Letting Carl Edwards Face Bryce Harper
In Saturday’s game against the Washington Nationals, manager Joe Maddon allowed right-handed pitcher Carl Edwards Jr. to face left-handed dynamo Bryce Harper with a runner on first and one out, his Cubs leading by two in the bottom of the eighth inning. Harper hit a two-run shot to tie the game. Three batters later, lefty Mike Montgomery would allow a three-run homer to righty Ryan Zimmerman, giving the Nationals a decisive lead. Managerial decisions that lead to playoff losses tend to draw scrutiny, but this particular set of moves seems pretty defensible.
To briefly review, a few relevant facts:
- In Game 1 on Friday, Kyle Hendricks pitched seven strong innings. Carl Edwards followed with a perfect eighth inning, striking out Trea Turner, inducing a weak blooper (for an out) from Bryce Harper, and then striking out Anthony Rendon. Wade Davis pitched the ninth.
- The Cubs bullpen has four right-handers: Wade Davis, Carl Edwards, John Lackey, and Pedro Strop.
- The Cubs bullpen has three left-handers: Brian Duensing, Montgomery, and Justin Wilson.
- In Game 2, Jon Lester pitched six innings, and Pedro Strop pitched the seventh inning.
Based on both the first game and then the first seven innings of Game 2, it seems as though Joe Maddon had established a pretty clear pecking order for his bullpen, featuring Wade Davis at the top of the depth chart and Carl Edwards just below. One might place Pedro Strop third on that list, although an argument could be made for Mike Montgomery, too, depending on the matchup. In any case, if there was any doubt regarding Maddon’s feelings about Edwards, his decision to use the the right-hander in the eighth inning of Game 2 erased it.
As for whether Edwards should have entered to start the eighth, that also seems pretty sound. A pinch-hitter was going to lead off the frame, regardless, and it was going to be either the left-handed Adam Lind or right-handed Victor Robles. That pinch-hitter was to be followed by righty Trea Turner, lefty Bryce Harper, and righty Anthony Rendon — the exact same trio, in other words, whom Edwards had set down the day before.
The lack of rest likely wasn’t an issue for Edwards. On 12 occasions this year, he pitched on consecutive days; he performed at basically the same level in that sample as in his normal regular-season outings. Moreover, there appear not to have been any discernible physical effects from Edwards’ back-to-back appearances: his fastball velocity on Saturday was just as good as it was on Friday. Again, no real arguments against the decision.
In reality, Lind led off the eighth with an opposite-field single. Edwards responded by striking out Turner. That brought the lefty Harper, the possible tying run, up to the plate. Edwards gave up the homer to Harper and walked Rendon. Mike Montgomery then entered the game to face a left-handed Daniel Murphy.
Because of the way the Cubs have constructed their bullpen, using a lefty to face Harper, then a righty to face Rendon, and another lefty against Murphy wasn’t really feasible — and that doesn’t even account for the presence of the right-handed Zimmerman. At this point in Saturday’s game, Strop had already pitched and was thus unavailable. Lackey, having worked as a starter all season, was an unlikely choice to enter mid-inning. And the prospect of a multi-inning appearance from Davis was unlikely after he’d closed out Game 1 the day before.
In any event, it isn’t entirely clear that Chicago has a LOOGY-type to face Harper. The table below shows the number of high-leverage plate appearances against left-handed hitters this season for Cubs relievers.
Name | High-Leverage PA v LH |
---|---|
Wade Davis | 42 |
Carl Edwards | 22 |
Mike Montgomery | 16 |
Pedro Strop | 16 |
Koji Uehara | 15 |
Justin Grimm | 9 |
Brian Duensing | 9 |
Hector Rondon | 6 |
Justin Wilson | 4 |
Davis is the best reliever the Cubs have, closing out games against whichever batters are up in the ninth inning. After him, Edwards had the second-most important plate appearances against lefties. Montgomery was a starter for much of the year, so it’s possible he might have come out ahead if he were in the bullpen all season, but he doesn’t really have that lefty-killing arsenal. He uses a few different fastballs plus a change, cutter, and curve, lacking a typical LOOGY slider. Duensing has been with the Cubs all season and only faced nine lefties in big situations. This clearly is not his job. Justin Wilson goes fastball-cutter but has also been dealing with a sore neck and a tough September during which he struck 13 but walked 10 of 38 batters faced.
So the realistic options against Harper were either (a) sticking with Edwards, (b) going with Montgomery for the rest of the inning, or (c) bringing in Davis earlier. Given Edwards’ success the day before, the lack of a major advantage with Montgomery, and the risk of bringing in Davis early, Edwards still seems like a reasonable play. As for Harper, he does have career platoon splits, with a 112 wRC+ against lefties and a 154 wRC+ against righties. However, much of that gap is due to a fairly rough first few seasons, when Harper had very little experience against top-flight left-handers. Over the past four seasons, he has moved his numbers against same-sided pitchers up to 126 while his numbers against righties have been roughly the same. In an ideal situation, yes, you bring in a lefty to face Harper, but it’s better to have your best pitcher available pitch rather than go with a platoon just because you can.
Here are the platoon numbers over the last three years for relievers.
Year | RHP | LHP | Difference |
---|---|---|---|
2015 | .310 | .290 | .020 |
2016 | .317 | .291 | .026 |
2017 | .320 | .289 | .031 |
Avg. | .316 | .290 | .026 |
Seems pretty standard. When relief pitchers have the platoon advantage, they get more batters out. This is very general, though. What happens when we restrict the encounters only to high-leverage situations? Limiting the sample in this way removes a lot of unimportant moments and likely a lot of poor relievers, too, who aren’t as likely to pitch in these situations.
Year | RHP | LHP | Difference |
---|---|---|---|
2015 | .299 | .293 | .006 |
2016 | .314 | .307 | .007 |
2017 | .316 | .304 | .012 |
Avg. | .310 | .301 | .008 |
The difference dwindles. Does this mean the platoon advantage doesn’t exist? Of course not. Some of the difference is likely due to the fact that some of the left-handed relievers in this second sample have entered a high-leverage moment specifically to face a talented left-handed hitter. So, even if they’re creating better outcomes than their right-handed counterparts would, their raw numbers are likely to suffer a bit. But lefties aren’t the only pitchers facing lefties: good righty relievers are also facing good lefty batters, too, particularly in your standard closing situations in the ninth. The same is true against right-handed batters.
Year | RHP | LHP | Difference | HL RHP | HL LHP | Difference |
---|---|---|---|---|---|---|
2015 | .300 | .318 | -.018 | .291 | .288 | .003 |
2016 | .302 | .322 | -.020 | .298 | .311 | -.013 |
2017 | .305 | .322 | -.017 | .301 | .299 | .002 |
Avg. | .302 | .321 | -.018 | .297 | .299 | -.003 |
Getting a great platoon matchup is ideal. Absent a reliable LOOGY option, however, a talented right-hander is the best alternative — and Carl Edwards fits that description. While we could argue the merits of Wade Davis entering important situations regardless of inning, a la Andrew Miller, that seems like an unlikely deployment given the way the Cubs have done things all year long. The ninth-inning closer versus any-inning relief ace is a debate that has gone on for a while and will likely continue to, but if we already know Davis is going to be saved for the ninth, letting Edwards take the eighth and pitch to Harper seems like the right move.
Craig Edwards can be found on twitter @craigjedwards.
What Maddon should have done is let Strop start the inning. When Lind is sent up as a PH, get a LOOGY in. Then Edwards can come in for the next batter, but you get the platoon advantage over Lind (who has a big split).