I had fun last year looking at the largest DRA-RA9 differences in baseball, so I decided to do it again using the DRA run values chart at Baseball Prospectus. Similar to what I did last year, I will stick to pitchers who have pitched at least 120 innings. Obviously this will mean all full-time relievers will be excluded, but I think it would be better to focus on larger sample sizes here to minimize sample size errors.
So why is this important? DRA-RA9 is the improved ERA-FIP. RA9 is just ERA without the silly, arbitrary distinction between earned and unearned runs. It counts everything. You almost never see it in articles, but it is always used instead of ERA in serious sabermetric research.
Baseball Reference uses RA9 to calculate its version of WAR. DRA takes a pitcher’s RA9 and adjusts for everything imaginable that pitchers can and cannot control. Naturally, it factors in strikeouts, walks, and home runs, so you could say it has a FIP component to it. Unlike FIP, however, DRA is on the RA9 scale, not the ERA scale, so comparisons must always be made to RA9.
If you want to learn more about DRA, I suggest you clear some time from your schedule. DRA does have its detractors, and I once wrote a response to those detractors. You can check out my article on the overperformers here.
Sale got off to an awful start this season, settled down through May and June, and then struggled again from July through mid-August when he was put on the IL for the rest of the season due to left elbow inflammation. Thankfully, it was reported that he will not need Tommy John surgery. He finished the season with a 4.89 RA9, more than double his RA9 for the previous season, and the highest of his career by over a run.
Because of the hitter friendly confines of Fenway Park and the strong competition he faced, Sale still finished the season with a respectable, but hardly Sale-ian 2.3 WAR. I am sure the Red Sox were expecting more out of him, but at only 84 wins this past season, vintage Sale would not have made a difference.
The funny thing is that Sale’s strikeout and walk rates were in line with what one would expect. He had a 35.6 K% and 6.1 BB%, and only Gerrit Cole had a higher strikeout rate among pitchers with at least 140 IP. The juiced ball was a problem, but not an especially big one for Sale. He gave up home runs to 3.9 percent of batters faced, which is just a little higher than the league average of 3.7 percent. He has always been a little homer prone anyway, with a career HR/FB ratio of 11.2 percent going into the 2019 season.
Sale had an excellent 2.93 DRA in 2019, almost two runs lower than his RA9. Obviously his excellent strikeout and walk rates are factoring into that. In addition to that, DRA is claiming that he should have given up fewer runs given the quality of contact he gave up. Originally I was struggling with this, because according to Statcast, he had a 36.3 percent hard-hit rate, and his .407 xwOBA on contact was way higher than his actual .294 wOBA.
The component of DRA that assesses how many runs a pitcher deserved to save or give up over the average is called Hit Runs. Sale has -14.2 Hit Runs despite his hard-hit stats, so I reached out to Jonathan Judge of Baseball Prospectus for clarification. Basically, you have to keep in mind that hard-hit rate and xwOBAcon measure what actually happened on the field, while DRA determines what should have happened.
To quote Judge, unlike the stats that measured what actually happened on the field, DRA “[controls] for the inputs or quality of pitches thrown.” In other words, he still has great stuff, and that stuff deserved to get better results than it did.
Given Sale’s track record, do you think his true talent is closer to his 4.89 RA9 or his 2.93 DRA? I’d definitely go with the latter.
A 4.97 RA9 might not look very good, but it looks a lot better for a pitcher in Colorado. That was not an outlier season, either, as he has a 4.53 RA9 over the past three seasons and has averaged 3.8 WAR per year. His strikeout rate was not much better than average, but he walked only 4.9 percent of batters faced, down from 7.0 percent each of the two previous seasons. He didn’t get killed by the juiced ball, either, having given up home runs against 4.0 percent of hitters faced. If there was a HR%+ stat that adjusted for ballpark, it would probably be below 100. (FanGraphs does have HR/9+, but it is not park-adjusted.)
Márquez posted a 3.26 DRA this year. The ballpark is obviously a factor here, and he is also being credited for his low walk rate. He suffered from a low strand rate, too, and apparently the Rockies’ defense was not so great when he was pitching (DRA assesses defense only for when the respective pitcher is pitching). As with Sale, DRA determined that his stuff was better than the quality of contact that he gave up would dictate, hence the -8.2 Hit runs.
Musgrove has never been more than a back of the rotation starter. His 5.18 RA9 is not very good, but he was also pitching in front of a poor Pirates’ defense. As for his peripherals, he had a subpar strikeout rate, but he did demonstrate good control (he walked only 5.4 percent of batters faced). The juiced ball did not a appear to affect him at all, either.
DRA really liked Musgrove, giving him a 3.59 DRA. That also more than doubled his WAR, going from 1.7 at B-Ref to 4.0 at Baseball Prospectus. He is clearly getting credit for the good walk rate, home run rate, and poor defense behind him. Interestingly enough, while he did have a low strand rate this season, he also has a low strand rate of 68.8 percent for his career. It might be that he has more problems than most pitchers when he pitches from the stretch.
I was surprised that he got credited for almost 10 Not-in-Play (NIP) runs when his strikeout rate was below average, and he also hit nine batters on top of that. Also, as seems to commonly be the case, he was unlucky with the quality of contact he gave up with respect to his stuff.
This one is probably really good news for Cubs fans, because Darvish was a replacement level player in the first year of a six-year, $126 million deal. He pitched pretty well in 2019, with a 4.13 RA9 and 3.3 WAR. That is easily worth $20 million for a wealthy team trying to contend.
Darvish had an excellent 2.69 DRA in 2019. He struck out nearly a third of batters faced, but the juiced ball led to him struggling more with the long ball. He has always been homer-prone, but this year he gave up 33 HR on 4.5 percent rate. His nearly 23 percent HR/FB ratio is eye-popping and certainly fluky.
DRA gave Darvish a ton of credit for NIP Runs and Hit Runs. As mentioned, he did strikeout a ton of batters, but his 7.7 percent walk rate is not too much better than average, and he hit 11 batters. This is probably getting old now, but he also had a whopping -20.2 Hit Runs. His stuff should normally get much better results.
Thor had a down-year by runs allowed, but his 4.60 RA9 in front of that awful Mets’ defense is pretty solid. Interestingly enough, his 3.40 DRA does not give him much back for that defense, possibly because it might not have been that bad when he pitched. He does continue to have a maddeningly low strikeout rate given his velocity, with a 24.5 K% that is not much better than average.
DRA gave Syndergaard significant deductions for NIP Runs and Hit Runs. He continues to have strong walk rates, and even though he was majorly affected by the rabbit ball compared to last year, he only gave up home runs to 2.9 percent of batters faced. He suffered from a low strand rate, too. Syndergaard had -12.3 Hit Runs, but he might be the least surprising pitcher on this list in that regard given his fastball velocity. It’s nice to see that his DRA rated him so much better than his RA9, but that 3.40 DRA is also roughly a run higher than it was in previous seasons.
Special thanks to Jonathan Judge for providing helpful clarifications on DRA.
. . .
Luis Torres is a Featured Writer at Beyond the Box Score. He is a medicinal chemist by day, baseball analyst by night. You can follow him on Twitter at @Chemtorres21.