Investigating Luis Perdomo’s strange projections

Welcome back to Mediating Projections, the recurring series where we examine players with divergent projections and determine how well they’ll do in the coming season. (Trust me, it’s more fun than it sounds.) For last week’s edition, featuring Red Sox lefty Chris Sale, click here. This week, we’re taking a look at Luis Perdomo, the Padres right-hander who had an up-and-down rookie year in 2016.

As a Rule 5 pick from the prior offseason, Perdomo remained in the big leagues with San Diego for all of 2016. He struggled on the surface, with a 5.71 ERA in 146 ²⁄₃ innings; meanwhile, his peripherals — a 98 DRA- and 99 cFIP — hinted at a respectable pitcher. For 2017, Steamer and ZiPS can’t figure out which metrics represent his true talent:

Steamer has an optimistic perspective, expecting Perdomo to take a step forward in 2017. ZiPS, by contrast, predicts a nasty sophomore slump. Which should we believe?

Since both systems project identical strikeout rates, we won’t need to analyze that. For this edition, we’ll examine the four areas of disagreement — walks, home runs, batting average on balls in play, and strand rate — and come to a conclusion about which projected outcome is correct.

Walks — 7.4 percent (Steamer) vs. 7.9 percent (ZiPS)

Perdomo had decent control coming up through the St. Louis system. During his minor-league career as a whole, he walked 6.7 percent of the batters he faced; in 2015, across two levels of Single-A ball, he doled them out at a 6.8 percent clip. And last year for the Padres, 7.0 percent of opposing hitters worked a free pass off Perdomo. If he does get worse from there, how much will he fall off?

To avoid walks, a pitcher needs to accomplish two things: throw strikes (obviously), and avoid deep counts. Perdomo didn’t have much trouble with the former last year — his 64.3 percent strike rate was above the MLB average of 63.8 percent. And he excelled at the latter: 33.3 percent of those strikes were balls in play, well higher than the 28.3 percent major-league baseline. Primarily on the strength of those metrics, he has an expected walk rate of 6.1 percent, according to Alex Chamberlain’s research.

But Perdomo didn’t pull that off alone — he had some help from his backstop. Last year, Derek Norris caught 110 ¹⁄₃ of Perdomo’s 146 ²⁄₃ innings; Norris ranked 10th in the majors with 9.2 framing runs, per Baseball Prospectus. Thanks to that chicanery, Perdomo overcame mediocre command (he finished in the middle of the pack in CSAA, BP’s new measure of that skill) to tally the aforementioned strike rate. Since San Diego flipped Norris to Washington in December, we might conclude Perdomo would struggle in 2016 without his battery mate.

This season, though, the Padres catching situation should be even better. Over the last three years in the minors, Austin Hedges has compiled an astounding 44.3 framing runs. During 2015 and 2016, he accumulated 7.9 framing runs for the Padres, in less than half as many innings (412 ¹⁄₃) as Norris got in 2016 (983). Indeed, BP’s Dustin Palmateer posited that Hedges “might be the leading candidate for best defensive catcher in the majors by the end of the decade—shoot, he might be the answer by Opening Day.” With the 24-year-old wunderkind poised to take over behind the plate, Perdomo should keep cruising on along, racking up strikes and pitching efficiently.

Perdomo’s not a control wizard by any stretch of the imagination. He’ll give out walks time and again, and sometimes — as happened quite often in 2016 — those walks will come around to score. But he doesn’t have any issues finding the strike zone, and he’ll have a rock behind the plate if his command starts to fray. When it comes to bases on balls, the slight optimism of Steamer is the logical pick.

Home runs — 2.2 percent (Steamer) vs. 3.2 percent (ZiPS)

For Perdomo, this is easily the area of greatest import. Over a span of 850 batters faced — roughly 200 innings — the difference between ZiPS and Steamer amounts to about eight home runs. With a round-tripper worth, on average, about two runs, that means the projections are apart by more than a win and a half based on dingers alone. In 2016, 3.5 percent of opposing hitters clubbed a long ball off Perdomo; how far will he come down from that?

Perdomo’s defining characteristic is — and always has been — grounders. Last year, 59.0 percent of the balls put in play against him stayed on the ground; out of the 142 pitchers with at least 100 innings, only Marcus Stroman did better. Perdomo’s 2016 woes came with the ball in the air: Jon Niese was the sole pitcher in that sample with a higher home run/fly ball rate than Perdomo’s 21.7 percent. The questions here are twofold: Will Perdomo keep getting that many grounders? And will that many fly balls keep leaving the yard?

While that ground ball rate is extremely high, that doesn’t mean Perdomo can’t sustain it. As my colleague Spencer Bingol wrote in August, Perdomo started throwing his sinker a lot more when he moved to the rotation, and that pitch gets grounders like few others:

Because of the high velocity on the sinker — just 18 starters topped Perdomo’s 94.3-mph average — hitters have trouble squaring it up; because of its consistent location down in the zone, they’ll typically pound it into the ground. With this pitch in his arsenal, Perdomo should continue to pile up the grounders.

