That’s right, the “+” statistics exist in the realm of pitching. Park adjustments and a percentage normalization is a natural extension of ERA, and makes it more meaningful for directly comparing pitchers, even across generations.
We know generally how this will work. Park and league adjustment factors are computed, and we use those to slightly augment a pitcher’s ERA based on the environment in which each run was scored. Then, we scale everything so that league-average is 100. However, there is a trick: ERA is a statistic where lower is better, so our fractions are getting a little inverted. An ERA+ of 150 would mean that the league’s average ERA is 50% higher than the pitcher’s. It’s the difference between saying that 5 is 25% greater than 4, while 4 is 20% less than 5.1Later on, we’ll look at the more “proper” way to do it, where percentage points mean something for the pitcher, not the league.
Instead of diving into further mathematical details, let’s use ERA+ to explore a comparison I made in the ERA post. There, I pointed out that Bob Gibson’s 1.12 ERA in 1968 is the best by a qualified starter in MLB history, followed by a 37% increase in Dwight Gooden’s 1.53 ERA in 1985. But that’s nearly a twenty year difference, and I explicitly called out that pitchers had an abnormally good season at-large in 1968.
We can use ERA+ to see how much better these two seasons were compared to the average of their own seasons, and thus get a good indication of which season was actually better overall.
Do you have any guesses based on the hand-wavy nature of how I’ve discussed this?
In 1968, Bob Gibson’s 1.12 ERA was quite exceptional compared to even the elevated pitching of the league. His ERA+ was 258. That means the league was 158% worse than Bob Gibson on average.2Throwing back to yesterday: his WHIP was an elite 0.853.
Dwight Gooden’s 1985 was outstanding if only because he was just 20 years old when it began. It was his second season in the league, and he posted a 1.53 ERA (along with a 0.965 WHIP). His ERA+ was 229, meaning the league was 129% worse than him in terms of ERA.
Since the league’s ability compared to Bob Gibson was lower than compared to Dwight Gooden, we can still confidently say that Bob Gibson’s season was overall more impressive, given the context of the time in which he played.3Notice how clunky it is to say “the league is x% worse than this.” We’ll fix that in a couple of days.
Let’s take a look at some other interesting ERA+ numbers.
- Since 1920, Bob Gibson’s 258+ is actually only the the sixth best value among starters. The top mark is Pedro Martînez’s 291 ERA+ in 2000. In a league with steroid-ridden sluggers, his 1.74 ERA was magnificent.
- Greg Maddux has two seasons higher than Gibson’s: a 260 and 271 ERA+ in 1995 and 1994 respectively.
- Relievers live much more interesting lives, because a few bad outings can have an outsized impact on ERA due to fewer innings pitched. As such, Zack Britton’s 2016 campaign of 69 games pitched for the Orioles led him to a 0.54 ERA, good for an 803 ERA+. Amazing.4We’ll explore this season tomorrow, because some of his advanced statistics suggest this was a lucky, or at least incredibly unlikely, season for him to have. This absurd number also suggests that ERA+ is not nearly as useful as something like OPS+. The converse nature of it is confusing.
Continue to Day 15 – Fielding Independent Pitching
- 1Later on, we’ll look at the more “proper” way to do it, where percentage points mean something for the pitcher, not the league.
- 2Throwing back to yesterday: his WHIP was an elite 0.853.
- 3Notice how clunky it is to say “the league is x% worse than this.” We’ll fix that in a couple of days.
- 4We’ll explore this season tomorrow, because some of his advanced statistics suggest this was a lucky, or at least incredibly unlikely, season for him to have. This absurd number also suggests that ERA+ is not nearly as useful as something like OPS+. The converse nature of it is confusing.