Yesterday's game was painfully familiar for Rays fans of late. A strong performance by a starting pitcher was lost when the bullpen—and in particular the closer, Brad Boxberger—coughed up the lead in the ninth. Jared will have a more in depth look at Boxberger's struggles coming out tomorrow, but I wanted to whet your appetite with a quick visit to the WPA leaderboards.
The purpose of this visit is simple. Who on the Rays staff gets most credit for the games they have won? Who gets the blame for the games they have lost?
What is WPA?
First, a note about Win Probability Added (WPA). Of all of the stats in the sabermetric toolbox, it's probably the most easy to misuse. That's because it does a lot of things poorly. It doesn't tell you how good a pitcher is. It doesn't tell you how good he will be. All it does is tell a story.
Given his ability to affect the game (based on the situations in which he's pitched), how much has the pitcher contributed to either the wins or the losses? It's very context-dependent. For instance, striking out the side with the winning run at third base counts for a lot of WPA. Striking out the side with the Rays already down by five runs results in almost none.
What WPA is really good for is checking on our memory. It tells us who was pitching in the moments when the Rays won or lost the game; who should get the credit and who should get the blame. If you want to know, in most direct sense, why a team either has a winning or a losing record, you need WPA.
Leverage is related to WPA, in that it tells you how important the playing time of each player has been to the outcome of the game. Average leverage is one, higher numbers are more important situations and lower numbers less important. All stats are from FanGraphs.
Pitchers Who Have Won Games
Chris Archer: 1.65 WPA (13.76 positive, 12.11 negative, .99 leverage)
The first name on this list is the least surprising. Chris Archer has been very good this season, and his WPA is merely a function of that. Like most starting pitchers, his leverage hasn't been very high, but because he's been consistently good and has pitched a bunch of innings, he's been the pitcher who has most helped the Rays to win as a team.
Brandon Gomes: 1.57 WPA (5.68 positive, 4.11 negative, 1.41 leverage)
I bet most of you didn't see this one coming. I didn't. Brandon Gomes has been a favorite whipping boy for the commenters on this site this year, but while it's true that his FIP and xFIP have been unimpressive, his 3.31 ERA does just fine. What's more, he's generally been called on to pitch in high-leverage situations, and with the pressure on, he's succeeded.
I'm not going to argue that Brandon Gomes should be the closer or anything crazy like that, but he's done his job well enough (better than most), and there are other metrics that don't think his ERA is entirely off-base. Maybe people should stop asking for him to be DFA'd all the time.
Xavier Cedeno: 1.09 WPA (3.19 positive, 2.10 negative, 1.04 leverage)
This one is less surprising than Gomes, because specialists tend to post good WPA marks, and Cedeno has been a very solid lefty-specialist out of the Rays bullpen. His average leverage mark doesn't tell the full story, because the average leverage when Cedeno has entered the game is a bit higher at 1.34. That means that he's been called on to face a lefty with the game on the line, and more often than not has sat that lefty down, and the fact that he's gotten that out has caused his next couple batters (or his second inning) to be less pivotal.
Both Cedeno and to a lesser extent Gomes are arguments for relief specialists. The manager has control over who his relief pitchers face, so having guys who are able to succeed in one specific situation, and then putting them in that situation, wins baseball games.
Jake Odorizzi: 0.98 WPA (10.91 positive, 9.93 negative, 1.01 leverage)
Odorizzi his essentially in the same position as Archer. He's a starter, so his high WPA mark comes from consistent good pitching, rather than a small number of successful high-leverage outcomes. Maybe there's someone out there who still likes to use win-loss record and points to Odorizzi's 7-8 mark as evidence that he's not a winner.
Don't do that. If you insist on considering context, at least consider it accurately. Based on what he's done and on how he's "pitched to the situation," Odorizzi should have a winning record right now.
Jake McGee: 0.81 WPA (4.40 positive, 3.60 negative, 1.73 leverage)
Jake McGee is the best reliever in the Rays bullpen (not to mention one of the best relievers in baseball), and he's pitched like it this season. His overall leverage, though, is only third-highest on the team behind Boxberger and Kevin Jepsen. This is where the bullpen management should get called into question. Boxy's troubles are a whole other issue that can be discussed irrespective of McGee, but simply switching their usage would likely have squeezed another win or so out of this season.
Alex Colome: 0.69 WPA (7.31 positive, 6.66 negative, .99 leverage)
Colome's position here is misleading, because it's actually been the tale of two seasons. As a starter he threw 69 innings of 4.70 ERA ball with peripherals to match. He cost the Rays 0.41 games compared to an average pitcher, according to WPA. But since being moved to the bullpen in early July he's been a different pitcher, with a 1.09 ERA, a 1.69 FIP, and a 2.98 xFIP.
Over that time in the bullpen he's won the Rays 1.10 games above what an average pitcher would have (that would place him third on this list). Over that time, his average leverage has only been slightly above average (1.14). It's time to get this guy into the game when it's on the line.
Kevin Jepsen 0.49 WPA (6.41 positive, 5.92 negative, 2.03 leverage)
Jepsen (these stats include his time with the Twins) has been Brandon Gomes writ large. His leverage mark is the second highest on the team behind only that of "closer" Brad Boxberger. His peripherals are not actually that good, but a 2.35 ERA gets the job done.
After Jepsen comes a bunch of guys, including Nathan Karns, Erasmo Ramirez, and Drew Smyly, whose WPA is near average on the season, but let's skip down to where to start assigning blame.
Pitchers Who Have Lost Games
Enny Romero: -0.47 WPA (0.39 positive, 0.86 negative, 0.36 leverage)
This one feels like picking on him, because Romero has only thrown 22.0 innings, and they've mostly been in mop-up duty. Still, he's managed to hurt the Rays chances of winning even while mopping things up.
That being said, Romero's strikeout and walk rates are better than his 5.32 ERA, so hopefully he continues to get chances and to adjust as the year goes on.
Steve Geltz: -0.89 WPA (5.46 positive, 6.34 negative, 1.25 leverage)
Steve Geltz is an average pitcher. Both his 3.78 ERA and his 3.94 FIP are right on the league average. But after a clutch start to the season, Geltz has tended to be a bit worse in the more important situations, and a bit better in the less important ones, and that's why he comes in here as losing nearly a full game compared to the average expectation. Sequencing like that is one reason why the Rays won't be making the playoffs this year.
On the other hand, giving above-average leverage to an average pitcher also seems like a good way to not make the playoffs.
Matt Andriese: -1.00 WPA (3.74 positive, 4.73 negative, 0.90 leverage)
It was never in the plan for Andriese to pitch in the majors this season, so I'm not going to lose too much sleep over him losing the Rays games.
Brad Boxberger: -1.98 WPA (8.68 positive, -10.65 negative, 2.10 leverage)
This one though, I will lose sleep over. Brad Boxberger has been given the highest leverage on the team, and while his 3.81 ERA isn't terrible, it's not good either, and it might have been better than he deserved.
As I said earlier, we'll have a full analysis of Boxberger's struggles coming out tomorrow, but this reinforces the point that those struggles are a big reason as to why the Rays will be watching on the TV from home come October.
Matt Moore: -2.03 WPA (2.24 positive, -4.45 negative, 1.01 leverage)
This is just sad. Remember when we thought, "Well, all of these injuries to our pitchers are rough, but at least we'll get Matt Moore back late in the season, and if we're still in contention, that will help?"
This is not "help."