Big Data Baseball Page 8
The TrackMan system was installed for Arizona Fall League play in 2013, quantifying that Angels lefty Michael Roth had the top extension on his delivery of all pitchers in the league, releasing his pitches seven feet five inches from the pitching rubber, according to MLB.com, shrinking the distance from the mound to home plate. Perhaps this extension explains how Roth helped the South Carolina Gamecocks to two College World Series titles and rocketed through the Angels system although his fastball averaged only 85 mph.
“The offsets aren’t as large as you’d think they’d be, but they can add two miles per hour. That’s pretty substantial,” Fitzgerald said. “If you’re looking at batting averages against ninety-four [mph] compared to ninety-seven, you are seeing some significant differences.”
Fitzgerald worked with TrackMan for only six months, but his time there would prove critical. In the 2011–12 offseason, the Pirates were installing TrackMan at PNC Park when Dan Fox asked TrackMan officials if they knew of any bright young minds they had worked with who could help the Pirates. Fox was seeking a full-time assistant analyst to help him deal with the incredible amount of data flowing into the game. Fitzgerald’s name was dropped. The Pirates interviewed the tall, slim, green-eyed, dark-haired, fast-thinking, and quick-speaking Fitzgerald at the MIT Sloan Analytics Conference in early March of 2012. Fox liked that Fitzgerald had never worked in baseball, was a true outsider, and was willing to be more radical.
The twenty-three-year-old was hired, and Fitzgerald would quickly add value to the club. Just as he had become curious about the seemingly insignificant aspect of a pitcher’s extension with TrackMan, Fitzgerald had become fascinated by the impact of pitch framing.
“That was one of the first projects [Fox] and I kind of tackled,” Fitzgerald said. “How do we implement it? How do we improve upon it? What do we look for? We had a model we thought we liked to accurately evaluate pitch framing.… We looked at it for two weeks, just trying to poke holes in it.”
At first, like Turkenkopf, they couldn’t believe pitch framing could create that great of an impact. They had conversations in Fox’s office asking each other, “Can it really be that big? Can the effect be that significant?” They kept vetting their model. They tried to account for all factors that could play a role in a catcher’s opportunity and ability to frame pitches. For instance, what types of pitchers, and umpires, were catchers paired with? They kept being surprised at how significant and important this skill was.
Whom did their model keep identifying as by far and away the best pitch-framing catcher on the 2012–13 free agent market? Russell Martin.
“We said, ‘Okay, we have to get this guy,’” Fitzgerald said.
Huntington encourages subordinates to “pound the table” for players or ideas they believe in. By the second half of the 2012 season, a sanguine Fitzgerald was pounding the table for Martin. There was no need for formal presentations. Each night during home games, Huntington opened his suite in the club level behind home plate to his front-office staff. Everyone from the assistant general manager to the veteran graybeard scout, or fresh-faced, quantitative-minded intern, was welcome. Front-office staffers filled up to-go boxes in the media dining hall and headed up to watch the game and throw around ideas. They were free to talk openly and often, as just watching the game would spark a thought.
This in-person exchange of ideas and collaboration was important. The intersection of people and their thoughts is where the next big thing is often born. So in Huntington’s suite Fitzgerald began to make the case for Martin late in 2012 as the Pirates unraveled toward another losing season.
Fitzgerald can be animated when he’s excited, speaking quickly and enthusiastically, and he was adamant about the value of pitch framing. He said pitch framing was the biggest bang for the buck and completely undervalued. Heck, no one seemed to be making player-acquisition decisions based upon it. He was emphatic that Martin was the best value they could target in free agency. Fitzgerald and Fox argued that Martin should be their top free agent target, and he would represent a massive upgrade over the two starting Pirates catchers over the previous five seasons. They argued Martin would make every pitcher on the staff better.
