Mediating Projections, Part VI: Phil Hughes

Welcome back to Mediating Projections, wherein I look at the largest differences between ZiPS and Steamer. After a two-month hiatus, the series returns with its sixth and final edition for this offseason. (Here are installments I, II, III, IV, and V.)

Our subject today is Phil Hughes, the Twins starting pitcher whose performance has fluctuated as of late. He dominated in 2014, earning a lucrative contract extension, then collapsed in 2015. While ZiPS and Steamer pretty much agree on his ERA for 2016, they diverge massively on nearly everything else:

System	IP	K%	BB%	HR%	BABIP	LOB%	ERA	FIP	fWAR	fWAR/200
Steamer	198.0	15.9%	3.6%	3.5%	.294	71.7%	4.20	4.26	2.0	2.0
ZiPS	167.0	18.5%	2.3%	2.9%	.320	68.6%	4.10	3.49	3.2	3.8

ZiPS has an optimistic view when it comes to the three true outcomes (which is to say, the areas that pitchers theoretically control); Steamer has an optimistic view when it comes to balls in play and runners on base (which is to say, the areas that pitchers theoretically don't control). This presents an intriguing contrast and one that we'll attempt to investigate in more detail.

Strikeouts

Perhaps the biggest reason for Hughes's 2014 explosion was a high strikeout rate — he fanned 21.8 percent of the batters he faced. That mark deflated to 14.4 percent in 2015, which likewise played a large role in Hughes' tanking  that year. ZiPS expects he'll rebound significantly, whereas Steamer thinks 2014 could happen again. Where should we place our faith?

When Hughes surged two years ago, he did so because of two metrics: strike rate and swinging-strike rate. Pitchers who don't throw balls will get into a lot of two-strike counts, and pitchers who rack up the whiffs will usually set hitters down on strikes. Hughes's strike rate flew to a career high in 2014, before depreciating in 2015, as we'll discuss in a moment. More importantly, though, he gained far fewer swinging strikes: His rate of those dipped from 10.3 percent to 6.4 percent.

Why did Hughes lose the ability to finish batters and generate those strikeouts? His fastball velocity fell to a career low, and his softer pitches remained stable:

The lack of heat on the four-seamer sunk its swinging-strike rate from 11.0 percent in 2014 to 5.9 percent in 2015. The smaller gap between that pitch and the curveball made the latter's whiff rate drop as well, from 9.3 to 5.1 percent. Even some modest progress from the changeup couldn't compensate for those dual declines.

In June, Hughes talked about his absent velocity, failing to come up with any real reason for it. Starting pitchers in their late 20s will tend to lose power on their pitches, simply because they've gotten older, so the fact that Hughes couldn't really come up with a solution makes some sense. He turns 30 in June, and the likeliest scenario going forward is him never gaining back that velocity. Unless Hughes can figure out some way to collect whiffs — or goes back to extreme strike-throwing — he'll fulfill Steamer's projection.

Walks

If Hughes has made his name on one thing, it's evading bases on balls. In 2014, his 1.9 percent walk rate led qualified pitchers by more than a full percentage point. In 2015, his 2.5 percent walk rate again paced the major leagues (among hurlers with 100 innings), although the immortal Bartolo Colon threatened to take his crown. Just how low will his free pass clip be in 2016 — Steamer low, or ZiPS low?

At first glance, Hughes declined significantly between those two years. According to FanGraphs' PITCHf/x data, his zone rate plunged from 61.1 percent to 56.3 percent, while his O-Swing rate dropped from 35.7 percent to 29.9 percent. The effect: he threw 73.1 percent strikes in 2014, and 69.4 percent strikes in 2015. In the most important area for walk avoidance — getting strikes — Hughes got much worse. Maybe he'll give out more free trips to first as his luck disappears.

Strikes don't explain all of a pitcher's free passes however. The rate of balls in play correlates well with walk rate, since longer plate appearances mean more chances for ball four. Over that same span, Hughes's rate of in-play strikes increased dramatically, from 29.1 percent to 34.4 percent. In the end, Hughes actually didn't fall too far behind. If anything, Hughes has gotten unlucky as of late: Per Alex Chamberlain's expected walk rate, he should have issued free passes to 1.4 percent of batters in 2014 and 2.0 percent in 2015.

Does that mean he'll keep up this rate of output? Regression to the mean would suggest not. Last offseason, I observed that players who put up incredibly low walk rates tend to get a bit worse, simply because few players can sustain amazing production like that. At that point, however, Hughes only had one ridiculous campaign to his name. He's now tallied a second consecutive one, which would seem to give him a better chance at maintaining it. It's a bit debatable, but I feel like ZiPS takes the cake here.

