It’s dangerous to believe anything players say in spring training, where mechanical tweaks are common, batters daydream about bunting to beat the shift, and pitchers fiddle with their arsenals. Some spring experiments stick. Others end before Opening Day. Amid the usual empty predictions this spring, though, Masahiro Tanaka’s sounded more credible than most, if only because unlike the others, it didn’t promise improvement. Tanaka, who started and lost the Yankees’ home opener against Toronto on Monday, had told reporters days earlier that he intended to throw more sinkers and fewer four-seamers this season, and that the four-seamers he did throw would be slower than their predecessors from last season. It’s not that he couldn’t throw pitches at the same old speed, Tanaka clarified; rather, it’s that, like Bartleby the big leaguer, he would prefer not to.

Under normal circumstances, few pitchers with above-average velocity would willingly give batters more time to react. Professional pitchers have every reason to throw as hard as humanly possible, and the incentives are simple: Hard-throwing pitchers tend to miss more bats, allow fewer runs, and get GIF’ed more often. Teams (other than the Twins) recognize this relationship and give preference to guys with good fastballs. Pitchers, in turn, try to throw hard in order to make themselves more attractive to teams. It’s a process that weeds out all but a few special soft-tossers, causing pitch speeds and strikeout rates to climb over time.

Under different conditions, throwing at less than maximum effort was a baseball best practice. “Some pitchers will put all that they have on each ball,” Hall of Famer Christy Mathewson (or ghostwriter John Wheeler) wrote in his 1912 book, Pitching in a Pinch. “This is foolish for two reasons. In the first place, it exhausts the man physically and, when the pinch comes, he has not the strength to last it out. But second and more important, it shows the batters everything that he has, which is senseless. A man should always hold something in reserve, a surprise to spring when things get tight.”

To some extent, pitchers still work the way Mathewson did. With runners on base, or when facing an above-average batter, pitchers throw slightly harder. On 3-0 counts, or when facing opposing pitchers, they throw slightly slower. But today’s major league pitchers never really relax on the mound. Starting pitchers vary their velocity from fastball to fastball more than relievers, whose outings rarely last long enough for them to tire, but even starters pace themselves less than they did in Mathewson’s day. Modern starters have less reason to save something for a pinch, because bullpens are bigger than ever. By the time the pinch appears, it’s usually a relief pitcher’s problem.

Tanaka isn’t trying to bring back the dead ball era, though. He’s channeling Mathewson just to survive the season. The 26-year-old partially tore his ulnar collateral ligament — the fragile link in the elbow that, when severed, often leads to Tommy John surgery — last season, and he’s spent the last nine months trying to prevent “partial” from becoming “complete.”

“Partial tear” is a nebulous term that can describe anything from light fraying to a state just short of snapping. “Some partial tears need surgery no matter what,” says Glenn Fleisig, research director at the American Sports Medicine Institute. “Some partial tears it would be ridiculous to have surgery — they could easily get better from rehab. But then there’s a lot in the middle, which is, the doctor and the athlete and the trainers make a decision to give [rehab] a try.”

Tanaka is one of those middle men, the athletic equivalent of a patient diagnosed with a terminal illness who outlives his initial prognosis. We can’t call him cured, but by lasting this long, he’s making us question the dire predictions. Lingering indefinitely in an “every day off the DL is a gift” phase, Tanaka has different priorities than the typical pitcher. He’s a star with financial security, in the second season of a seven-year, $155 million deal (with an opt-out after 2017). He doesn’t have to prove he has major league talent, and he doesn’t have a competitor pushing him: Even a moderately diminished Tanaka would be better than the Yankees’ best possible replacement. The wisdom of continuing to play him, then, depends on the answers to two questions: Will the switch to sinkers keep Tanaka intact? And how effective can he be without the four-seamer?

“Biomechanically, it makes sense,” Fleisig says, referring to Tanaka’s reduced-effort delivery. When ASMI conducted a study on rehabbing pitchers who were told to throw at 75 percent effort, they found that those pitchers put “85 percent as much stress on their arm as a full-effort pitch.” ASMI doesn’t have data on arm stress at 90 percent effort, but Fleisig estimates that it would be “between 90 and 100 percent as much stress on your joint as 100 percent throwing.” Theoretically, that could be enough to keep Tanaka trucking until the offseason, when he could either continue to heal or elect to have surgery at a time when he would almost certainly miss only one season, returning at full strength in 2017.

