Mediating projections, Part II: T.J. House

Welcome back to Mediating Projections, wherein I apply shoddy methods and reach arguable conclusions regarding the forecasted futures of certain players. For more on my purpose here, see the first installment. Today, I'll look at T.J. House, who may or may not sustain his 2014 breakout. Unlike Julio Teheran, whom ZiPS loved and Steamer didn't particularly like, House receives a much better appraisal from Steamer than ZiPS:

Projection	IP	K%	BB%	HR%	BABIP	LOB%	ERA	FIP	WAR	WAR/200
Steamer	96.0	16.9%	7.0%	2.0%	0.305	69.8%	3.91	3.73	1.1	2.3
ZiPS	123.2	15.4%	7.8%	2.5%	0.320	68.2%	4.80	4.47	0.3	0.5

Will House have average numbers in 2014, or replacement-level ones? Let's investigate!

Strikeouts

As a minor leaguer, House fanned a decent amount of batters: In 198.2 Triple-A innings across 2013 and 2014, he posted a 17.6% K%. When he arrived at the show, however, his 18.7% clip handily surpassed that. Because most pitchers get worse at a higher level, the projections prognosticate some regression this year. But the extent to which he'll do so varies, as ZiPS thinks he'll plummet much more than Steamer does.

To solve this dispute, let's look into his peripherals from 2014.  As John Sickels noted, velocity gains on his primary pitch precipitated House's breakout:

House's fastball has great sinking action and always has, but it's shown more velocity this year, clocked as high as 94 MPH and averaging 90-92. This is three or four MPH higher than the reports from 2011, 2012, and 2013, and closer to what he was doing back in 2009 and 2010 when he was pitching well in A-ball.

Now, here's the funny thing about that velocity. As the season went along, it started to dissipate:

You'd think this would coincide with fewer strikeouts, right? That's not the case, though — he actually punched out more batters later in the year:

Month	K%	xK%
May	16.4%	18.4%
June	14.1%	15.5%
July	13.3%	16.3%
August	22.6%	20.2%
September	23.9%	19.8%

As the third column shows, fortune (both good and bad) effected that late-season surge; nevertheless, there's still some change there.

So he can blow batters away without velocity on his primary pitch. How? With another pitch — one that didn't appear until 2014. I'll refer again to the wisdom of Sickels:

...it looks like he's made substantial progress with his breaking stuff, refining a previously non-distinct breaker into a hard slider that eats lefties and an occasional softer curve.

That slider caused a whiff 22.2% of the time he threw it, much higher than the 13.0% baseline for the pitch. In no small part because of that, it was worth 1.32 runs above average, 19th in baseball among those who threw it often*. With that deadly pitch sticking around, House strikes (no pun intended) me as someone who would live up to Steamer's expectation.

Walks

During his smashing debut, House didn't just improve vis-à-vis strikeouts: His walk rate — which sat at a mediocre 8.1% in Triple-A — suddenly dropped to 5.1%. Given that he got rather lucky in attaining the latter mark (his expected walk rate was 6.0%), it probably won't happen again; however, Steamer and ZiPS don't entirely agree on how much wilder he'll become.

In my view, free passes won't plague House much in 2015. For one thing, he's avoided them in the past (the aforementioned AAA number notwithstanding): Chris St. John's JAVIER gives him a walk rate z-score of -0.13 for his minor-league career. While that obviously doesn't best the average by much, it still shows that he has the ability to outperform his peers. Additionally, he'll get some help from his battery mates. Yan Gomes and Roberto Perez — both of whom grade out as above-average pitch framers — will return this year, so the strike zone will remain large.

But perhaps most important is the pitches with which he posted that low walk rate last year. He didn't derive it from pounding the plate, since his 47.5% Zone% was actually below the average for starters. Rather, he made hitters swing when he threw unhittable pitches, with a 35.3% O-Swing% (well above average). And the sinker that lost all that velocity didn't really contribute to it — hitters offered at it only 22.0% of the time it was out of the zone. Instead, that dastardly slider, and a surprisingly effective changeup, came in to save him. With incredible O-Swing rates of 51.0% and 50.9%, respectively, these pitches should give him a good shot at fulfilling Steamer's projection.

