James Shields, Better But Different

We’ve talked about James Shields a lot this postseason, and most of it hasn’t been all that positive. “Big Game” James had been lousy throughout this magical Kansas City run, and he was particularly bad in Game 1 of the World Series, getting pounded with line drives left and right and leaving after three innings. In Game 5, he went up against Madison Bumgarner again, and again he lost, mainly because Bumgarner was outstanding, to the point that we’re now talking about him in a historic context.

The big picture view there is that Shields has started two World Series games against Bumgarner and lost both, likely ending his Royals career and helping to put his team in a 3-2 hole headed back to Kansas City. That’s a factual statement, but it also misses something that was largely overshadowed by Bumgarner’s dominance and more confounding Ned Yost decisions: Shields was actually pretty good last night. As the indispensable Daren Willman of Baseball Savant noted, Shields’ 21.2% swinging strike rate in Game 5 was the best any starter had this postseason, topping Zack Greinke’s NLDS start.

It’s fair to note that Lorenzo Cain’s fantastic catch on a Hunter Pence ball in right field saved Shields at least one run and perhaps two, but there were also some questionable plays by Alcides Escobar and Jarrod Dyson that didn’t go down as errors, so, noted and moving on. Shields can’t control his defense, so let’s focus on what he could control.

Here’s what that meant: Shields pitched differently than we’d seen him pitch as a Royal, and perhaps differently than he ever has. It’s actually a little terrifying to think that a pitcher who has been very good for many years would change his approach in Game 5 of the World Series. Fortunately for him, it worked.

* * *

We’ve known for a while that the current Shields hasn’t been the “classic” Shields. For years, he’d been known for his killer changeup, the one that had so often made quality hitters look silly. From 2011-13, it was rated as the best righty change in baseball, and it regularly appeared at the top in those individual seasons or even going back further than that. But this year, the change fell apart. It was awful for most of the year, as Jeff wrote, before seemingly coming back in September. It’s probably a very overrated pitch, as Eno wrote last week.

That’s echoed by unnamed scouts from a Nick Cafardo Boston Globe column over the weekend:

Scouts who have seen Shields over his career feel he’s changed from a fastball/changeup pitcher to a fastball/cutter pitcher. Scouts remember the changeup being unhittable and now the cutter has taken over, and at times it is hittable. Still effective, but there is some bewilderment over Shields’s repertoire and where he goes from here after a poor postseason.

“Changed to a fastball/cutter pitcher,” said the scout. Apparently so. Last night, Shields threw the cutter at a higher percentage than he ever had in his entire career, per Brooks:

Shields highest cutter usage (%) games, 2008-14
Game	Fourseam	Sinker	Cutter	Curve	Change
KCA@SFN (10/26/14)	19.57	16.3	38.04	13.04	13.04
CHA@KCA (4/6/14)	21.78	4.95	37.62	12.87	22.77
KCA@BAL (4/27/14)	20.79	20.79	35.64	6.93	15.84
TBA@BOS (4/16/12)	15.24	0	35.24	21.9	27.62
KCA@BAL (10/10/14)	27.78	16.67	33.33	4.44	17.78
KCA@CLE (9/11/13)	34.58	2.8	32.71	6.54	23.36
CHA@KCA (5/6/13)	24.51	13.73	32.35	8.82	20.59
KCA@MIN (4/12/14)	19.13	10.43	32.17	7.83	30.43
TBA@TOR (7/28/08)	26.6	8.51	31.91	4.26	28.72
TBA@BAL (5/13/12)	23.08	1.92	31.73	20.19	23.08

He didn’t quite throw the change less than he ever had before, but considering that he’s made 244 starts (postseason included) since 2008, that we’re talking about a bottom-six game here is still pretty noteworthy.