“If you have a partially torn ligament or tendon, throwing at reduced effort won’t help some people who are too far gone,” Fleisig explains. “But a guy who’s borderline, it’s a reasonable approach, because what happens is your ligament and tendons are full of … countless fibers, and when you throw at reduced effort, it can reduce how many of these fibers get torn.”

ASMI doesn’t have data on two-seamer and four-seamer stresses, specifically, but lower velocities generally mean less stress and a lower risk of a ligament tear. From a medical perspective, Tanaka’s efforts aren’t futile.

Even better: From a performance perspective, Tanaka is about the best possible candidate for four-seamer removal, a procedure that would be fatal for most pitchers. Pitch-type run values — stats that tell us how much damage was done against a given offering over the course of a season — are overly simplistic, in that they treat each pitch as an isolated event instead of as part of a sequence that sets up a subsequent pitch or capitalizes on a previous one. But Tanaka’s four-seamer, in isolation, was a below-average pitch in 2014, getting fewer whiffs and grounders than the typical four-seamer. Conversely, Tanaka’s splitter, his signature pitch, was almost two standard deviations above the mean in both whiff rate and ground ball rate. His four-seamers and sinkers were mostly placeholders that primed batters for his better pitches.

Sort the list of pitchers with at least 100 innings thrown last season by four-seamer run value per 100 pitches, and Tanaka ranks seventh from the bottom. He’s the only one of the bottom 15 pitchers to have posted a league-average park-adjusted ERA. That’s not surprising: Most pitchers work off their fastballs, so a fastball that gets hit hard is often the sign of a pitcher who gets hit hard. Tanaka was the exception, an ace who made his mark with off-speed stuff. Last month, I wrote about the ways in which some starters mix their pitches to minimize their vulnerability to repeated exposure to the same batters as games go on. Tanaka was the only full-time starter last season who didn’t throw any one pitch type more than 26 percent of the time, making his pitch assortment the most evenly distributed in the 104-pitcher sample. And in theory, a pitcher who relies less on his four-seamer than most starters should miss it less than the average pitcher would if it went away.

Opening Day gave us a real game to go on, and Tanaka was true to his word. On their first pitch of any given game, starting pitchers tend to throw fastballs at “an exceedingly high rate,” in part because leadoff batters tend to take the first pitch of a game more often than usual. The first pitch of Tanaka’s major league career was a four-seam fastball: typical. The batter, Melky Cabrera, took that pitch for a called strike: also typical. From the very first pitch of Tanaka’s start on Monday, though — a slider for a called strike, followed by back-to-back splitters that resulted in a Jose Reyes swinging strikeout — Tanaka was atypical.

[mlbvideo id=”62424983″ width=”510″ height=”286″ /]

The righty threw 82 pitches over four innings, and only five of them (6.1 percent) were four-seam fastballs, according to classifications provided by Pitch Info. That’s a lower four-seamer percentage than Tanaka recorded in any start last season. He replaced the missing four-seamers with sliders, splitters, and sinkers. The table below shows the usage percentage of each of his four most common pitch types on Monday, with the usage rank among his 21 career big league starts in parentheses, and the table after shows Tanaka’s game-by-game pitch selection.

Tanaka’s five four-seamers averaged 92.6 miles per hour — lower than his April 2014 average of 93.5, but higher than his average in any subsequent month last season. That’s encouraging, although there was no sign of his top-shelf stuff. In the first three months of last season, Tanaka maxed out at close to 97 mph, but his fastest delivery on Monday came in at 94.6. Without as much leeway to make mistakes, Tanaka will have to keep the ball down, which he did a decent job of on Opening Day:

Tanaka threw 66 percent of his pitches in the bottom third of the strike zone or below, an improvement over his 53 percent rate from last season, which itself ranked ninth-highest out of 139 pitchers with at least 100 innings pitched. Three of the five runs (four earned) that Toronto scored on Tanaka came on ground balls, one of which deflected off Chase Headley’s glove, and another of which was a bunt that Headley threw away. The fourth, though, came on a 91 mph sinker to Edwin Encarnacion, a pitch that Tanaka located only a little better than the fan who tried to throw the ball back from the bleachers:

