The MVP Machine

by Ben Lindbergh

  According to a 2018 report by the Aspen Institute, only 34 percent of children from families earning less than $25,000 played a team sport in 2017, compared to 69 percent from homes earning more than $100,000. Between 2011 and 2017, the participation rates in households earning at least $75,000 increased, while the participation rates in lower-earning households decreased. Organized sports are often an unsupportable expense.

  Baseball, wrote star outfielder Andrew McCutchen in a 2015 piece for the Players’ Tribune, “used to be a way out for poor kids. Now it’s a sport that increasingly freezes out kids whose parents don’t have the income to finance the travel baseball circuit.” Deliberate practice is powerful regardless of one’s economic origins, but it requires time, technology, and instruction, all of which may be in short supply for children from lower-income families. Similarly, tools like TrackMan help amateur talents get spotted and drafted, but only if they can afford to play on teams and in tournaments where those systems are installed. In the post-Moneyball era, the have-nots are only falling further behind the curve.

  In late 2016, MLB and USA Baseball began to combat the pay-to-play structure of youth baseball by forming the Prospect Development Pipeline, a series of free, team-attended, invitation-only events for high-school players. In late 2018, MLB expanded the program by forming the PDP League, a Royals Academy-esque “development and showcase” experience scheduled for mid-June through early July 2019. The announcement on MLB.com promised that the eighty participants would receive “evaluation using modern technologies to tailor an individualized development curriculum for each player.”

  MLB and USA Baseball also teamed up in 2018 to launch the Trainer Partnership Program, which in part will bring “all of the state-of-the-art evaluation tools that have become so common at the major-league level to the international market.” In November, the program held a three-day showcase in Boca Chica that featured 120 Dominican players who’ll be eligible to sign in 2019 and 2020. Every pitch, batted ball, and run time was meticulously tracked.

  These initiatives are addressing real problems for young players, but MLB isn’t funding them out of purely altruistic motives. Both programs are rectifying a problem for teams by feeding decision-makers’ insatiable desire for data. On the international market, teams can verbally commit to players up to three years before they’re eligible to sign, but they can’t bring those players to their academies until one year before. That leaves a window during which teams are dependent on trainers for data, and the trainers have taken advantage. “It’s getting wild out there,” Bannister says. “The earlier you can get data on [a player, the better], but you can’t bring him into your academy. So now it’s all about setting stuff up in a portable format, or now buying it from the [trainers].” Or luring players to a free showcase and letting the cameras roll.

  For good or ill, the youth baseball experience worldwide is about to be more information-rich than ever. If anything, though, objective measurements provide fuel for competitiveness and a new way to hook kids on an old game. The same people who say stats suck the joy out of following baseball may also say that launch angles and spin rates suck the joy out of playing it, but just as there’s nothing joyless about understanding the sport in a different and deeper way, there’s nothing joyless about being better at it.

  Eventually, early exposure to data may help restore a species that Moneyball made almost extinct: the MLB player turned general manager.

  According to data provided by Baseball Prospectus writer Dustin Palmateer, 44.1 percent of GMs hired in the 1980s were former MLB players. Among GMs hired in the 2010s, only two, Dave Stewart and Jerry Dipoto, have been big leaguers. The percentage of new GMs who were once minor leaguers has fallen from 67.6 percent in the ’80s to 20.6 percent in the 2010s. Almost 40 percent of GMs hired in the 2010s have been Ivy Leaguers; from the ’70s to the ’90s, that rate never rose above 3 percent. Although it’s a sign of progress that nonplayers are no longer excluded from the team-running ranks, front offices have swung so far in the other direction that they’ve merely traded one type of homogeneity for another, morphing into a slightly younger and far nerdier brand of old boys’ club. GMs are exclusively male, overwhelmingly white, and increasingly Ivy League educated. That lack of demographic diversity is likely leading to a lack of diversity of thought.

