Mediating projections, Part III: Leonys Martin

Welcome back to Mediating Projections, the series in which I attempt to solve disputes between various projection systems (the operative word there is "attempt"). Check out Parts I and II for more hot forecasting action, as well as the inaugural post from last offseason. Today, we'll scrutinize Leonys Martin, an elite defender whose offense collapsed this year following a couple of solid campaigns. The former Ranger and current Mariner has different outlooks from different systems:

System	PA	K%	BB%	ISO	BABIP	wOBA	wRC+	Def	fWAR	fWAR/600
Steamer	529	21.0%	6.2%	.108	.295	.282	79	4.8	1.2	1.4
ZiPS	446	22.2%	5.8%	.120	.306	.301	88*	11.0	2.5	3.4

*ZiPS projects OPS+ instead of wRC+; the difference should be minimal.

Steamer thinks Martin will remain a poor hitter and lose a step in the field; ZiPS thinks his offense will bounce back and his defense will continue to elevate him. Which projection should we believe? Let's break down the differences between the two of them.

Strikeouts

Interestingly enough, ZiPS takes a more pessimistic view on Martin's plate discipline metrics, predicting worse marks than Steamer in both areas. The strikeouts stand out as the most notable resting disagreement, because of the extent — the systems differ by more than a full percentage point — and because of Martin's history. Before 2015, he'd gone down on strikes in 20.1 percent of his plate appearances; he followed that up with a 22.3 percent clip this year. Will he keep that pace up, or will he regress toward his career norm?

2015, as stated previously, went rather poorly for Martin. He paired a new-found distaste for pitches in the strike zone with a loss of contact, causing both his looking and swinging strike rates to spike:

Year(s)	Z-Swing%	Contact%	Look%	Whiff%
2011-2014	64.1%	78.6%	16.6%	12.0%
2015	60.1%	74.1%	19.2%	12.7%

That dropoff substantiates the added strikeouts. But how do we know Martin won't improve again in these regards? If he altered his approach in 2015, maybe he'll change it back, after seeing how badly things turned out.

In terms of swings, Martin clearly made a significant adjustment:

High strikes, low strikes, and everything in between — Martin took them all much more often. It wouldn't be crazy to expect him to revert to his previous strategy, which had yielded fewer called strikes (and better hitting overall).

In terms of whiffs, though, he has a considerably gloomier outlook. Here, the change came on the other side of the equation, as pitchers targeted him differently:

Before, Martin had seen 25.4 percent of his pitches arrive in the lowest fifth; that grew to 30.1 percent this season. Since he's always whiffed the most in that area — with a lifetime swinging-strike there of 20.3 percent — this means pitchers have finally caught on to his primary weakness.

Of course, he could counter-adjust, as most solid baseball players do. If he refines his vision on balls in the dirt, he would probably shave off some punchouts, perhaps even returning to his earlier levels. Without that kind of a change, though, the strikeouts will likely stay, just as ZiPS predicts.

Walks

While the gap here doesn't yawn as wide as the strikeout one does, it still looks significant. This year saw Martin's free pass rate drop to 5.2 percent, from a previous career mark of 6.2 percent; Steamer thinks it'll jump back up to the latter level, and ZiPS splits the difference. Which forecast makes the most sense?

Oddly enough, Martin didn't chase more often in 2015. His swing rate on pitches outside the strike zone actually fell, from 34.2 to 31.2 percent. Pitchers attacked him more often, upping their zone rate from 49.1 to 50.3 percent, but he still ended up taking more balls — his strike rate decreased from 67.6 percent to 67.0 percent. Yet somehow, he earned fewer bases on balls.

Situational aggression seems to bear the blame here. Of Martin's 310 plate appearances this season, 12.6 percent went to three balls, compared to 14.2 percent across the previous three seasons. And his strike rate in those scenarios didn't move very much:

Year(s)	0-Ball Strike%	1-Ball Strike%	2-Ball Strike%	3-Ball Strike%
2011-2014	63.7%	69.9%	74.1%	74.7%
2015	64.5%	68.9%	71.6%	74.1%

What should we make of all of this? With injuries draining his hitting (which we'll get to shortly), Martin appeared to press at the plate. He held onto his patience early on, but as the count worked in his favor, he threw it out the window, hoping to break out of his season-long slump. This meant he entered fewer three-ball counts and didn't take advantage of those to a greater degree, which sank his walk rate.

With that said, Martin can probably improve again when it comes to free passes. On a new team that stresses plate discipline, he'll have an environment enabling him to succeed. (And, as we'll see in a moment, his hitting should rebound, eliminating the urge to push at the plate.) All in all, the optimism of Steamer looks like the more reasonable projection.

Isolated power

Even before 2015, Martin had never distinguished himself when it came to clout, with a meager .110 ISO. That deflated further to .094 this season, and the projections don't agree on the extent to which it'll come back — ZiPS thinks he'll pound out more extra bases than Steamer does. In which system should we trust?

There's first the matter of ballpark, which doesn't help Martin's case. According to research from my colleague John Choiniere, Safeco Field has repressed left-handed power by six to eight percent in recent seasons. Playing half his games there, as opposed to the narrow dimensions of Globe Life Park, will limit Martin's ability to hit for power. The stadium matters more when we take into account Martin's power profile — he doesn't hit home runs, but rather smacks doubles and triples, which Seattle crushes for lefties.

