Big Data Baseball Read online

  On February 17, 2012, the New York Yankees so wanted to be rid of the final two years of A. J. Burnett’s five-year, $82 million deal they agreed to trade Burnett to the Pirates—and pay half his salary—for two marginal prospects. Burnett had not fit well in Yankee Stadium or the American League East Division. He had posted ERAs above 5.00 in back-to-back seasons in 2010 and 2011. He was stubborn with his approach. And he had lost a tick on the fastball that had made him one of the game’s top young flamethrowers with the Marlins earlier in his career. Thanks in part to their database, the Pirates’ analytics department saw a buy-low opportunity in a pitcher who had an underutilized but effective sinking fastball. They saw a pitcher who had perhaps become too leery of pitching to contact in the cramped ballparks, against the power-packed lineups, of the AL East, meaning Burnett had focused on generating swings and misses from opponents, trying to blow four-seam fastballs by them, rather than attempting to induce weak contact and ground balls with slower, sinking fastballs.

  “Being able to try to add that to a pitcher’s arsenal was a concerted effort,” Dan Fox said of the ground ball and two-seamer. “Targeting guys who you think might have that ability is one part [of the plan]. The ground-ball rate [increasing], the difference we saw with players we already had … that’s all Ray [Searage] and Jim Benedict being able to teach it.”

  The Pirates’ analytics department thought Burnett could have more success by leaning on his two-seam fastball, but getting him to employ it fell to the coaching staff. On Burnett’s right arm is tattooed in Latin STRENGTH THROUGH LOYALTY, a tribute to his wife and family. Burnett had also become incredibly loyal to his simple, two-pitch mix—a four-seam fastball and a curveball. Unlike other aging pitchers he had never developed a third pitch. But he had now posted three straight seasons of 10-plus percent home-run rates.

  Searage had his work cut out for him. His plan was to sell Burnett with a simple message: Searage was not looking to overhaul pitchers, he was trying to improve one or two elements. He was into tweaks, not reconstruction.

  Did Burnett buy in? Check out his pitch usage with the Yankees in 2011 versus his pitch usage with the Pirates in 2012:

  Like Morton, Burnett’s pitch mix radically changed with the Pirates. And like Morton, Burnett’s results also evolved. Burnett’s ground-ball rate soared. He went from posting a below-average ground-ball rate in all three of his seasons in New York to the game’s second-highest ground-ball rate—56.9 percent—among qualifiers in 2012. He began to resemble the dominant pitcher he was earlier in his career.

  That Searage and Hurdle could get buy-in from Morton was one thing. Morton was an amiable, thoughtful pitcher in need of help. That Searage and Hurdle could get buy-in from Burnett was remarkable. Burnett was a stubborn, headstrong veteran who had had success. He could be difficult to manage. His own teammate and friend Jeff Locke told the Tribune-Review there was “A.J.” and there was “J.A.” The J.A. was for “jackass.” But somehow Searage and Hurdle were able to connect with all sorts, from the headstrong and macho Burnett to the analytical Morton. If they could get buy-in from Burnett and Morton, from both poles of the cooperation spectrum, could they get buy-in from everyone else?

  Communication was key with the pitching philosophy. And the philosophy wasn’t just about throwing a lot more two-seam fastballs, it was also about location. Pirates pitchers were told they were not just going to change what they threw, they were going to change where they threw. Unlike the shifts and pitch-framing philosophies, where the data was discovered or produced at the top of the front office and delivered down to the coaching staff and players, the third prong of the run-prevention strategy, the approach on the mound, was driven from the bottom up.

