Following up on my previous post, Who Is the Shane Battier of Your Faculty?: Wall Street Journal Book Review: Paul Dickson, Lone Star Turnaround (reviewing Ben Reiter, Astroball: The New Way to Win It All (2018)):
Mr. Reiter now has written a full account of the remarkable story of how one of the greatest turnarounds in modern baseball history was engineered. ... Houston had looked at the processes that Oakland A’s general manager Billy Beane had used early in the 21st century. That team’s methods—sophisticated statistical analyses and attention to “undervalued” measuring sticks (like on-base percentage)—were detailed in Michael Lewis’s Moneyball (2003), and they changed the way baseball front offices operated. But Mr. Lewis’s book also portrayed a somewhat fraught internal organization, with old-fashioned scouts in one corner and the analytic nerds in the other, often disagreeing about players and prospects and resenting one another as well.
Astros general manager Jeff Luhnow wanted to figure out how to get scouting and analytics to work together and eventually produce an internal metric that would render a decision on a player as simple as the one in blackjack: hit or stay, keep or trade, play or bench. The blackjack analogy is apt, since Mr. Luhnow’s leading partner in all of this was Sig Mejdal, a former blackjack dealer and NASA scientist who became the head of the Astros’ Nerd Cave or, as the Astros named him, “director of decision sciences.”
Under Mr. Luhnow, scouts not only made subjective judgments about a prospect’s talent but also collected unique data that they fed to the folks in the Nerd Cave. And the nerds began listening to the scouts. All of this was easier said than done, but it was done, and the team made a series of sound, even brilliant, choices as it drafted, traded and signed players. ...
[R]oster-creation, all by itself, did not bring home the championship. Building an exceptional team is one thing, but making it work as a team is another. “Fault lines” exist in all complex organizations—including baseball teams. If these lines can be bridged or eradicated, a team is likely to win more ball games. To use another bit of old-fashioned terminology, a team needs chemistry.
Carlos Beltrán, the veteran outfielder signed by the Astros after the 2016 season, immediately took on the role of chief chemist. Among other things, he created a postgame ceremony that awarded prizes for excellence in the field and instituted a postgame “court” for those who failed to attend: The fine was $500. Mr. Beltrán also had a singular ability to study opposing pitchers and determine their “tells”—gestures and small changes in behavior that signaled whether or not the next pitch would be, for example, a breaking ball or a fast ball. Finally, Mr. Beltrán had a strong desire to close the gap between the English and Spanish speakers. ...
Mr. Reiter’s superb narrative of how the team got there provides powerful insights into how organizations—not just baseball clubs—work best.
Sports Illustrated, Why Carlos Beltrán Was the Perfect Addition to Aid the Astros' Journey to the World Series:
Luhnow also felt that Beltrán could imbue a club with something else, a variable that neither Statcast nor any of Sig’s other metrics could begin to track.
In 2015 the Astros became a contender sooner than expected, finishing the regular season 86–76, a 35-win improvement in the span of just two years. But in the ALDS the Astros fell to an opponent that was not only finally good, but cohesive and relentless in a way that was hard to explain. The opponent was the Royals, Beltrán’s old team, who were on their way to winning their first World Series in three decades.
The next season the Astros regressed. They won two fewer games than in 2015, and missed a wild-card spot. Maybe, after the outlier that was ’15, they simply returned to the normalized improvement curve Luhnow and Sig had once imagined when they began rebuilding the club in ’12. But perhaps, they speculated, their clubhouse was missing something crucial in both seasons. They had turned the Astros around by ridding the roster of expensive veterans and focusing on acquiring the right young players. But that meant they had almost no veterans. “We had some veterans in there, but they weren’t necessarily the types of guys that create followership,” Luhnow said. He reflected on the ’15 ALDS and the intangible dynamic the experienced Royals had, one that his precocious Astros lacked.
Perhaps the club was missing a player who could not only hit home runs but who had experienced virtually everything a player could in professional baseball, one who knew what it was like to be very young and very old, to make the league’s minimum salary and more than almost anyone, to make All-Star games, to win playoff series and to lose them, to play like Superman and to be too injured to play at all for great swaths of a few seasons. A player who had faced 1,498 pitchers in his career and shared a dugout with nearly 700 teammates.
Perhaps it was missing Carlos Beltrán.