Shields lowest change usage (%) games, 2008-14
Game	Fourseam	Sinker	Cutter	Curve	Change
KCA@SLN (6/3/14)	35.42	21.88	30.21	4.17	8.33
KCA@CHA (6/15/14)	27.27	24.55	30	8.18	10
TBA@DET (9/28/08)	70	0	10	10	10
KCA@TBA (8/2/09)	38.83	6.8	24.27	19.42	10.68
TOR@TBA (7/7/09)	42.16	8.82	29.41	7.84	11.76
KCA@SFN (10/26/14)	19.57	16.3	38.04	13.04	13.04
SEA@TBA (7/20/12)	16.98	24.53	16.98	28.3	13.21
CHA@TBA (10/2/08)	31.71	9.76	24.39	20.73	13.41
BAL@TBA (8/31/08)	26.25	22.5	20	17.5	13.75

So it was different. But was it better? One way to evaluate that is by whiff percentage. Other than what looks like an outlier game against the A’s back in 2010, he’s never topped the success his cutter showed last night:

Shields top whiff percentage games, cutter, 2008-14
Game	Fourseam	Sinker	Change	Curve	Cutter
OAK@TBA (4/28/10)	5.13	16.67	45.16	7.14	36.36
KCA@SFN (10/26/14)	5.56	6.67	25	25	34.29
TBA@BOS (6/5/08)	7.69	0	0	0	33.33
NYA@TBA (9/15/10)	13.89	8.33	20.83	11.76	33.33
ANA@TBA (4/30/11)	13.33	0	22.86	0	33.33
CHN@KCA (3/15/14)	4.35	0	40	14.29	31.25
TBA@OAK (7/31/12)	0	10.34	46.67	0	25
BAL@TBA (10/2/12)	6.9	11.76	36.67	27.27	25
KCA@DET (9/10/14)	3.45	0	20	33.33	25
TBA@CHA (7/22/09)	3.13	0	9.52	8.33	22.73

Of course, the cutters he threw in Game 1 weren’t very good. He threw it 13 times and got exactly zero swing-and-misses on it. After that game, he made several vague allusions to “adjustments,” and pitching coach Dave Eiland said the same during an in-game interview last night, talking about where Shields positioned his hands. Jim Bowden tweeted that Shields was “starting hands lower & bigger turn resulting in better north south direction & downward plane.” 

Maybe that’s all true. Nothing obvious shows up in the horizontal/vertical positioning charts, and it’s pretty difficult to say with certainty how pre-pitch hand placement affects a pitcher. Of course, even if there is something to all of that, there’s something more painfully obvious: Shields simply put the cutters in better spots. Below are three heat maps, from his last three postseason starts, showing where the cutters were placed against righty hitters. (These are from the catcher’s perspective.) In Game 1 of the ALCS against the Orioles, he was catching a ton of the plate. In Game 1 of the World Series against the Giants, he didn’t throw it very much, but when he did, it was all right down the pipe. And last night?

That’s how a pitch like that is supposed to work. Enough of the strike zone to make it interesting, and enough of the edge of the plate to make it hard to square up. That’s how it worked in the bottom of the fourth against his opposite number, Bumgarner:

So it’s a great sign that after a rough few weeks, Shields was able to once again be pretty effective, even though it wasn’t quite good enough against Bumgarner. He made some possible mechanical adjustments, he located better, and he used a pitch that was working far more than one that wasn’t working. Though it didn’t result in a win, this isn’t so much on Shields, like it probably was after Game 1. Bumgarner was just that good, and Yost didn’t always put his team in the best position to do anything to change that.

But no matter what happens when the Series goes back to Kansas City, Shields has likely thrown his final pitch as a Royal, barring an unexpected relief appearance in Game 7. The question now, as his 33rd birthday looms in December, is what exactly he’ll be going forward. Shields’ free agency has the potential to be one of the more fascinating of the winter. Unlike Jon Lester, he’ll have a qualifying offer attached. Unlike Max Scherzer, he’s likely past his peak, and he’s also not the Shields we thought we knew. The changeup might not be the weapon it once was. The cutter, which had so long been a negative for him, was actually one of the best in baseball this year. Trading one weapon for another isn’t necessarily a bad thing. It just might add uncertainty at a time when teams, more than anything, want to have confidence in a player they’re about to spend tens of millions on.