by John Viril—
Kansas City Royals ace James Shields has been struggling for awhile now.
He dominated opponents until this May 27 start against St. Louis. However, he’s had a much harder time ever since. Though he’s had good overall statistics until Tuesday’s 6-run debacle against the Twins, his numbers indicate declining effectiveness.
Up to that St. Louis start, James Shields had a 2.47 ERA, .247 BA against, and a .580 OPS against. In the 14 starts since, Shields has a 4.11 ERA, .292 BA against and .793 OPS against.
That’s a big decline.
Both Jeff Zimmerman of Royals Review and Ben Nielson have written about this issue, but both failed to identify the cause. Nielson wrote about the problem, but admited he could not find a satisfactory answer. Zimmerman tabbed a varying release point as the culprit, while noting that there had been no decline in velocity.
Neither, however, mentioned that Shields has lost about 4 mph of velocity on his cut fastball—which we can see on this velocity chart from fangraphs.com.
The chart shows the velocity of James Shields’ cut fastball as it progresses through the last 3 seasons. Not only do we see a steep decline through 2013, we also notice that this decline is not normal for Shields because we don’t see the same thing happening in either 2011 or 2012.
If we look at his game logs, we see that his declining cut fastball velocity correlates very well with his declining strikeout rates:
Month K% Cut FB Vel
April 23.6 90.7
May 23.3 88.5
June 18.8 87.1
July 14.7 85.9
Now, this evidence looks pretty convincing. But, there’s a good reason Zimmerman does not dig into individual pitch velocities. Pitchf/x is a relatively new technology and the pitch identification algorithms are getting regularly tweaked. As a result, pitch f/x often identifies the same pitch differently—even within the same season.
In short, pitchf/x data could be showing a velocity decline on James Shields’ cutter not because of any real change, but instead because it’s identifying a slower pitch as a cut fastball. Indeed, if we look at his pitch type frequency data, it shows that Shields threw 5% cut fastballs and 15.4% sliders in 2012 while he’s used 14.5% cut fastballs and 4.6% sliders in 2013. That very much looks like pitchf/x is shifting how it recognizes Shields’ “slider” vs. a “cut fastball”.
Except for one thing: dig into Shields’ velocity data on Texas Leaguers and you will find that pitches identified as cutters or sliders tend to have an identical velocity. The only pitch slow enough to pull down average velocity while having a trajectory that could be confused with a cut fastball is Shields’ vaunted changeup.
I suspect some of Shields’ changeups aren’t moving enough and are getting recognized as cut fastballs by pitchf/x. That might explain his declining strikeout rates. Hitters aren’t whiffing as much because James Shields is throwing too many slow pitches that are relatively straight. Fangraphs shows that opponents’ out-of-the strike zone contact rate (O-Contact%) has zoomed to 67.4% in 2013 as compared to Shields’ career norm of 58.4%.
The bottom line is that hitters aren’t chasing and missing against James Shields right now. Whether he can return to form down the stretch remains to be seen.
Get more Kansas City gear in the Royals Store