The analytics community’s view of the impact of a team’s chemistry on its bottom line performance had evolved from the days in which the conventional wisdom held that it was so squishy that it probably didn’t matter much. “Whether you sell insurance or you’re a schoolteacher, obviously the people you work with can make you more productive or less productive,” the sabermetrics pioneer Bill James toldThe Seattle Timesin 2010, eight years after he’d been hired as a Red Sox adviser. “Baseball would be quite a remarkable activity if it was the one place in the world where your coworkers didn’t have any impact on how productive you were.”
James had less patience than ever for the argument that while chemistry might exist, it was so difficult for anyone to measure that it was best to ignore it. “If you divide the world into shit that you know and shit that you don’t know, and you study the stuff that you know, then you’re not going to learn very much,” he said.
Luhnow and Sig had spent their careers in baseball trying to devise ways to turn the shit they didn’t know into shit they did. By the winter before the 2017 season the value of team chemistry remained in the first bucket. Still, as Luhnow said, “Just because you can’t quantify it doesn’t mean it doesn’t exist.”
For Sig the idea of paying a lot of money for something he agreed was real but couldn’t value, funds that could otherwise be spent on qualities that he definitely could, made him a bit squeamish, if he was being honest. What was chemistry, anyway? Was it how players got along, how they shared information to attain a common goal? Was having experienced players always beneficial, or was there such a thing as too much experience held by players who pushed outmoded ideas on their younger teammates? Might it be better if players didn’t get along perfectly, to drive healthy internecine competition? And while it certainly seemed as if good chemistry mattered, and provided a team like the 2015 Royals with an edge, was that an availability heuristic—essentially, a memorable outlier? Not only did Sig have zero predictive information related to chemistry, but he didn’t even know what he might try to predict it. How could you try to reach a goal if you couldn’t define its parameters?
But Luhnow made the decisions. ... On Dec. 5, 2016, the Astros signed Carlos Beltrán. Luhnow gave him a one-year deal worth $16 million, committing to compensate him more in ’17 than anyone else on an Opening Day payroll that had nearly quintupled in just four years, to $124 million. Beltrán, in turn, would receive more than just $16 million. He would, he believed, get a real chance—perhaps his last one—to experience one of the only things he hadn’t during his 19 years in the majors: a World Series victory.
One night in 2009 an academic named Kate Bezrukova was watching a Yomiuri Giants baseball game in Tokyo with a colleague named Chester Spell. Bezrukova was an assistant professor in the psychology department of Santa Clara University, and she and Spell were in Japan attending a conference. The Siberia-born Bezrukova had always loved sports and was excited to see how the Japanese played baseball. To her surprise, not all the players were Japanese. Most were, but others were Korean, American, Venezuelan and even Australian. And yet they still combined to form a successful team. That year the Giants would go on to win their record 21st Japan Series. How did they manage to play so well together?
In the Tokyo Dome, Bezrukova and Spell realized that they had happened upon the ideal type of organization to study how demographic differences, called fault lines, could affect performance. In a lab it was difficult to measure performance, but baseball teams provided not only demographically diverse groups to analyze but externally valid results, most of all wins. Bezrukova and Spell directed their research assistants to perform fault line analysis on the demographics of all 30 major league teams, between the 2004 and ’08 seasons.
Those rosters revealed several different potential fault lines that could divide a team into ossified factions, called in-groups and out-groups, and hinder its performance. They could run between position groups: not just pitchers and hitters, but starters and relievers, and every-day players and reserves. They might be based on status, pitting older and better-compensated subgroups against younger, underpaid factions. And they could result from nationality, which could divide a clubhouse along fault lines of culture and particularly language. Such divisions could shift a team’s focus from winning to what the researchers called task-irrelevant cues, like competition and distrust between isolated subgroups, as well as restricted communication of actionable information and advice.
Bezrukova and Spell estimated that a major league team’s fault lines could account for up to three extra wins, or three extra losses, in a given season. A six-win swing could mean the difference between a club that won a World Series and one that didn’t even make the playoffs. Intriguingly, the teams that performed the best in Bezrukova and Spell’s fault-line analysis were not those who were the most demographically similar—mostly young, low-wage Hispanics, say, or older, highly compensated whites. They were instead those who had players who could cross-cut between a mix of subgroups, who could facilitate a complementarity, as opposed to a rivalry, based on their differences. Perhaps they had a player who was older and American, but made relatively little money. Or perhaps they had someone who was older and highly paid, but also Hispanic, and was particularly motivated to, in the academic parlance, deactivate his club’s fault lines. ...