  Hurdle wanted and needed his assistant coaches to be on board and feel ownership of this entire philosophy: from the defensive alignment to the pitching plan. That’s the only way the effect would be maximized. He wanted his assistants to feel empowered. He wanted there to be collaboration. Hurdle asked the coaches and pitchers to buy in to the three-pitches-or-less doctrine. Searage and Benedict were on board with some of the concepts, they had their own ideas to add, and they had questions. Hurdle wanted his assistants—from Leyva to Searage—to ask questions. Hurdle didn’t want to just hand down orders. He wanted his assistants to improve the big-picture plan. That would foster a sense of ownership and improve commitment and execution. Hurdle preached to his players, coaches, and front office staff that “We’re trying to all get it right together.” During the off-season and that spring, Hurdle wanted his staff to think of ways to enhance the plan, to test and share theories, and to make these plans better and more refined. More than anything else, he wanted them to ask questions of not just him but of the analytics staff, which was sending down an avalanche of data from the front office. He wanted his coaching staff to have the analytics staff test theories the coaches had, not just subscribe to those handed down to them. He wanted the coaches and analytics team to respect each other. That was the only way they would maximize and integrate the data. To get his position players to follow along, to get his pitchers to buy in, Hurdle had to present a united front. He had to have his assistant coaches on board. He had to have them accept not just the data-based message but the messengers as well. Creating a culture of respect and communication was still a formidable barrier.

  Hurdle’s assistants were all veteran coaches. Searage’s hairline had receded after four decades of watching hundreds upon hundreds of pitchers. Nick Leyva’s dark hair had grayed. Hurdle’s bench coach, Jeff Banister, was a baseball lifer, a longtime minor league catcher and Texan, tall and broad-shouldered, his skin creased after thousands of hours under the sun on baseball fields. Banister had been part of the organization longer than anyone else. He was a twenty-fifth-round draft pick by the Pirates in the 1986 draft. He had only one major league at bat in his career, resulting in a basehit in 1991 against Dan Petry. Banister began coaching in the minor leagues in 1993. Entering 2013, he had been in the Pirates system for twenty-eight years, and mostly all he had experienced was losing.

  For these coaches to lose any disdain they held for the analytics staff’s invasion of the sacred space of their clubhouse and for the data to flow smoothly from the front office to the field, their voices had to be heard and their questions answered by the club’s analysts. This was why the Pirates had Fox and Fitzgerald travel to spring training in 2013 and made them fixtures in the clubhouse. The two men had to be available to answer questions by players and the staff. They were there to become familiar faces throughout the season. Just as Hurdle had become more comfortable with and had gained trust in Fox and Fitzgerald in 2012, now he wanted the same for his entire staff and players in 2013.

  Some of the assistant coaches were skeptical of having every pitcher trade in his trusted four-seam fastball for a two-seam fastball against every major league hitter. They felt to convince their pitchers to adopt this approach they would need evidence. And they didn’t want generic, one-size-fits-all evidence; they wanted data tailored to each individual hitter-pitcher matchup. So the staff asked for the analysts to identify the pitch types, locations, and velocities that every major league hitter was most likely to beat into the ground.

  Fox and his team had another project, too. Searage and Benedict had always believed in pitching inside, but most pitchers resisted buying in to this for several reasons. For starters, if a pitcher missed the inside of the plate, leaving the ball over the center of the plate, then the batter was likely to hit the ball deep for a home run in this modern era of shrinking ballpark dimensions and stronger hitters. That was perhaps the most humiliating scene in baseball, to stand in the center of the field alone and wait for a new baseball from the home-plate umpire while your opponent gleefully floats around the bases. Moreover, most pitchers did not enjoy missing inside toward the batter. If you hit a batter, retaliation against one of your own teammates was likely. Still, Searage and Benedict felt pitching inside was important and could
accentuate the ground ball and a shift plan. They just needed some statistical evidence to show to their pitchers. The coaches wanted to see every major league hitter’s performance against pitches on the outer half of the plate after being pitched inside earlier in an at bat. What was the psychological and performance effect of pitching inside?

  When the coaches went to Fox and Fitzgerald with these questions, the coaches were struck by how receptive the two were to them. The analysts did not brush off the questions or roll their eyes; they were interested in the coaches’ theories. Fox and Fitzgerald did not believe they and their algorithms had all the answers. They wanted to hear the questions.

  “I have to find a way to not just encourage but to inspire and challenge these guys in different areas so I can help match them up in certain areas where their skill-sets will shine,” Hurdle said of his staff. “Some of them found working with these analytics people was really fun and they probably didn’t have that feeling going in.”

