So Shaun Marcum wasn’t very good last night.
There’s been a powerful counterargument to the simple idea “Shaun Marcum’s been terrible,” which I highlighted in this Jon Presser piece earlier this week. And it still holds: his FIP is just 3.27, and his peripherals are quite good, 2.3 walks per nine, and 8 strikeouts per nine.
However, another stat caught my eye on Twitter last night: Marcum has allowed hitters a paltry .548 OPS with nobody on. But with runners on? 1.073.
That’s a huge gap, and one that doesn’t remotely represent a typical split. The National League, as a whole, is allowing a .686 OPS to hitters with nobody on, .735 with men on base.
But is this second stat just more indication that Marcum has been unlucky? Or is there some degeneration in his ability to pitch from the stretch?
Last season, interestingly, the split wasn’t as big, but it was still fairly massive. Marcum allowed a .641 to hitters with no one on, .845 with runners on base. And Marcum’s line drive rate has steadily increased over four years, but stayed fairly steady through 2011, as did his none on/runners on split. Then, at the same time his runners on numbers cratered, his line drives increased dramatically, to 23.1 percent in 2012 and 24.4% in 2013. For his career, he’s at 19.3 percent.
So what we’ve seen from Marcum is a significant increase in line drives, and they’ve been happening with men on base, and they’ve been happening for two years now. That could be luck. But it feels to me like something more, and I’d like to have a better sense of what changes for Marcum out of the stretch.
As far as bouncing him from the rotation: Adam Rubin intelligently points out doing so saves the Mets money by denying incentives, rather than a move that they’d make to justify his contract, which itself is a relatively small $4 million guaranteed. The smart money is on the Mets not getting the chance, given the inherent fragility of pitchers, and that both Dillon Gee and Jon Niese are pitching through injuries.
But simply hoping Marcum turns it around, thanks to FIP, might not be the logical conclusion, either.