From there, we can note the major differences between the projections when it comes to power. Steamer actually prognosticates a higher double rate (4.2 percent) than ZiPS does (3.8 percent). The latter, by contrast, thinks he'll hit much more three- and four-baggers — at clips of 0.9 and 1.8 percent, respectively — whereas the former anticipates a triple rate of 0.4 percent and a home run rate of 1.5 percent.

The aforementioned park factor — Seattle depressed left-handed triples more than any other AL ballpark last season — as well as the innate flukiness of triples, means Martin probably won't make it to third base that often. As for long balls, consider this: He's hit one in 1.4 percent of his lifetime chances, roughly half of which have come in one of the most long ball-friendly parks in the league. The odds of him achieving a new career high in his new setting don't seem very high to me.

Even when Martin held his own at the plate, he didn't hit for much power. His resurrection, if it happens, won't come because of clout; with his stadium inhibiting him, his ISO will likely live up to Steamer's prediction.

Batting average on balls in play

Martin's biggest decline at the plate came from his BABIP, which fell from .325 across the prior four years to .270 in this one. Although both Steamer and ZiPS think he won't reach that level again, the latter expects him to come closer to it than the former does. Will his hits make a comeback?

As with power, the ballpark will work against Martin. Steve Staude's research shows, unsurprisingly, that Safeco Field harms BABIP, to a notably greater degree than Globe Life Park does. That explains why both projections foresee a deterioration from his Rangers mark as a Mariner. But the real question concerns his 2015 season: What went wrong during it to cause his BABIP to dive?

Part of the problem came from the fact that he simply didn't make solid contact frequently enough. His hard-hit rate fell from 25.7 percent across the preceding four years to 23.3 percent in this one, while his soft-hit rate increased from 20.1 percent to 23.8 percent. This likely arose from his lingering hand injuries, which should convalesce given a full offseason.

Let's not stop there, though. Martin's batted-ball profile also shows a significant change. His ground ball rate didn't budge in 2015, but his other metrics did: He hit 15.3 percent line drives and 33.0 percent fly balls, compared to respective rates of 21.7 and 27.9 percent prior. Over the course of his career, Martin's compiled a .721 line drive BABIP, ranking 67th in baseball out of 418 qualifiers. Compare that to his .075 fly ball BABIP — which places 373rd among 379 of his peers — and you'll understand why he needs liners to stay afloat.

Line drives and fly balls are always tricky to separate. Scoring bias can impact them, and they can vacillate a lot from year to year. For Martin to regain his (relative) offensive prowess, he'll need to keep his batted-ball angle low, a task that may prove diffcult — especially in Seattle, which helps flies and hurts liners. While a clean bill of health could clearly aid that cause, it might not be enough.

In the end, this one could go either way. Because Martin will play under a venerable former hitter in 2016, I'll choose ZiPS, but realistically he could maintain his subpar 2015 production. The difference between the two doesn't stand out that much, and whatever happens could have as much to do with luck (will Martin remain healthy, and will his air balls stay low?) as it will with skill.

Defense

By a tremendous margin, this constitutes the largest difference between the two projection. On a per-600 plate appearance basis, they diverge by 9.4 runs — about the value of a full win in 2015. Both think he'll play in center field, so positional adjustments aren't the question. ZiPS just happens to love Martin's glove, whereas Steamer eyes it more skeptically. Which will win out?

For his career, Martin has saved 26.4 runs, according to UZR, giving him a career Def/600 of 12.4. That comes close to the level ZiPS predicts, but it doesn't necessarily mean he'll fulfill that. With defensive numbers, we need to regress heavily, to account for the sample issues and general inconsistencies that plague the metrics. And with Martin, that regression doesn't leave him looking too appealing.

Martin's never derived many runs from range or error avoidance — in fact, UZR has him as a slight negative in both of those areas. His incredible arm sets him apart, by netting him an astounding 28.6 runs above average. To get a better idea of where his defense will go from here, we need to focus in on this facet of it. For that, we'll turn to Jeff Sullivan, who analyzed the similar case of Juan Lagares two offseasons ago.

Sullivan discovered something interesting, and relevant to Martin's future: Arms don't correlate as well from year to year as range does. Players who show off their godly cannons in one campaign can find themselves looking more mortal in the next one. Just look at Lagares — in just two seasons, he went from 13.8 runs above average to 1.7 runs below average. Martin has posted high arm ratings in three consecutive years, so he has a broader resume to work with, but the principle remains the same. Even with the strongest of arms, we should assume the player will get worse.

Martin will turn 28 in March, so he still has some youth left. Nevertheless, his one-dimensional defense probably won't progress any further from here. Atop the majors when it comes to gunning down runners over the past three years, he has nowhere to go but down. That isn't to say he'll become a negative in the field — just that he'll likely play at a more pedestrian level, to the tune of Steamer.

Conclusion

Steamer wins out in three of the five debates, and ZiPS takes the crown in the other two. Martin, in other words, should be an average player who falls short of the stardom he attained in prior years. While he still possesses the upside to reach another echelon, his plate discipline woes, lack of power, and potential defensive issues will stand in his way. Of course, we don't know what 2016 will bring — maybe Martin will break out and help the Mariners finally realize their potential. Projections can only go so far.

. . .

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) and about the Brewers on BP Milwaukee.