  The assistant coaches’ questions led to research that might not otherwise have been considered by Fox and Fitzgerald. If every analyst in the sport is operating off the same big data—such as PITCHf/x and TrackMan information—then asking the right questions is paramount.

  “A lot of times a coach or somebody will have something in the back of their head, but they won’t write an e-mail. They are not going to compose something and give it to you,” Fox said. “But they’ll come over and say, ‘From a player-development perspective, track this metric. I’d like to look at it in terms of player [performance].’ I just try to be as available and visible as possible because you never know what you are going to get. They all have these vast databases of player comparisons and situations and strategy that I don’t have.”

  Fox and Fitzgerald looked into the coaching staff’s questions about pitching inside, and prior to the 2013 season they found that pitching inside would indeed have a psychological effect on batters that would create even more ground balls and further enhance the plan. The numbers showed opponents were more likely to pull outside pitches on the ground after being pitched inside earlier at bat. After being pitched inside, players were less willing to aggressively lunge at outside pitches. Now the coaching staff had the data they needed to get their pitchers to pitch inside, but would the pitchers execute the plan?

  On June 13, 2013, the Pirates held their collective breath as Morton returned to the major leagues. He wasn’t great in his first start since returning from surgery. He allowed 4 runs—2 earned—over 5 innings. The bull pen was worse and the Pirates were routed 10–0. But Morton did offer some evidence that the improvements he had made pre-surgery had stuck: he produced 2 fly outs and 6 groundouts. His velocity was steady, in the low 90s. And although not observable in the box score, Morton’s movement had returned.

  The ground-ball trend continued on June 18 at the home-run-friendly park in Cincinnati. In the first inning, Morton induced Zack Cozart to pound a 92 mph sinker into the ground to third baseman Pedro Alvarez, who began a key double play. In the fourth, Morton’s sinker reached 94 mph as Brandon Phillips slammed it into the ground back to Morton, who flipped to first base for the out. With the Pirates leading 3–0 in the fifth, Morton found himself in his first jam of the game: runners on second and third with two outs, and the left-handed Jack Hannahan coming to bat. Morton threw a wicked 93 mph sinker that darted away from Hannahan, who grounded sharply to shortstop Jordy Mercer, who threw across the diamond to end the inning.

  Morton’s first two starts began a season-long trend of his opponents chewing up the turf with ground balls. He became perhaps the best ground-ball pitcher in the major leagues, and as the season went along, he became even stronger. Had he not missed the first two months of the season and had pitched enough innings to qualify, Morton would have led baseball in ground-ball rate at 62.9 percent. In a 5-start stretch in August, Morton posted an unheard of 43-to-5 ground-ball-to-fly-ball ratio.

  And it wasn’t just Morton who went on to have a career-best ground-ball season; it was nearly every Pirates starter. Almost every Pirates pitcher decreased his four-seam usage and increased his two-seam frequency, and nearly every Pirates starting pitcher produced a career-high ground-ball rate. According to PITCHf/x data analyzed by BrooksBaseball.net, Liriano did not throw a single four-seam fastball in 2013. Only sinkers.

  On June 14 Jeff Locke took the mound at PNC Park against the Los Angeles Dodgers. Locke was in the midst of a surprising breakout year after entering the season with only 10 major league starts and a 6.00 ERA. After allowing a leadoff single to Yasiel Puig, Locke induced Nick Punto to pound a sinking fastball to Clint Barmes, who began a double play. In the third, a running two-seam fastball got in on the hands of Puig, the Dodgers’ rookie sensation, who grounded out weakly to second base to end the inning. After crossing first base, Puig tossed his hands, rolled his eyes upward to the sky, and shook his head in frustration.

  With the Pirates leading 2–0 in the fourth, Hanley Ramirez came to bat with one on and one out for the Dodgers. Ramirez beat another Locke two-seamer into the ground, leading to another inning-ending and threat-ending double play.

  During a 3-start span that included 16 straight scoreless innings, Locke produced 23 ground-ball outs. The approach made him an All-Star.

  “Our pitching program has by far been the strongest of the [plan’s] outcomes,” said Huntington. “It’s a front-office [preference], but it’s also a philosophical focus. We like guys that throw strikes and get ground balls.… If the ball is not leaving the [ballpark], you have a chance to make outs.”

