Debate around pitchers tends to center around ERA and FIP, and their component parts. However, there are a few other interesting tools for examining a pitcher’s overall performance in a season. One element that isn’t captured by either statistic is how often a given starting pitcher was, or wasn’t, bailed out of tight situations by their relief corps.
For this, we’ll turn to our old friend, deserved run average (DRA). DRA you ask? Another statistic? We’ve used DRA in the past in a few pieces, as it provides an interesting perspective on pitchers and pitching staff management, but let’s briefly re-introduce the concept.
Created by the most excellent folks over at Baseball Prospectus, a site that has produced quite a lot of front office talent over the years, DRA attempts to look at how much a given pitcher’s ERA was affected by the quality of relief work that followed.
It accomplishes this by assigning run expectancy values to a pitcher for the situation in which they left the game, rather than the ERA produced in those situations by the success or failure of their reliever to strand inherited runners. In the process it gives a clearer picture of what a pitcher’s ERA “should be” instead of allowing the success or failure of their relief to color their ERA.
For example, say Matthew Boyd leaves the game in the sixth inning after allowing a single and a walk, with one out recorded. In that specific situation, .91 runs can be expected to score. What actually happens is entirely dependent on the reliever that takes over, but this creates a problem. Teams with better relief corps are going to make the pitcher they’re taking over for look better than he actually is.
Let’s say Boyd leaves five different games in that specific game situation over the course of the season, and in all five cases, the reliever manages to get out of the inning without allowing a run. Does Boyd deserve credit for having better relief work behind him than another starter? Deserved run average would argue that no, he does not. Instead, Boyd would be charged .91 runs for leaving the game in that situation, adjusting his ERA upwards.
Of course, this can also work to a pitcher’s benefit in terms of adjusting his ERA this way. Let’s consider a scarier hypothetical situation, where Boyd leaves the game with runners on first and third and one out. In that scenario, the run expectancy is 1.14. At least one run is a given, on average, in that situation. But let’s say the reliever who takes over immediately surrenders a three-run home run. By ERA, Boyd would be charged two runs for the two men he left on base. However, DRA would only charge him for the state of the game as he left it, not the bad result generated by the reliever. He would still only be charged for 1.14 runs in DRA’s calculation.
So you get the general idea. DRA is used to moderate ERA and remove the work of the relievers from the equation. We’ll get to “what it means” in a moment. So without further ado, let’s compare every pitcher who started five games or more for the Detroit Tigers in 2021, and see what DRA says in relation to their actual ERA mark.
Tigers starters DRA 2021
The key thing to remember, is that DRA helps to tell the story of a pitcher’s season, but should not be thought of as a pitcher’s “real” ERA. There’s a good reason managers turn to the bullpen when their starters are tiring or getting into trouble late in an outing. Your starters’ ERA should be better than their DRA. The whole point of going to the bullpen mid-inning is to shut down a developing rally.
If your manager is pulling starters from the game at the right times, and inserting the right reliever, in the best matchups possible, that fresh reliever should be better equipped to handle the situation than a starter the opposition has likely seen twice already, and who may have already thrown 90 or more pitches. So to what extent we can take something from ERA to DRA comparisons, it may say more about how a team’s manager handled the transitional innings between the starter and the bullpen than it does about the pitchers themselves.
As you might expect, whether he may have tried to push a starter too far or not, A.J. Hinch pulled the right strings most of the time when they did get into a jam late in an outing. Every starting pitcher’s ERA was better than their DRA, several of them drastically. When he took them out of games with runners on base, the relief corps clearly did a really good job with those inherited runners.
However, this isn’t totally without meaning for the individual starters either. It’s just a bit too vague an indicator to project from. Those with the biggest separation between ERA and DRA, like Casey Mize, Tyler Alexander, Wily Peralta, or Matthew Boyd, left the game in some trouble fairly often, and were picked up by the bullpen cleaning up inherited runners quite a bit. Perhaps if they stayed in those situations, they would’ve done as well as their relief did, but there’s no way to know. It’s also possible that in some cases, they were left in games too long and allowed to start innings they couldn’t finish.
You can take this a number of ways, and it generally references only a small amount of appearances in which a starter was replaced mid-inning. Mize, for example, came out of just six games of his 30 total starts mid-inning which is why it isn’t something we tend to put a lot of weight on. It’s just an interesting comparison. We can literally home in on those six games and see what specifically happened, if we want to.
In the specific case of Mize’s ERA/DRA difference, we’d mostly point to the fifth inning during their August 10 matchup with the Baltimore Orioles. Mize left the game in the fifth with men on first and third with one out, and Kyle Funkhouser stranded both runners with a pair of ground balls to third base. The run expectancy in the situation Funk inherited was 1.14. Five other times Mize came out of the game in a lower leverage scenario.
Still, some of these DRA marks do suggest that Mize, for example, was pretty fortunate to post a 3.71 ERA. FIP would say the same thing, but based on his strikeouts, walks, and home runs allowed, rather than by presuming that he won’t always have so many good relief performances to back him up.
Sifting the DRA of relievers seems less interesting than with starting pitchers. The samples are much smaller, and we find less utility in ERA there to begin with, let alone the type of adjustment DRA makes. One meltdown inning for a starter is much more quickly erased than for a reliever, and so typically we lean even more on the peripherals. Run expectancy metrics like run expectancy 24 (RE24) are also useful.
Like any good statistic, DRA helps you tell a player’s story. Mostly, it’s just interesting to compare. Perhaps DRA speaks more to overall management of the pitching staff more than it means for any individual starting pitcher. We’re still going to use fielding independent pitching (FIP), seasoned with batted ball data to project pitchers for the 2022 season, of course, as it’s proven to be effective at projecting a pitcher’s future ERA, certainly far more than ERA in one season predicts ERA in the next. But as a measure of a bullpen’s depth and quality, and a manager’s facility for picking the right reliever in the right moment, to avoid his starter’s blowing up late in an outing, DRA suggests that the Tigers’ middle relief, and Hinch’s timing in inserting them into the game was pretty good overall.