When the Detroit Tigers first hired Brad Ausmus, many fans thought, “he’s young and highly educated, I’ll bet he’s into sabermetrics!” That lasted about a month before Ausmus went on the radio in January of 2014 and said, “people got the thought that, because of my age and my background, I’m a sabermetrician…I’m really not.” His in-game strategical decisions over the past two seasons have more than proved that he wasn’t joking.
But with the promotion of Al Avila to the general manager position, there’s new hope. Avila wants to get the organization into the big data game, the Tigers have hired a slew of analysts, and now there’s a new question on the table: can Brad Ausmus be taught to embrace the #Mathletics revolution?
There’s at least one example of where it might matter. Avila said recently that he’s looking for the best relievers available, not based on whether they’re right- or left-handed. In contrast, Ausmus loves him some LOOGY action, and it doesn’t always matter whether the data supports the matchup. There’s some potential for not-so-hilarious hijinks here, if, for example, Avila stocks the 2016 bullpen with a high-quality lefty of the sort that should be facing three batters minimum, and Ausmus regularly blows his wad early by using the guy for LOOGY-type, one-batter scenarios.
How easy is it, really, to get the advanced metrics off the spreadsheets and onto the playing field? The story of the 2013 Pittsburgh Pirates, as recounted in the book Big Data Baseball, by Travis Sawchik, contains a real-world blueprint — and the blueprint smells faintly of hope for Tigers fans who want to see the team making smarter decisions on the field.
The will not to lose
Grit, heart, hustle, and the will to win may not be actual factors in baseball — at least, not dependable, repeatable, measurable factors — but for the Pirates and their front office, there was certainly a will not to lose anymore seasons. The 2012 Pirates finished the season at 79-83, which marked their 20th losing season in a row.
Their payroll was the fifth lowest in baseball in 2012, and it only increased by about $16 million in 2013, good for the number 20 spot out of 30 teams. They weren’t going to buy their way to a winning season, and the will not to lose forced the front office to try something different.
A saber-savvy GM
Neal Huntington was part of that group of stats-lovers in the early- to mid-1990s that helped the Cleveland Indians get from 105 losses in 1991 to five straight division titles during the 1995 through 1999 seasons (that group also included Paul DePodesta, later of Moneyball fame, as well as current Blue Jays president Mark Shapiro). Huntington was no stranger to advanced metrics, as he had seen first-hand how they could help struggling teams return to contention. For the Pirates to start taking some radical steps, it needed to start at the top, with the GM.
One of the first things Huntington did when he became the Pirates’ GM was to start building an analytics team, beginning with Dan Fox, an analyst and software developer who had written for Hardball Times and Baseball Prospectus, a guy who loved Bill James’ work and wanted to take it further. Fox built the Pirates a new system from scratch for collecting and processing advanced metrics, and gave them the foundation they needed.
Old dogs that want to learn new tricks
Manager Clint Hurdle wasn’t a sabermetrician by any stretch. He was an old-school baseball guy who came up through the system in the late 1970s and played through the late ’80s, well before the analytics revolution. But he was fired as the Colorado Rockies manager in 2009, ended up working as an analyst for MLB Network, and that’s where he was introduced to the world of “big data.” Sawchik says in his book that Clint Hurdle’s colleagues at MLBN “directed him to the Web site Fangraphs.com…where Hurdle began conducting his own investigations, often looking for statistical data to support a theory or hunch he had for an on-air segment.”
The net effect? Hurdle is quoted in the book as saying, “It was kind of like being in a Wizard of Oz setting with no ramifications…Once I got started, it was hard to stop.”
When it came time for GM Neal Huntington and his stats-obsessed partner Dan Fox to introduce some changes in the way the Pirates were going to play the game, they had a manager in Clint Hurdle who was at least open to new ideas, and who could help them communicate those ideas to the players.
Try it in a sandbox first
So what was the big change that was going to save the Pirates’ 2013 season? Defensive shifting. A looooot of defensive shifting. They also wanted their starting pitchers to start throwing more two-seam fastballs and sinkers, to induce more ground balls. Pitching for weak contact instead of strikeouts was the plan, which would simultaneously result in lower pitch counts that would let their starters pitch deeper into games, and fit neatly with the new defensive strategy.
But before they dropped this information bomb on the major league team, they experimented with it at the minor league levels first. When they saw the results they wanted, it was much easier to promote the theory to the big leagues. It’s always easier to get others to accept major changes when you have tangible results to back it up.
Everybody get along now
This may be one of the most critical steps that led to the Pirates having success integrating advanced metrics with on-field strategy. Huntington made sure that Fox and the rest of the analytics team had regular contact with the manager and the players. The mountains of data were distilled into visual charts and presented to the team, and the analytics group encouraged collaboration and challenge.
Players and coaches were invited to question the metrics and help refine the results — and the guys crunching the numbers were grateful for the input. “The game is played on the field and not on a spreadsheet” may be a line designed to discount sabermetrics, but there is a lot of truth to it, and in this case, the Pirates found a way to blend bits and bytes with the real-world experience of the guys on the field.
Players have to buy in
A major league shortstop who has spent years learning to play his position isn’t necessarily going to be comfortable shifting to shallow right field just because the math says he should. Pitchers who are trying to keep their stats respectable aren’t going to be thrilled when a weak grounder bounces through an un-guarded left side of the infield and a run scores. Some of the Pirates’ players definitely had objections to radically breaking from traditional baseball strategy based on what a computer was telling them to do.
If the players aren’t on board, the manager will hear about it. If the manager isn’t on board, the GM will hear about it. The whole thing can go horribly wrong if there isn’t buy-in across the board. The Pirates went about things in the right way, though, getting Clint Hurdle’s support early on, and making sure the analytics gurus and actual players were working closely together and in collaboration.
In the end, the slow but steady stream of results proved that the data-driven theories were right, and the Pirates went from being a team worth -25 defensive runs saved in 2012, to being a team worth 68 defensive runs saved in 2013. They finished the season with a record of 94-68 and broke the 20-year losing streak. It wasn’t good enough to beat the St. Louis Cardinals and their 97-65 record for a division title, but it did win them a wild card spot and a chance to play in the postseason for the first time since 1992.
Big Data in Detroit
The Tigers aren’t coming off a 20-year losing record, but they are coming off a very disappointing last-place finish, and they are playing their game on the dime of an elderly man who has made it clear that he wants a World Series Championship, not “now,” but “right now.”
And while the Tigers are certainly not ranked among the lowest-spending teams in baseball, they also don’t have a ton of available payroll after letting Miguel Cabrera, Justin Verlander, Victor Martinez, and Anibal Sanchez swim in the gold vault with permission to take as much as they could carry. If the organization is going to go worst-to-first in 2016, they’ll have to be very precise and calculating in how they spend what’s left of the budget.
They have a saber-savvy GM, they have an analytics team, and they have the will not to lose. Can they get buy-in from Ausmus and the guys on the field to bring big data to the diamond?