When Beltrán arrived for his first spring training with the Astros in February 2017, he knew that he appeared as intimidating to his young teammates as any nine-time All-Star once had to him. Beltrán had long ago made a promise to himself. When he was a veteran, no young teammate would have to seek him out to mine him for his knowledge about how to prepare. Further, no young teammate of his would ever feel lost and alone, simply because of the language he spoke. Beltrán would always make himself available, if someone wanted his help. He also thought that he’d be doing something for himself: fostering a team that had a better chance of winning.
During his first days with the Astros, he approached each one of his new teammates—everyone, pitchers included. “My friend, I am here to help you,” he said. “Even if it looks like I’m busy, you won’t bother me. If you sit down next to me and ask me a question, I would be more than happy to give you the time that you need.”
By 2017 the Astros’ young players had a world of tools at their disposal that Beltrán hadn’t had as a young player. The club’s video room hummed with computers loaded with clips that could reveal pitchers’ tendencies to the percentile. A given night’s starter might throw a first-pitch fastball 75% of the time, and 85% of those first-pitch fastballs came in on the inner half of the plate.
It was useful information, particularly as pitchers, on average, threw much harder than they had when Beltrán was young. Hitters without a plan of attack, who intended to simply react to what the pitchers threw them, no longer had a chance. But the analytical information didn’t capture the other side of hitting.
“Analytics people, they understand the statistics,” Beltrán said. “But they don’t understand what the player is thinking.” What beyond the data could give a player the confidence to believe that every time he stepped to the plate, it was the pitcher who was in trouble, not him? And what could someone like Beltrán detect in a pitcher’s habits that could equip a teammate to understand not just what the probabilities suggested he would likely throw, but to know for certain?
Beltrán’s impact was impossible for the Nerd Cave—which Sig’s analytics group had nicknamed itself—to quantify, but the young shortstop Carlos Correa attempted to attach a number to it: seven. Of the 24 home runs he hit in 2017, by the end of the regular season Correa attributed precisely seven to Beltrán’s influence, to his showing him how to use video to break down opposing pitchers to a depth he had never before imagined, to his identifying their tells. Beltrán aided the Astros’ pitchers, as well. In ’16, the season after Dallas Keuchel broke out to win the Cy Young Award, his performance had disintegrated along with his team’s. His ERA went from 2.48 to 4.55. The first time Beltrán met Keuchel, in spring training, he gently suggested one reason why that had happened.
“Sometimes you held your hand above your glove last year before a pitch,” Beltrán told Keuchel. “If the ball showed, it was a fastball. If it didn’t, it was an off-speed pitch.”
“I appreciate you telling me that,” Keuchel said. Keuchel threw seven shutout innings against the Mariners on Opening Day. By the end of July his ERA remained below 2.00. Had Beltrán, in reality, accounted for just a fraction of the 40% boost in production that Correa attributed to him—not just for Correa, but for all the Astros he mentored, even the pitchers—then he would have been worth far more than that $16 million the club paid him. After 19 years in the big leagues, Beltrán knew that he might help Houston win in other ways too.
Even though Beltrán had helped to persuade Major League Baseball to require its clubs to hire full-time Spanish translators starting in 2016, when you walked into any big league locker room during any given club’s downtime, you almost always found two groups hanging out separate from one another: the English speakers and the Spanish speakers. There might have been a jagged fracture running down the middle of the clubhouse’s low-pile carpet. Beltrán thought there had to be a better way.
In Houston, Beltrán wanted to create not only an environment in which useful information could freely flow between players, but also the type of inclusive culture he longed for when he was young. He focused much of his effort on mending the natural division between its players who grew up speaking Spanish and those who did not. The Astros’ Opening Day roster included 17 Americans and seven Latinos. While many of the club’s native Spanish speakers, including Correa, José Altuvé and Marwin González, had become bilingual at younger ages than Beltrán once had, in previous seasons they tended to stick together and to speak in their natural tongue, while the Americans did the same. When an American player asked him a question, though, Beltrán intentionally often answered in Spanish before repeating his reply in English. He wanted to normalize both languages. “When Beltrán came over, that merged the clubhouse,” said third baseman Alex Bregman. “We’re all just way closer.”
