RHP Joe Kelly
Joe Kelly would have found his name in the Locks section if he were still pitching the way he did in the first half when he posted a 1.49 ERA before the break. Kelly hit the disabled list in July and hasn’t been the same pitcher since. He owns a 4.98 ERA over 21 2/3 innings since returning. His strikeout rate has actually gone up but he also walked 11 batters over that span. Kelly still throws an imposing fastball that can reach triple-digits on the radar gun but he needs to show better control. His month of September was much better than August when he initially returned from the DL, so at least he’s trending in the right direction.
If Kelly can’t be trusted, who else would take this final spot? Brandon Workman was a candidate before stumbling down the stretch. He allowed a pair of runs in each of his last two appearances and owns a 6.10 ERA since the start of September. Heath Hembree is a serviceable right-hander who has logged a significant workload this season but his ERA is nearly a full run higher than last year. Matt Barnes has logged more innings than any Red Sox reliever but he’s posted a 4.30 ERA in the second half and a 6.75 ERA in September. It’s been over a month since Farrell trusted Barnes enough to use him in a tight ballgame that wasn’t decided in extra-innings.
Any of those pitchers could be options to fill the final bullpen spot but I’m leaning toward Kelly for the upside he brings if he can recapture his early season form.
The wildcard in this scenario is that since the Red Sox will only need four starting pitchers, the odd man out of the rotation could end up in the bullpen. Between Eduardo Rodriguez, Doug Fister and Rick Porcello, would any of them prove useful in a relief role? Possibly, although I don’t view any of them as a better option than Kelly.
The only reason to keep one of them in the bullpen is if Farrell expects to have a short leash for their Game 4 starter, in which case having a pitcher on standby who can handle a starter’s workload could prove essential.