Home runs

Throughout his career, Phil Hughes has struggled with the long ball. Homers plagued him in New York, and when he moved to the more pitcher-friendly confines of Minnesota, they still pestered him. Hughes has a lifetime round-tripper rate of 3.2 percent; ZiPS takes the under on that, Steamer the over. How often will his pitches land up in the outfield seats?

Hughes's batted-ball profile leaves him vulnerable to dingers. For his career, 34.4 percent of the balls in play against him have stayed on the ground, which trails the major-league average by nearly ten percentage points. With all of those air balls, a fair amount will inevitably leave the ballpark. His respectable fly ball distance of 278.2 feet — about an average mark, although it's hard to tell with Baseball Heat Maps — can't really negate it. Hughes can only eschew the four-bagger, as he did in 2014, with a glut of strikeouts and some HR/FB rate magic; neither of those look likely for 2015.

As mentioned previously, Hughes does seem to have one thing on his side: his home stadium. FG's park factors assert that Target Field was the seventh-worst for home runs in 2015, while Yankee Stadium was the third-best. To this point, he hasn't taken advantage of that, but he could begin to capitalize on it this season. Of course, an extensive outfield can only do so much. When a hurler allows as many long fly balls as Hughes does, the negativity of Steamer appears the most likely.

Batting average on balls in play

Historically, a high BABIP has also haunted Hughes, albeit to a less notable degree than his home run hardships. He has a career BABIP of .301; since 2013, he's posted a mark of .318, higher than any other pitcher with as many innings. So will he continue to struggle with this in 2016, as ZiPS predicts, or will he begin to see his luck turn around, as Steamer foresees?

Hughes certainly seems to have earned his high BABIP. During that three-year period, he's allowed 30.4 percent hard contact, significantly above the major-league line. Although line drives have some subjectivity to them and can be a bit unsteady, Hughes's 23.4 percent mark over that span doesn't inspire much hope. Even his 4.2 percent popup rate can't make up for all of those easy hits. Hughes just can't seem to manage contact that well.

There's also the matter of team fielding. Last season, the Twins didn't do that well when it came to defense, finishing 22nd in DRS and 18th in UZR. The FG depth charts expect them to save 2.5 runs this year, a satisfactory sum and a bit of an improvement. Along with the general idea of regression to the mean — which posits that a more normal BABIP would await Hughes in the future — this appears to work in Hughes's favor.

Here, neither Steamer nor ZiPS is the clear winner. I wouldn't foresee a totally awful BABIP from Hughes, but an average mark just strikes me as unrealistic. In the end, Hughes's track record makes ZiPS the more likely option; even if his team catches the ball better, there's only so much they can do to combat his shortcomings.

Strand rate

In the final category, things become a little foggy. Hughes has always left a fair amount of runners on base — 71.6 percent of them, to be exact. Steamer feels that clip represents his true talent level, and ZiPS thinks he can't live up to it. Will Hughes maintain or sacrifice this facet of his game?

During good times and bad, Hughes has consistently performed better with runners on base, in every regard:

Situation	K%	BB%	HR%	BABIP	wOBA
Bases Empty	19.2%	5.7%	3.2%	.302	.325
Men On	19.6%	5.4%	3.2%	.299	.323

An average pitcher will have significantly worse strikeout and walk rates with runners on base. The fact that Hughes has done just as well regardless of the situation reflects well on him (as do the similarities in home runs and BABIP). Altogether, Hughes has the same wOBA with no one one and with men behind him, while pitchers MLB-wide will lose about ten points of wOBA in the latter situations.

What's more, Hughes has held runners pretty well. Even pitchers who fare well to the batter at the plate can come undone if the men behind them move up, but Hughes hasn't allowed that to happen. According to Baseball-Reference, runners have tried to steal in 5.3 percent of their opportunities against him, which is about the major-league average. In pretty much everything that could give a pitcher a low strand rate, Hughes has excelled, meaning Steamer wins out here.

Conclusion

Hughes gets three victories from Steamer and two from ZiPS. More accurately, though, he gets three pessimiastic takes and two optimistic ones. With a lower strikeout rate, low walk rate, and high home run rate, Hughes should work his way to a decent FIP; with a subpar BABIP and solid strand rate, he could underperform and post a higher ERA. It's not the best package out there, but on a Twins club that has an ungodly amount of prospects and future talent, it'll suffice, at least for now.

. . .

Ryan Romano is a contributing editor for Beyond the Box Score. He also writes about the Orioles on Camden Depot(and on Camden Chat that one time), and about the Brewers on BP Milwaukee.