really simple and easy to understand stats are good for communicating information quickly or communicating to inexperienced users.
On the other hand, complex stats are good for extracting maximal information but sacrifice the fact that they can be interpreted easily.
ERA succeeds and is so often used because it is an extraordinarily easy stat to use. Looking at ERA tells us quickly roughly how good someone is. And that is a valuable thing to have.