  Dipoto says that in contrast to a previous generation of players that was left behind by baseball’s statistical revolution, today’s players are “learning as they go, and as a result, I think you’re going to see some players start to matriculate back toward front-office or player-acquisition-type roles like they did twenty-five, thirty-five years ago.” Rays conduit Cole Figueroa adds, “I truly believe there is a harmonious middle ground where one day we will see players running teams again, and even more progressive-thinking people who didn’t play at all [sitting] in the dugout.” In December 2018, Figueroa’s team took a step along the latter path by installing its director of analytics, Jonathan Erlichman, as the majors’ first “analytics coach.” Erlichman, who never played the game beyond T-ball and majored in math at Princeton (where he wrote his senior thesis on “Gravitational Redshifts in Galaxy Clusters”), will be in uniform as a member of manager Kevin Cash’s staff in 2019, a stathead in coach’s clothing. “People are just excited for what we’re potentially going to learn from this,” Erlichman told the Tampa Bay Times.

  In 2015, Bill James said, “My view on the world is we have an ocean of ignorance and a small island of knowledge.” He wasn’t speaking specifically about baseball, but the observation applies, even to one of the most obsessively chronicled and comprehensively quantified human hobbies. “I feel like, yeah, our island is one hundred times bigger than it used to be,” Fast says. “But does that make the ocean that much smaller?”

  It’s scary not to know things. But it’s also exciting, because it means there’s more to learn.

  Inside the self-contained bubble of baseball, being better can accomplish only so much. Expansion aside, there’s a finite number of roster spots in the majors and a finite number of wins to be distributed among the thirty (or, in the future, thirty-two or thirty-four) franchises. No matter how good teams get at developing players, the average team will finish at .500, the average player will be worth no more than before, and the team that wins the World Series will still start out 0–0 the next spring.

  Outside of sports, though, there are no such constraints. We can be better without making something else in the world worse. Each of us individually, and all of us collectively, could be Justin Turner before changing his swing, Rich Hill before fully embracing his curveball, Trevor Bauer before redesigning his slider. Maybe humanity is about to break out. As James said, “We haven’t done anything yet to compare with potentially what we could do.”

  EPILOGUE

  THE RESIDUE OF DESIGN

  Success breeds complacency. Complacency breeds failure. Only the paranoid survive.

  —ANDY GROVE, former Intel CEO and semiconductor pioneer

  Brian Bannister’s father spent fifteen years in the majors and pitched in one playoff game. That was one more than Bannister himself experienced. Neither of them ever went to the World Series. So when the 2018 Red Sox made the ALCS against the Astros and then the World Series against the Dodgers, following back-to-back division series losses in the two preceding seasons and a playoff no-show the year before that, Bannister thought the same thing as the rest of the Red Sox: we can’t waste this opportunity.

  More so than any other major sport, modern baseball becomes a different beast in the playoffs, morphing from an everyday activity in which teams and players pace themselves to one in which they ratchet up their intensity in response to extra off days and higher stakes. Managers pull their starters early, furiously cycle through a parade of relievers (some of them regular-season starters), and ask closers to go longer and dig deeper than they do during most of the year. And as teams deploy their best pitchers, pitchers deploy their best pitche
s, throwing close to their maximum speeds at all times. The October mantra, Bannister says, is: “Best pitch quality, best pitch execution, best pitch mix.”

  All of the above is particularly true in the World Series, a best-of-seven sprint in which there’s no need to hold anything in reserve. “That’s when all the seeds that have been planted, in that year, in previous years on the development side, [come] together,” Bannister says. “This is no longer a development scenario. This is pure execution.”

  In a sense, though, the ongoing revolution is blurring the line between development and execution. Although big projects such as swing overhauls and new pitches are typically reserved for the off-season or spring training, today’s cutting-edge coaches and players are constantly tweaking and tuning in a more focused way than their predecessors. Bannister’s portable pitching lab and Fenway’s suite of sensors allowed the Red Sox to scout themselves, scouring their staff for any noncompliant pitches that could give a game away. “Maintaining breaking balls is a full-time job now,” Bannister says. “[A breaking ball] will waver between a 60 and a 70 pitch, just based on 15 degrees of spin axis. And that can be grip, that can be fatigue as the year goes on, it can just be getting out of whack slightly.”