The one big problem in convincing ownership that they should invest in Martin? In 2012 Martin had batted .211. While batting average is hardly an ideal tool to judge a player’s total worth, it was still by and large the number one statistic used by the general public in evaluating a position player’s value. Batting average was the one statistic displayed on every major league scoreboard when a player came to bat, and Martin’s was well below average. Huntington already liked Martin’s reputation as a quality defender and as a positive, veteran clubhouse presence, but the data from Fox and Fitzgerald would help sell the owners on a player with a questionable bat. Fox and Fitzgerald wrote a report on their findings on the value of pitch framing, and they also supplied anecdotal evidence of its value. Huntington took that report to the ownership of one of the lowest-spending teams in the game and asked for the most money the Pirates had ever given to a free agent.
“It helped that we had this information to supplement it,” Fitzgerald said. “I think it made it easier for [Huntington] to go to ownership and say, ‘Yes, I want to acquire a guy who hit .220 last year, but here is the reason why, and here is the value he brings.’”
On November 28, 2012, after a workout, Martin returned to his friend’s two-story town house in a hipster-populated neighborhood near downtown Montreal. The town house’s stone façade rested along a tree-lined street. On the first floor was a tattoo parlor, and above it were two levels of apartments with a rooftop terrace. Until work on his Montreal condo was completed, Martin was residing at his friend’s home. He liked it there. He liked the view of the skyline from the rooftop, where AstroTurf had been installed along with netting so one could hit golf balls. Here Martin made his decision.
Martin had been wined and dined in Texas by the Rangers, and the Yankees had made a halfhearted attempt to keep him. However, the Pirates had been the most enthusiastic and aggressive. Once he made his decision, Martin called his agent and related a story of how he had been at a Foot Locker earlier in the week to buy a new pair of workout shoes. The pair that caught his eye were black and gold: Pirates colors. He did not think of himself as superstitious, but he felt it was some sort of pull to Pittsburgh. In some ways, it was easy. The Pirates offered the most guaranteed money, and what if they were able to turn it around in Pittsburgh? That would be an amazing story to be a part of. At the least he wanted to rebuild his value.
The Pirates had just outbid the rest of the industry for Martin. Once the two-year, $17 million deal was agreed upon, an outcry came from the fans and media. Many in the local media thought it was a desperate move. Pittsburgh Tribune-Review columnist Dejan Kovacevic wrote in November 2012 of the Martin signing, “Let’s not pretend this transaction was anything other than what it was: An overpriced desperation move that’s going to hurt the Pirates in more ways than one.” By June, Kovacevic admitted his error noting Martin had been “masterful” and worth “every penny.”
But initially, the public did not understand, did not see Martin’s pitch-framing ability. And the rest of the industry wasn’t quite ready to buy into pitch framing the way the Pirates had. What the public wasn’t able to see was that Martin had just improved the Pirates by nearly 40 runs per season over their previous catchers, and it had nothing to do with his bat—which everyone kept judging him by. It had nothing to do with performance stats that could be found on the back of a bubble-gum trading card. It had everything to do with his framing skills. According to BaseballProspectus.com, in 2011 Martin saved 32 runs for the Yankees through framing. In 2012, he saved 23 runs. In 2012, Barajas had cost the Pirates 9 runs through framing. In 2011, Doumit’s glove had cost the club 15 runs. This value was hidden, and that’s the only way the Pirates could afford it.
But Martin’s would not be the only free agent signing by the Pirates tha
t caused head-scratching that off-season.
* * *
Back in 2001, baseball hobbyist and blogger Voros McCracken published what at the time was a radical idea. Wrote McCracken in an article published at BaseballProspectus.com, “‘You’re insane.’ That’s generally the response I get when I present the information you’re about to read. I’ve been accused of being the epitome of ‘pseudo-stat fan gibberish.’ I’ve even been accused of being Aaron Sele (a woeful Boston Red Sox pitcher) writing under a pseudonym. I’m not entirely sure why my little way of doing things stirs the emotions of people to such a large extent, but apparently it does. My belief? Well, simply that hits allowed are not a particularly meaningful statistic in the evaluation of pitchers.”