[mlbvideo id=”63291883″ width=”510″ height=”286″ /]

Before the homer, Tanaka had thrown 20 fastballs — four-seamer or sinker — in 49 pitches (40.8 percent). After the homer, he threw six out of 33 (18.1 percent). That pitch selection stinks of fear. Still, Tanaka got whiffs with his slider, and batters missed his splitter on 60 percent of their swings. Even without as much separation in speed between his hard and soft stuff, Tanaka struck out six batters in an imposing Toronto lineup.

[mlbvideo id=”62527483″ width=”510″ height=”286″ /]

Although on the surface the box score seems like a confirmation of Yankees fans’ worst fears, it’s possible to put a positive spin on Tanaka’s inaugural outing in 2015. Aside from his voluntary renouncement of the four-seamer, Tanaka’s stuff seemed almost the same as it was before we were worried about his elbow. He stills throws hard enough to get swings and misses with his off-speed stuff. And while Tanaka doesn’t have a great sinker, his forsaken four-seamer was worse. It’s probably too much to hope for addition by subtraction, but if Tanaka’s partially torn ligament doesn’t hurt his command,¹ the new Tanaka should be a useful starter (if not an ace) for as long as his ligament lasts, even if he nibbles too much to go deep into games. Tanaka has already defied one of our fears — and his own — by avoiding surgery and surviving the spring UCL reaping so far. Now he has a chance to chart the non-fastballs frontier.

♦♦♦

Most pitchers who lose velocity in the spring don’t do so in a conscious attempt to take it easy, but some unintended velocity trends are as telling as Tanaka’s. Small-sample pitch speeds aren’t as easily dismissed as some other spring training stats: Last spring, reliever Hector Rondon’s four-seamer velocity showed a gain of 1.6 mph compared to his average from the 2013 regular season; Rondon would enjoy an additional uptick after Opening Day, which contributed to a successful season as the Cubs closer. Budding Angels ace Garrett Richards also added speed in the spring. Meanwhile, future injury cases Martin Perez and Bronson Arroyo showed significant velocity slippage.

It’s easy to cherry-pick examples that seem useful in retrospect, but more thorough research has revealed a moderate correlation between velocity changes in spring training and subsequent readings during the regular season. This year, we have spring velocity data from Peoria, Surprise, and Salt River, as well as from some MLB stadiums in exhibition games. The embedded table below shows the difference between 2014 regular-season velocity and 2015 spring velocity for almost 300 pitchers for whom we have four-seam fastball velocities from both periods. Caveats: The readings aren’t corrected for temperature or calibrated by ballpark, so the numbers aren’t precise down to the decimal point, although those adjustments are usually small.

On the positive side, the Indians’ Zach McAllister, whose stats stood out in March, also showed one of the month’s largest velocity gains. The Padres’ Andrew Cashner, a top-10 starting pitcher by four-seamer velocity last season, threw almost 2 mph harder this spring. Trevor Cahill’s velocity picked up before the Braves stole him away from the Diamondbacks, and Dodgers addition Brandon McCarthy, whose velo was way up last year, might have even greater gains to make.

On the other end of the scale, Angels starter Jered Weaver and Matt Shoemaker showed large declines, although Weaver’s fastball bounced back to the high 80s in his Opening Day loss. Tommy John victims Yu Darvish and Tim Collins posted slower speeds before they were shut down, and the Royals’ Greg Holland–Wade Davis–Kelvin Herrera trio had trouble replicating their intimidating 2014 fastball figures. Dellin Betances was down (as he suspected), as was Matt Cain, who’s suffering from forearm tightness. Keep in mind that velocities tend to rise until August; spring speeds, intriguing as they are for arms we haven’t yet seen in the regular season, are only early looks that don’t always predict regular-season results.

Thanks to Harry Pavlidis for research assistance.

1. It’s too soon to tell, and ASMI doesn’t have any answers, although somewhat worrisomely only Francisco Liriano has a lower zone rate since Tanaka’s September comeback (minimum 10 IP).