  Entering the playoffs, the Red Sox faced this problem with right-handed relievers Joe Kelly, thirty, and Heath Hembree, twenty-nine, two key cogs in a bullpen that projected to be Boston’s biggest vulnerability. Both Kelly and Hembree have sliders that rate as their best secondary pitches when they’re working well. In the 2016 playoffs, Bannister says, Kelly’s slider had been “as close to throwing an 80 as you could throw,” coming in at up to 93.5 mph with a wicked shape. In his three postseason outings that year, Kelly threw the slider 41 percent of the time, more often than any other pitch. At the start of 2018, Kelly’s slider was still formidable, as was Hembree’s, whose movement Bannister describes as “perfect.”

  Over the course of the regular season, though, both Kelly and Hembree had slowly lost their nasty sliders. Relievers who throw a lot of sliders, Bannister says, tend to release them with their hands slightly turned to the side, as if they’re throwing a football. That hand position imparts extra spin to the slider compared to holding the hand more square to home plate, which adds extra hop to the fastball. When the season started, Kelly and Hembree had their hands at an angle and threw spin-efficient sliders with less hop on their heaters, but they both began to suffer from the same affliction. “As the year went on, for whatever reason, their hand[s] [were] getting more and more square on their fastball[s], and they were getting more and more carry on their four-seamer[s], and we started to see them lose their sliders,” Bannister says. The wipeout pitches they’d been throwing in April had devolved into “baby cutters” that weren’t missing many bats.

  There was one happy by-product of this unintended alteration: pitchers who throw with their hands square to home also tend to have better curveballs. And so, Bannister says, “We actually saw Joe and Heath’s curveballs getting better while their sliders were getting worse.”

  Thanks to sensitive tools like Edgertronic and KinaTrax, the Red Sox learned in 2018 that there’s value to placing pitchers into buckets based on their arm actions. Two pitchers with identical release points, as measured by TrackMan, can reach those release points via very different mechanisms, and that sequencing of actions—the relationship between shoulder, elbow, and hand—can be used to group pitchers together. “You can only throw a ball so many ways,” says Bannister, who uses golf analogies to describe the angle at which pitchers’ hands come through the ball: hook, draw, square, fade, and slice. Classifying pitchers, he says, “allows us to be better coaches because we can address their needs very specifically based on those specific angles and ratios.”

  The Red Sox discovered that once a pitcher’s arm action starts drifting in a certain direction, it’s difficult to reverse the trend, just as it is to change a golf swing overnight. Kelly and Hembree had morphed from “draw” pitchers to “square” pitchers. At first, the team had tried to fight it, but accepting it soon seemed the smarter course. “Once we identified, ‘Hey, they’re switching what bucket they’re in’… we knew we needed to address their needs differently and come up with a different game plan,” Bannister says.

  With the playoffs approaching and both pitchers unable to trust their sliders, Red Sox pitching coach Dana LeVangie, Bannister says, made a bold decision: “Let’s just completely eliminate the slider, and let’s go to [the] curveball.” Although both pitchers’ ideal sliders were better than their ideal curves, they couldn’t execute the sliders, and they could execute the curves.

  Kelly threw his last slider of the season on September 19. Hembree threw his on September 29. The chart shows how Kelly (solid lines) and Hembree (dotted lines) saw their slider whiff rates sink month by month through September. It also reveals how their usage of sliders gradually gave way to curves, culminating in a slider-less October.

  The strategy worked wonders. Hembree allowed zero runs in his four playoff appearances, including a routine 11th inning in an epic World Series Game 3. The Red Sox reliever with the next-lowest ERA was Kelly, who led the bullpen with 11 1/3 innings pitched over nine games and allowed only one earned run, striking out 13 without issuing a single free pass—an extraordinary feat for a pitcher who’s normally not stingy with walks. Despite his pedestrian regular-season stats, he signed a three-year, $25-million deal with the Dodgers in December.