  In June, the Pirates staff produced an amazing 53 percent ground-ball rate, and the club had gone 17-9 in the month, its third straight winning month to open the season, allowing just 0.7 home runs per 9 innings. The ground-ball trend continued all season. In 2013, the Pirates staff combined for the highest ground-ball rate— 52.7 percent—since groundball-fly ball data began being recorded. The data is available back to 1988 at BaseballReference.com. In 2010 the Pirates ranked fifteenth in baseball in ground-ball rate at 44 percent. That ranking climbed to seventh in Hurdle’s first year in 2011 and sixth in 2012, and in 2013, with the Pirates’ concerted focus, the rate was 4 percent greater than that of the next-closest team, the St. Louis Cardinals.

  Consider the remarkable changes by individual pitchers in the following charts:

  Four-seam fastballs were the traditional go-to pitch for major league pitchers. In 2008, the first year PITCHf/x began tracking and labeling every pitch in baseball, the vast majority of fastballs thrown were four-seam, accounting for more than 50 percent of all pitches. By 2014 the rate had decreased to 34.6 percent—a dramatic decline, in part tied to the Pirates. In 2008, 3.8 percent of all pitches in baseball were two-seam fastballs, according to FanGraphs.com, but in 2014, that number spiked to 14.7 percent. The two-seam fastball gave the Pirates a way to reinvent pitchers and create something precious—pitching depth.

  There was one catch. To create as many ground balls as possible, the Pirates threw inside more than any other team. Their pitchers hit an MLB-high 70 batters in 2013. The Pirates batters were, not coincidentally, hit by an MLB-high 88 pitches. By late June the approach was clearly bothering opponents.

  In the ninth inning of a game at Cincinnati on June 17, Reds closer Aroldis Chapman buzzed Pirates second baseman Neil Walker with an up-and-in, 100 mph fastball that dropped Walker to the dirt.

  “You try to give players the benefit of the doubt, but given the history between us and Cincinnati, it makes you wonder,” Walker told reporters afterward. “Balls should not be anywhere near anybody’s head. I was hit by [Mat] Latos last time at our place. I was fine with it because he hit me right in the butt.… It’s tough to get me fired up about things, but when you see a hundred-miles-per-hour fastball come near your face, you have to wonder a little bit.”

  The next night Charlie Morton began his start by drilling Reds leadoff hitter Shin-Soo Choo in the calf with a fastball. It contin
ued a yearlong feud between the Reds and the Pirates. Still, Pirates batters kept mostly quiet about being retaliatory targets. They knew the plan on the mound was working. They saw it was helping them win.

  The Pirates led baseball for much of 2013 in defensive efficiency, which is the number of balls hit in play that are converted into outs. They finished fifth for the season, a meaningful improvement over the previous seasons of the Huntington Era, when the Pirates ranked in the twenties.

  “Given the ground-ball nature of our staff, the difference shifts could make in the number of balls in play we turn into outs, and the impact it could have, was dramatic,” Huntington said.

  More ground balls being hit into the shifts helped the Pirates convert an extra 2 per cent of batted balls into outs throughout the season’s first half compared to their 2012 level. This percentage might seem insignificant, but over the course of a season, when 4,500 balls are put in play that’s equal to 90 hits being turned into outs. In the first half, the Pirates ranked first in baseball in defensive efficiency, the rate of converting batted balls into outs. They had ranked last, thirtieth in the sport in defensive efficiency in 2010 and twenty-fifth in 2011. By keeping more balls on the ground, the Pirates allowed a major league-low 101 home runs in 2013, 52 fewer home runs than they surrendered a year earlier. Thanks to the shifts, despite not having a single Gold Glove-caliber infielder, the Pirates limited opponents to a .224 batting average on ground balls, the fifth best mark in baseball. This was one of the greatest defensive improvements in baseball history. Ten runs added, or saved, is roughly equal to 1 win. With a 93-run improvement—from 68 defensive runs saved in 2013 compared to -25 in 2012—that converted to roughly 9.3 wins added.