As a kid, if one of my favorite pitchers (read: Bobby Jones ... yes, both of them*), had an off game, I'd often default to "well, the team he played against was really good" when explaining away a poor performance. It's logical, and it gives a great excuse. And it wasn't just useful for me! Analysts everywhere have used the quality of competition to give guys more or less credit, saying stuff like "well, if he can pitch in the AL East, then he can pitch anywhere" or "Roy Halladay should get extra credit for pitching in the AL East" during the AL East's heyday.**
* - Note: 2000! Great year for Bobby Joneses (Joneii?) on the Mets. At least in terms of sheer number of Joneii, if not performance.
** - Note: This heyday is officially over. Until probably 2015.
Digressions aside, time has passed and I've learned more about baseball. With that, comes a little more understanding. Of course pitchers need more credit for pitching in the AL, versus the DH-less NL. There's even more to accounting for the quality of opponents than this -- our own Stephen Loftus does this at a granular level for most of the starting pitchers in baseball.
Stephen's metrics -- oaRA and oaFIP -- are useful tools in our attempt to quantify pitcher performance. What would Clayton Kershaw's RA9* have been if you take into account how good (or bad) his opposition was?
* - Note: RA9 is a much, much better version of ERA. Just sayin'.
According to this data, Kershaw has his RA9 adjusted down slightly due to his performance against the quality of the opponents he faced. The teams he pitched against were slightly better offensive teams than Kershaw's "true" RA9 gives him credit for -- he faced tough competition. So when we adjust for opposition, we have to give him a little extra credit.
As far as oaFIP goes, Kershaw pitched against teams that were slightly more likely to perform worse in the three areas FIP accounts for: strikeouts, home runs, and walks. Kershaw, for that reason, loses a little credit in oaFIP, and sees his FIP of 1.81 rise up to an oaFIP of 2.12.
The gory, Bayesian math of how Stephen assigns credit for a pitcher's peripherals can be found in either of his two original articles on the subject, and they're well worth a read. But for now we'll focus on how we can apply these to use in pitcher comparisons ... and comparison is the best way to use oaRA9 and oaFIP.
Why? Because, as Stephen has pointed out in his previous work, the best use for oaRA is to compare two (or more) pitchers with comparable rate stats, to take a look at which ones may have benefitted from facing weaker competition. Just using oaRA and oaFIP in a vacuum can lead to an issue or two, given that pitchers who give up more runs may see their oa-metrics skew a little more heavily.
Anyways, how about a quick comparison of two pitchers with identical RA9 and FIP numbers?
James Shields vs. Francisco Liriano
A great example of how two pitchers may differ, is the case of "Big Game" James Shields vs. "No Real Games Of Consequence" Francisco Liriano. These two pitchers had identical rate stats for RA9 and FIP during 2014 -- both guys sport a 3.59 FIP and a 3.77 RA9. Their respective FIPs don't differ at all, even when adjusting for the quality of their opponents. Both pitchers see a small shift between FIP and oaFIP to the tune of about 0.1 "runs" of oaFIP.
As you can see, their adjusted FIP remains the same when accounting for the quality of the opposition -- that means you'd expect the strikeouts, the walks, and the home runs not to be too terribly different based on the quality of the teams and hitters they battled.
But their runs allowed? That's a different story entirely. According to Mr. Loftus's formula, Shields had faced slightly more difficult offensive teams, and if you adjust his runs against to account for that, he could get about a fifth of a run worth of credit. Liriano, on the other hand, faced inferior offensive competition. Adjusting his RA9 adds something like half a run to his rate per nine. And in the end? That's an oaRA difference of just under a full run -- quite the difference between two players with exactly the same FIP and ERA.
Perhaps you can chalk the difference up to AL vs. NL, or perhaps it has to do with Shields facing the Tigers five times this season -- they were the No. 1 offense in baseball by wRC+. Regardless, if you were shopping for a free agent starting pitcher this offseason, and comparing Shields to Liriano, the raw rate stats of ERA and FIP might indicate some level of similarity, when really there might be a bigger difference, even when it comes to 2014 performance, than one might have imagined.
Scott Kazmir vs. Sonny Gray
This one's a fun one. Two pitchers for the same team: both Gray and Kaz logged more than 190 innings for the
Sabermetric World Champion Oakland Athletics. Of course, the Athletics played the same teams throughout the season, but that doesn't mean that these two pitched against the same teams and lineups. Both starters had good RA9 marks, but when adjusting for opposition, well, wow Scott Kazmir. Kaz pitched against teams who were counted on to score many, many fewer runs, making Kazmir's oaRA vault up to something over 5.00. Meanwhile, Gray's oaRA remains flat. It's a bit of a shock to see a run and a half of difference between Kazmir and Gray's oaRAs.
Adam Wainwright vs. Stephen Strasburg
The battle between ERA (or our stand-in, RA9) and FIP, in table form. Strasburg has the killer stathead reputation, striking out all comers, while Wainwright is a runs-allowed machine, keeping opposing bats quiet like some sort of evil librarian.
Waino and Stras already had very different RA9 numbers, before accounting for how much their opposing teams scored runs. But when you adjust? Whoa buddy. Wainwright's oaRA drops half a run from his RA9, while Strasburg sees his rise more than three-quarters of a run. The end result is a massive split between the two in oaRA ... more than two full runs per nine innings!
When it comes to adjusting the FIPs, there's no magic there. Wainwright sees a rise in his FIP, Strasburg's remains roughly the same. But those RA9 numbers made me think a little bit harder about which of the two pitchers really had the better season in terms of run prevention.
Chris Young vs. Roberto Hernandez
From the depths of the FIP leaderboards, they appear! Guys who get no strikeouts! The league's pre-eminent flyball pitcher and a guy who used to be a groundball wizard but now, well, let's just say the White Council broke his staff.* Where both of these pitchers had very low non-adjusted FIPs (and pitched a similar number of innings), they had very different RA9 numbers, as well as very different narratives through the season.
* - Note: Tolkien term.
As you can see, Hernandez doesn't see his numbers shift much when accounting for strength of opposition, but Chris Young really does. His vaunted RA9 -- a number than led many to believe he was actually a pretty good pitcher in 2014 -- shoots up nearly a full run when accounting for quality of opposing offenses. I suppose the teams Young pitched against were already pretty poor run-scoring teams. At the same time, adjusting FIP gives us a number that's not quite as bad as that un-adjusted 5.02 FIP might have you think.
Anyways, that's a little bit of fun that you can have with Stephen's oaRA and oaFIP leaderboards. Accounting for the strength of a team's opponents -- especially when done in a rigorous, mathematical way, instead of just ballparking* it -- is a great way to perform a quick sanity check on a pitcher's performance, or help you start to dig deeper into the "why" when looking at end-of-year stats.
* - Note: Baseball term.
. . .
All statistics courtesy of FanGraphs.
Bryan Grosnick is the Managing Editor of Beyond the Box Score. You can follow him on Twitter at @bgrosnick.