McCracken found that pitchers had significant control over their strikeout rates, walk rates, home runs allowed, and hit batsmen. However, his radical idea was that hits allowed by a pitcher were largely dependent upon the defenses behind him. That meant that earned run average, ERA, the gold standard to judge pitcher performance, was faulty.
McCracken continued, “I looked at the behavior of Hits Per Balls in Play. That’s where the trouble really started. I swear to you that I did everything within my power to come to a different conclusion than the one I did. I ran every test, checked every stat, divided this by that and multiplied one thing by another. Whatever I did, it kept leading back to the same conclusion: There is little if any difference among major league pitchers in their ability to prevent hits on balls hit in the field of play. It is a controversial statement, one that counters a significant portion of 110 years of pitcher evaluation.”
Hits not being dependent on a pitcher’s performance? That seemed counterintuitive. What was undoubtedly true was pitchers had control over the location of pitches—that is, whether their offerings were balls or strikes. Pitchers therefore had considerable control over whether they struck out or walked a batter. But if a ball was put into play, McCracken theorized a pitcher’s influence over whether a batted ball was converted into an out or a hit was largely subject to the ability and placement of the defenders behind them, while also tied to the dimensions of the ballpark they pitched in. McCracken was onto something as major league hitters had for decades produced a batting average around .300 on balls put in play. For instance, in 1920, major league hitters produced a .297 average on balls in play. Ninety-three years later, in 2013, major league hitters hit .297 on balls in play. Eight defenders in fair territory—not including the catcher—could only cover so much ground. With those defenders aligned nearly the same way for a hundred years, the amount of balls they were converting into outs had become something of a mathematical constant. McCracken found even the best pitchers of the modern era, pitchers such as Greg Maddux and Pedro Martinez, had little control over the outcome of a batted ball. The idea that some pitchers could allow fewer batted balls to become hits than others did not seem to hold up.
This thinking influenced teams from the Moneyball A’s to the Pirates, and outside hobbyists, to become more interested in measurements such as strikeouts and walks per inning, measurements they thought pitchers independently controlled and were more telling of true skills.
By the 2012–13 off-season, the Pirates had largely removed traditional statistics to measure pitchers’ performance. Neal Huntington rarely cited ERA, wins, or hits allowed when evaluating a pitcher. The Pirates tried to separate a pitcher’s performance from his environment as much as possible. Indicator became an often-used word by Huntington. The Pirates were thought to be paying particular attention to a new age metric created by baseball analyst Tom Tango and inspired by McCracken, something that was not found on the back of baseball cards, a metric focused on what pitchers did control and a better indicator of what a pitcher’s true ability was: fielding independent pitching or FIP. Dave Studeman, a writer and analyst for TheHardballTimes.com, took FIP a step further to estimate how many home runs a pitcher should have allowed. Home-run-to-fly-ball rates generally fluctuate over time, so one bad-luck home run year can also dramatically impact a pitcher’s ERA. Studeman called his refined FIP “xFIP,” a further attempt to isolate a pitcher’s true ability.
For example, in the season before the Pirates traded for A. J. Burnett, he had produced a 5.15 ERA with the Yankees, but his xFIP—on a scale designed to be similar to an ERA—was 3.86, which said he was unlucky because of factors out of his control. In his first year with the Pirates, Burnett’s ERA fell to 3.51 and his xFIP remained relatively constant at 3.40. It seemed like magic but it was simply math.
When the Pirates speak among themselves about a perfect pitcher, they talk about a big, physical pitcher who racks up strikeouts and ground balls and does not walk batters. They do not mention wins or losses or ERA, traditional standards for pitchers. The Pirates also knew they were not going to sign pitchers that carried all three elite skills—high strikeout rate, high ground-ball rate, and low walk rate—on the free agent market. Those types of pitchers were costly and contended for Cy Young Awards. So they asked themselves, What do you sacrifice? What’s easier to get? What’s harder to get? What’s harder to fix? What’s easier to fix?