Beltrán instituted other bonding strategies. He never understood why even the very good teams he had been on tended to treat regular-season wins as a matter of course. During spring training, he had enlisted the president of the World Boxing Organization, a friend, to commission two championship belts for him. After Keuchel pitched the Astros to a 3–0 win over the Mariners on Opening Day, Beltrán explained to his new teammates what they were going to do with them. After each victory, before anyone hit the showers, every member of the team would sit in his locker for the awarding of the belts, one to the hitter of the game and one to the pitcher of the game. Beltrán would distribute them that night, but thereafter the belt-holders from the previous victory would decide who got them. Each new awardee had to give credit to the other players who had performed well that game before beginning his own acceptance speech. Failure to participate would result in a $500 fine from the club’s kangaroo court, over which, of course, Beltrán presided.
The first night, a few players had watched the proceedings from the entrance to the shower room, towels around their waists. That was 500 bucks. “Baseball players, when you mess with their money, they listen,” Beltrán said. Nobody was fined after that. Soon the players acquired a new clubhouse sound system, to blast music during the ceremonies and whirling party lights. When reporters entered the room, as they were permitted to do after the ceremonies had concluded, they sometimes had trouble making out the faces of the players they were trying to interview. That was because of the fog machine.
One day in mid-July, Beltrán arrived at Minute Maid Park to find a curious message scrawled on the clubhouse’s whiteboard. funeral for carlos beltrán’s glove—3:30, it read. Though Beltrán still did fine in the outfield, and Statcast indicated that he got to precisely the percentage of balls he should have, it had been two months since he had played anything but designated hitter. He walked onto the field at the appointed time to find the rest of the Astros solemnly kneeling in a semicircle in the outfield around a box that contained the leather implement that had once won him three straight Gold Glove awards. Catcher Brian McCann stood in a priest’s robes next to faux tombstones, ready to deliver the eulogy.
Beltrán couldn’t stop laughing. It was funny, but also something more. It was one thing for him to try to create a winning culture by sharing the knowledge he had accumulated and demolishing demographic walls. But for his teammates to collectively concoct a way to tease him like this, someone who was far older and richer and more accomplished than any of them were? That was chemistry. Everything he had tried to do was working. ...
Luhnow believed that a clubhouse with good chemistry could persevere though periods of failure better than one without, and Kate Bezrukova and Chester Spell’s research independently confirmed that concept: deactivated fault lines could prevent poor results from snowballing. Nobody, aside from Jessica, knew that Carlos Beltrán was just three months from retiring. But he wasn’t done yet. In the World Series against the Dodgers, with the Astros’ first championship hanging in the balance, Beltrán would unleash his unquantifiable powers one final time [Beltrán discovered and told his teammates the "tell" that revelaed what pitch Dodger starter Yu Darvish would throw, and he de-escalated a racial incident involving teammate Yuli Gurriel.] ...
[After the Astros won the World Series, they realized they] had even further go to in understanding clubhouse chemistry. The Atros gave $16 million to Carlos Beltrán for reasons beyond his ability to hit home runs, but they had no way of predicting that without him, they would likely not have won a World Series in which he didn't have a hit. Kate Bezrukova and Chester Spell, the professors who conducted some of the first academic research into the effect of chemistry on a club's performance, had expanded their work for one of Sig's former employers: NASA. They were studying how factors like narcissism and aggression could impact the efficiency of a group of half a dozen high-performing people whi might one day be locked together in a shuttle the size of a one-bedroom apartment for more than a year, the time it would take to successfully complete a mission to Mars. Their techniques, they believed, could be applied to a baseball clubhouse. It would require the installation of cameras to analyze every interaction between players, both conversational and nonverbal, and biometric devices to record their stress responses, like their heart rates and cortisol levels. But it could be done.
There was always more data to be collected and harnessed. That underscored the inadequacy of the data you already had. ... Data could help guide best practices, but it was unwise to to confuse those with perfect practices. If people who denied the power of data could no longer compete, neither could those who believed that data alone provided an answer, not a tool. "All models are wrong," the British statistician George E. P. Box once wrote. "But some are useful."
As the data landscape flattened—as every organization gained access to the same set of numbers and technologies—then the new edge might come from the sources that were virtually impossible to quantify and to incorporate into statistical models, There would, in other words, always be a place for human intelligence alongside the artificial kind, and not just in baseball. There would always be a role for gut feels.