Home Runs

This constitutes the majority of the discrepancy between ZiPS and Steamer. Whereas the former thinks he'll allow them at the same rate in 2015 that he did in 2014, the latter predicts a significant increase in efficacy. He almost certainly will have a higher walk rate and a lower strikeout rate, so round-trippers could make or break his season.

House garners ground balls at an extraordinary rate: 60.9% of the balls hit off him in 2014 didn't go airborne. While regression to the mean tells us that he'll probably have a lower clip in the coming campaign, that sinker will keep the ball low and the number high. Moreover, a 17.9% HR/FB% — which House put up in 2014 — will almost certainly not occur again, giving more credence to the low-homer estimate.

On the other hand, we might not feel so confident, once we examine the batted-ball data off House. Against him last year, batters hit their fly balls an average of 301.4 feet; had he qualified, that would have led the majors. Despite the small sample size, that's a damning statistic for his care — and not an altogether surprising one. As a study by Matthew Murphy delineated, ground ball pitchers tend to sacrifice more long balls than we'd think on account of their inability to accumulate pop ups; House's 2014 IFFB% of 3.6% only reinforced that stereotype.

Even with all of those ground balls, House can't help what happens when hitters put it in the air. Together with a home ballpark that's been more homer-friendly in recent years, this has a good chance at turning him into a moon shot machine; for that reason, I'll pick ZiPS in this area.

Batting average on balls in play

When hitters didn't knock the ball out of the park against House, they still excelled — only eight other pitchers had higher BABIPs against than his .332 mark. Because luck strongly influences this statistic, a number that extreme most likely won't recur. So just how far should we expect it to fall?

Not very far, probably. House generally keeps the ball on the ground, which helps him avoid long balls, since only fly balls can leave the yard. The problem with that approach, though, is that grounders go for hits more often than flies do, meaning ground ball pitchers will post higher BABIPs. Murphy's study illustrated this, as did House in the minors, with a BABIP of .309 in his tenure there (.334 in Triple-A).

A quality defense behind a pitcher can make up for his deficiencies, but House won't benefit from that either. Cleveland's projected infield — which, obviously, tends to field ground balls — features nary a positive force with the glove: Carlos Santana, Brandon Moss, Jason Kipnis, Lonnie Chisenhall, and Mike Aviles have all struggled on the diamond. Only Jose Ramirez and Francisco Lindor can offer some hope, and even they might not hit well enough to justify a roster spot. With this unimpressive bunch backing him up, House will probably post a BABIP more in line with ZiPS.

Strand Rate

Lastly, we'll look at an underappreciated element of House's 2014 success: good results in the clutch. Among 149 pitchers with at least 100 innings, his 78.9% LOB% ranked 15th. Neither projection system thinks he has a chance of repeating that, but Steamer thinks he'll do much worse than ZiPS does. With which should we place our faith?

For one thing, House struck out the same amount of batters, regardless of the presence of men on the bases — and that works in his favor, because the average pitcher becomes less effective with the pressure on. He also walked more hitters with runners on, but so does everyone (even when controlling for the intentional ones). By output and by peripherals, he consistently did well when the going got tough.

Furthermore, House throws with his left hand, which gives him an advantage in the running game. Indeed, out of 164 possible opportunities to steal facing him, runners took off a mere eight times in 2014; that 4.9% rate sat comfortably below the 5.7% league-wide figure. And if they do try to swipe a bag or two, Gomes and Perez will be able to cut them down a fair amount of the time.

By one traditional measure of pitching misfortune, House will probably remain hapless; but by the other, he shouldn't fall quite as much. If his tremendous BABIP split (.384 with the bases empty, .263 with runners on — and that came with identical line drive rates and more popups for the former) crashes down to earth, as it probably should, it'll take his strand rate with it, but it should only be to the level that Steamer foresees.

Conclusion

Like last time, Steamer and ZiPS take three and two categories, respectively. Thus, I scientifically lazily conclude that House will probably fall closer to the former — which is to say, he'll regress some overall, but he'll produce nonetheless. The AL Central race might come down to the wire this year, and House could give the Tribe the extra boost it needs to rise above.

. . .

All data courtesy of FanGraphs, Baseball-Reference, Baseball Heat Maps, and Baseball Prospectus.

Ryan Romano is an editor for Beyond the Box Score. He also writes about the Orioles on Camden Depot and on Camden Chat that one time. Follow him on Twitter at @triple_r_ if you enjoy angry tweets about Maryland sports.