  Boston’s technology had failed to prevent Kelly and Hembree’s sliders from degrading. But by helping the Sox understand what went wrong—and, in the case of their curveballs, what went right—the data suggested a solution. “If we didn’t have the tools, and we didn’t have the knowledge of the biomechanics and the physics of what was actually happening, there was no way we would have ever come to that conclusion,” Bannister says.

  Bannister reached back to his 2015 playbook to boost Boston’s postseason staff in other ways. Another right-handed reliever, Matt Barnes, was already reliant on curves, throwing them 39.1 percent of the time during the regular season, which ranked eighth among pitchers with at least forty innings—one spot behind Rich Hill. But on one crucial occasion, Bannister convinced him to drastically ramp up that rate.

  “Every once in a while [Barnes] goes, ‘What do you got for me?’” Bannister says. One of those times came on October 14, in the pitching lab in right field before ALCS Game 2 against Houston, with the Sox trailing the Astros 1–0 in the series. “I just said, ‘[Lance] McCullers and [Ryan] Pressly are going out there and, for lack of a better term, McCullers-ing us to death with breaking balls,’” Bannister recalls. He suggested to Barnes, “Why don’t you just go out there and just mirror them and do it right back?”

  That night, Barnes relieved David Price with two outs in the fifth inning, runners on first and second, and Boston up by one run. Facing Marwin González, he threw four consecutive curves to end the threat: called strike, swinging strike, foul, swinging strike three. The next inning, he threw ten curves in eleven pitches, going groundout, pop fly, groundout. He’d thrown 93 percent curves, easily exceeding his career high in an outing of more than two pitches. “That just said [to the Astros], ‘Hey, if you’re going to do that, we have guys that can do that too,’” Bannister says.

  Not only was that appearance pivotal—Barnes got the win in Boston’s series-tying 7–5 victory—but it may have had a psychological impact on the rest of the ALCS. In an era of rigorous, data-driven scouting, one outlier outing in which a player strays from his history sows doubt. “Once you’ve established that you’ll do something that’s so many standard deviations away from what you’d normally do, I think it actually makes you better for the rest of that series and the rest of the playoff run,” Bannister says, adding, “All the advance reports go out the window.”

  Speaking of pitch selection and psychological impact: two innings after Barnes left the game, Rick Porcello entered it. Porcello, who was pitching out of
the pen between starts, threw his sinker more often than any other pitch during the regular season. Statistically speaking, it wasn’t one of his better pitches—his four-seamer and slider were better—but it helped him induce contact and create quicker at-bats, allowing him to stay in games longer. In a playoff relief role, though, he wouldn’t need to be economical; he’d only need to miss bats. “Three best pitches, three best locations, and just live or die by that,” Bannister says. “You’re not a sinkerballer anymore, you’re Rick Porcello, the eighth-inning setup guy.”

  His first playoff appearance of 2018, a short but sweet hold in the eighth inning of ALDS Game 1, marked the first time in his ten-year career that he’d ever appeared in an MLB game without throwing a sinker. ALCS Game 2 was the second. Porcello got Tony Kemp to ground out with a curve, struck out González with a four-seamer, and then ended the inning by convincing Carlos Correa to chase and swing through a slider down and away. As he walked off the mound, he screamed and pumped his arms.

  “I thought that was a defining moment of the whole run,” Bannister says. Before then, “the narrative had always been we were one reliever short, or the bullpen was the vulnerability of our team, and seeing Rick step up and be that stopgap… was just such a boost for the entire team. At that moment, we were like, ‘We’re going to do this.’”

  When they did do it, in Game 5, it was behind three homers off of Clayton Kershaw (including bombs by J.D. Martinez and impending MVP Mookie Betts) and a third consecutive strong start by David Price, who had entered the month with an ugly October track record. Price had only added to his choker reputation in his shaky first two postseason starts of 2018, allowing two homers on his cutter without recording any Ks, which prompted a recommendation.