For example, you can’t fix the height of a five-foot-nine player. You can’t make him six-three. You can’t make a right-handed pitcher left-handed, and you can’t make a pitcher throwing 85 mph increase his velocity to 95 mph. You typically can’t make a pitcher learn to develop an elite breaking ball. Of all the skills McCracken found to be independent in a pitcher, the ability to strike out batters was the most difficult to improve and typically most costly in free agency. The Pirates thought they could perhaps improve other independent skills of pitchers.
In the 2012–13 off-season, finding and signing a pitcher with upside was on the Pirates’ to-do list right after getting the team buy-in on the big data strategies and finding a free agent catcher. Specifically they required top-of-the rotation arms, the kind of pitchers that could consistently impact games. The kind of arms also labeled as aces, pitchers who struck out a lot of batters and did not walk many. But, of course, they ran into their usual problem of not being able to afford proven top-of-the-rotation arms, since starting pitchers enjoy the highest average salary in the game. For example, winter free agent ace Zack Greinke signed a $147-million deal with the Dodgers, roughly a third of the value of the Pirates franchise before the season, as estimated by Forbes. The game’s top pitchers were routinely earning $22–$25 million per year on the free agent market, or a third of the Pirates 2013 payroll.
The Pirates had just spent half of their available free agent funds on Martin, after which they retained the major league equivalent of pocket change. With their limited free agent resources, they had to make a choice between flawed but high-upside, bounce-back candidates, or durable but league-average, back-of-the-rotation arms. The Pirates opted for upside, and to find upside, as with Martin, they had to look beyond conventional numbers and take on risk. Huntington liked to say they had to pay for projected performance—really, projected improvement by a player—and not for successful history.
Some teams were still paying for successful history. They were paying in part for traditional, back-of-the-baseball-card statistics in the free agent marketplace, which meant that they were paying for wins, saves, innings, hits allowed, and earned run average. These results were in large part influenced by the defense behind a pitcher and the home park he pitched in. The Pirates and other analytically savvy teams went in search of something else. Value rested in the independent skills pitchers possessed. Forget earned run average and wins, the traditional statistics with which to judge pitching performance. As the Pirates analysts and scouts scoured the free agent marketplace for a starting pitcher with upside, they went back to one pitcher they could afford with intriguing indicators whom, they thought, their coaching staff could improve, and whom they specifically thought Russell Martin could improve. They kept coming back to Francisco Liriano.
Liriano was once one of the most prized young
arms in baseball. He had a six-foot-two frame, broad shoulders, and a slim waist, the ideal V-shaped torso scouts looked for in an athlete. The San Francisco Giants saw him work out first as an outfielder in the Dominican Republic, but were most intrigued with his rocket arm and that he threw left-handed. He converted to pitching and hit the upper 90s with his lightning-quick arm. He was more of a dream, a project, a raw and unrefined pitcher with potential, when he was traded to the Minnesota Twins in 2004 in a package for catcher A. J. Pierzynski. With the Twins he developed a wicked slider, polished his command, and demonstrated a precocious feel for a changeup. He rocketed up through the Twins system, overwhelming hitters in such places as Fort Myers, Florida, and New Britain, Connecticut, and was ranked as the sixth-best prospect by Baseball Prospectus and Baseball America entering 2006. He began the year with the Twins, featuring a fastball that averaged 95 mph along with a swing-and-miss slider, and a fall-off-the-table changeup. USA Today published a feature on Liriano headlined “Scarier than Santana,” as in Liriano’s older teammate Johan Santana, who would go on to win his second American League Cy Young Award in 2006. The excitement was washed away in August of 2006 when Liriano felt pain in his left elbow. He tried, unsuccessfully, to pitch through the injury. Doctors found he had torn the ulnar collateral ligament in his elbow, meaning he needed Tommy John surgery. The procedure is often viewed as an automatic cure for an elbow injury, with its sole negative side effect being the time a pitcher will miss while rehabbing, as if the surgery were merely an inconvenience. Liriano’s case was a reminder that the surgery is not a perfect science, and not every pitcher returns to his previous ability, at least not initially.