James Shields Has Been Messed Up for a While

One of the realities of the earlier part of the season is that we notice things we might not otherwise notice. A hot streak or a slump to begin the year stands out more than a hot streak or a slump in the middle of August, because at the beginning, everything starts fresh. This is one of the reasons why people tend to overreact to early results. The numbers make it look like they’re the only results, as new seasons stand out from prior ones. As others like to remind, players streak all the time, and we typically just accept it if we even notice at all.

So if it’s easier to notice a streak at the beginning, it follows that it’s harder to notice a streak in the middle. Which means when a streak in the middle does get noticed, that means something. We’ve all noticed James Shields‘ streak. James Shields’ streak is one of the very ugliest starting-pitching streaks on recent record. It is, very genuinely, just about unbelievable.

This could be a whole post of fun facts. The numbers are that extraordinary. I’ll try to limit myself, because the fun facts aren’t the point. But, all right: over his last four starts, Shields has allowed a total of 32 runs. Jake Arrieta has allowed a total of 32 runs over his last 30 starts, covering more than 200 innings. Shields, since his last game with San Diego, has yielded a 1.441 OPS. Barry Bonds, in 2004 — when he walked more than 200 times — finished with a 1.411 OPS. James Shields has strung together four starts of turning the opposition into prime Barry Bonds. This is James Shields, of the James Shields Trade.

It’s been impossible not to notice. Even the worst pitchers don’t bottom out like this, and this has become a serious problem for a team that’s trying to make the playoffs. Shields, of course, isn’t this bad — position players pitching aren’t this bad — but maybe the most troubling thing is this isn’t just a four-start slump. It’s been a horrible, unimaginable four starts, sure, but Shields hasn’t been quite right for some time.

Here, let’s have some fun, before the analysis. I guess this could pass for analysis? Not long ago, we released improved versions of our player graphs. Here’s James Shields’ rolling-average ERA for stretches of four starts:

Numbers naturally fluctuate. Numbers don’t naturally take off for the moon. If something like this were to show up on a plot at the USGS, cities would be evacuated. It’s not at all an exaggeration to say this has been historic, especially for a pitcher just acquired intentionally, with the postseason in mind. Whatever happens to be wrong with Shields is complicated. There’s usually not a smoking gun, and right now a lot of factors have come together to make Shields the worst version of himself possible. His mechanics presumably aren’t right. His head, now, presumably isn’t right, because, how could it be? These analyses are never simple, but one thing does stand out in Shields’ record. From Brooks Baseball, let’s look at Shields’ last three years or so of average pitch velocities:

I’m going to post that again, but now with a little addition, for clarity:

Focus on the fastballs. That’s kind of abrupt. Velocities do move around sometimes, and you can see that, say, in 2014, there was a little dip in July. But then Shields recovered some. If you turn to 2015, it’s not the same. Shields lost something between May and June of last season, and he hasn’t gotten it back. He pitched decently enough in the meantime, so people weren’t horribly alarmed, but now you have to wonder. In retrospect, it looks like Shields has been messed up for something like a year.

I don’t know why. I pretty much can’t know why, not from out here. Maybe his mechanics are just badly out of whack. Shields says he feels physically fine, but you do have to wonder if there’s an injury. Something like the above took place not long ago with Greg Holland. Holland had an injury. Shields has always had durability on his side, but durable pitchers are only durable until they aren’t. Think about Bronson Arroyo. Think about even Jon Garland. There could be something in there. One way or another, this table is telling.

Everything has gotten worse. It’s almost like a switch flipped. It wasn’t that obvious in real time but it looks that way now. Beginning around last June, Shields hasn’t thrown pitches in the zone, and he hasn’t gotten out ahead like he used to. Mix in reduced velocity and you have an inferior pitcher. I don’t know how much can be rescued, but it’s not as easy as trying to get Shields back to where he was a few weeks ago. It’s more like 13 or 14 months ago.

Here’s another rolling-average plot. I don’t know how much this adds, at this point, but I made it, so I’ll embed it. You see rolling strike rates, rolling zone rates, and the rolling difference between in-zone swings and out-of-zone swings:

Shields started to lose strikes last year. He started to lose the zone last year. Now he’s getting more swings at his strikes without a corresponding increase in swings at would-be balls. This is probably a consequence of Shields falling behind, and his fastball isn’t good enough for him to get back into counts. I’m not ready to say that James Shields is toast, but he’s in need of a makeover. Or surgery. I don’t know which. But if Shields can’t get ahead, Shields can’t get outs, and he’s never been particularly strong at managing contact.

Shields never had a great fastball. He was, for a while, carried by his changeup. Somewhere in 2014, the changeup started to get away from him, and he turned to his cutter. Now the fastball isn’t real good, and the changeup isn’t real good, and the cutter isn’t real good. All that’s left is a curveball Shields struggles to throw for consistent strikes. Gradually, baseball has eaten away at his arm, and now there might not be enough left to be functional. With luck, this is simply a mechanical thing, and Shields will straighten it out, and he’ll re-find some measure of success. With pitchers, however, you don’t always get lucky. Certainly not the good kind.