On the other hand, when hitters did manage to elevate against Perdomo, they inflicted a lot of damage. Per FanGraphs, 54.4 percent of Perdomo’s air balls were hard-hit, one of the highest marks in the majors. By Statcast, he had an average exit velocity of 93.8 mph on fly balls and line drives, which also put him among the major-league laggards. So those home runs weren’t entirely bad luck — Perdomo did get knocked around quite a bit.

Still, given his extreme ground-ball tilt, and the semi-cozy confines of Petco Park, I can’t see all these taters sticking around. While Perdomo’s home run/fly ball rate will most likely remain high, anything below 20 percent would lower his overall home run rate significantly. If we give Perdomo, say, a 15 percent HR/FB rate for 2016, his home run rate plummets to 2.3 percent, lining up with what Steamer anticipates. A nasty sinkerballer like him will usually manage to suppress power, one fluky year aside.

BABIP — .298 (Steamer) vs. .312 (ZiPS)

Four-baggers weren’t the only thing to give Perdomo trouble last year. He also gave up far too many three-, two-, and one-baggers (I think that last one is a thing). Among those 100-innings pitchers, just four had a higher BABIP than Perdomo’s .341. Because of regression toward the mean, that probably won’t repeat, but he could end up with an inflated BABIP again. Will he?

Perdomo doesn’t act on his BABIP alone. We have to take into account the gloves behind him — i.e., the ones whom he’ll trust to field those balls cleanly. At three of the four infield positions, FanGraphs projects the Padres to be subpar defensively:

The newly extended Myers should hold down the fort at first, but elsewhere on the diamond, San Diego has some defensive holes. The gruesome collection of scrubs the Friars have assembled can hardly stack up to the remaining free agents, much less convert a grounder into an out.

For Perdomo, though, things may not be as dire as they look. This infield would still be a major upgrade over last year’s — in 2016, Padres infielders were worth -35 DRS and -23.0 UZR, thanks mostly to the team’s black hole at shortstop. The incompetence behind him was the reason for Perdomo’s elevated BABIP: He had a ground ball hard-hit rate of 20.4 percent, analogous to the major-league average. With his routine grounders heading into more steady gloves, Perdomo’s BABIP should come down significantly.

In 2017, Perdomo should continue to garner weak ground balls (assuming that sinker retains its potency, of course). The fly balls might remain hard, and plenty of them will fall in for hits, but with three-fifths of his balls in play on the ground, that shouldn’t ding him too much. With Petco Park — which, according to Steve Staude’s research, is the least BABIP-friendly (to hitters) in the majors — on his side, Perdomo has a solid shot at fulfilling Steamer’s projection.

Strand rate — 69.6 percent (ZiPS) vs. 70.3 percent (Steamer)

Here, as with walk rate, Steamer and ZiPS concur for the most part. Neither expects Perdomo to leave very many runners on base (the MLB average in 2016 was 72.9 percent), and each foresees some improvement from 2016, when Perdomo stranded 67.9 percent of his baserunners. Still, the two can’t come to a firm verdict, so we’ll have to resolve this one ourselves.

As a rookie, Perdomo displayed two skills that, in theory, should have helped him strand runners. He rode his elevated ground ball rate to 20 double plays in 129 opportunities, a 15.5 percent clip that was far above the 11.0 percent major-league average. Additionally, he allowed 12 stolen-base attempts in 241 chances*, for a 5.0 percent clip that fell a little below the MLB-wide mark of 5.3 percent.

*Per Baseball-Reference: “Plate appearances through which a runner was on first or second with the next base open.”

So why did Perdomo finish in the bottom fifth of the majors in strand rate? He could turn two and hold runners in place, but when it came to pitching from the stretch, he had some serious problems:

Underneath that, though, things get a little murky. Perdomo’s strikeout rate was about the same with the bases empty (15.9 percent) and with runners on (15.8 percent); the same applied to his BABIP (.338 BE, .344 MoB). With intentional walks removed from the equation, his walk rate was actually better with men on base (5.2 percent) than with no one on (6.6 percent).

The only area in which Perdomo performed noticeably worse from the stretch was home runs. Once someone had reached base behind him, Perdomo’s home run rate nearly doubled, from 2.5 percent to 4.7 percent. As explained above, Perdomo should recover from his 2016 bout with gopheritis; after he’s put that behind him, he’ll start to leave some more runners on base.

Strand rate is a notoriously volatile metric, so Perdomo could plausibly go in either direction this year. He has all the makings of a clutch pitcher, though — he gets double plays, he prevents runners from advancing, and with the notable exception of home runs, he doesn’t get much worse when pitching from the stretch. Steamer gives the most realistic appraisal here.

So in the final analysis, all four of these areas go Steamer’s way. If we buy the projection, Perdomo will keep his walk rate respectable, stop allowing so many homers, and make progress on his BABIP and strand rate. It could make him an average pitcher for the Padres, which they’d be more than happy to have on staff.

And hey — in a rotation that’s inspired a number of jokes (and could be historically awful), he doesn’t have to do much to stand out. Even if his production declines to what ZiPS projects, he’ll still get a chance to prove himself and rebound. Either way, the Cardinals may be wishing they could bring this Rule 5 pick back.

Ryan Romano is the co-managing editor for Beyond the Box Score. He also writes about the Orioles for Camden Depot, and about politics for The Diamondback. Follow him on Twitter if you enjoy angry tweets about Maryland sports.