:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/22719741/20120918_kkt_ab9_346.0.jpg)
Baseball is a game that breathes statistics—more than any other sport. It inhales them in with every decision, from the most meaningless midseason pinch-hit to the franchise-altering trades. Then, it exhales them back out again, at every moment—with every at-bat, every out, every game and every season. But for as many numbers as we have, and how broadly we use them, there’s one area where we assume numbers play but a small role. Though, I couldn’t help but ask.
"So, looking at this managerial search, is that something you're putting data together for—does anyone ask you, for example, how the Tigers are hitting?"
As, unbeknownst to us, news was breaking elsewhere at the very time we were grabbing lunch on Tuesday—that the Mariners would hire Detroit Tigers hitting coach Lloyd McClendon to be their next manager—the individual sitting across from me at the small Thai restaurant in Seattle's Pioneer Square grinned, and didn't hesitate a moment before responding.
"Yeah I gotta stay away from that one."
But while this particular topic was out-of-bounds, for understandable reasons given the timing, I had the opportunity to sit down and learn a considerable deal about the Mariners' analytics efforts through a one-on-one interview with one of the baseball operations analysts, Jesse Smith.
Smith joined the Mariners organization in the fantastical way these things seem to happen with numbers guys. After studying psychology at the University of Chicago, and working on projects that included a website that tested how informed users' political opinions were and a computerized negotiation web tool used in one of his pscyh professors' classes, his passion for baseball turned him to a dataset he had a bit more interest in.
"It just occurred to me I wanted to work with baseball data," said Smith, who grew up a fan of the Oakland Athletics and their Moneyball strategies. "I wanted to try to build a projection model, just to crush my friends in fantasy baseball, just to see if I could."
Smith partnered with a friend and created the model from the free database available at the time from Major League Baseball's website. When it was complete, he wrote a paper on it and sent the piece to 20 of the 30 teams. It garnered some interest, and the Mariners brought him in for an interview. They saw the potential in Smith and his model, and hired him.
"Yeah, it’s been a pretty awesome ride so far," said Smith reflecting on it.
But while Smith's story is an enjoyable one, this story here is on the Mariners' analytics department and the role it plays within the franchise.
The department, of course, is bigger than just Smith, who works almost purely on the professional side. He teams up with Baseball Operations Analyst Wesley Battle, and when you're thinking of the people who spend the entirety of their days pouring over pro players' numbers in a constant quest for value, this is who you're thinking of.
But, naturally, there are more than two guys working the numbers.
The organization also has Anthony Aloisi parsing through amateur data—though, naturally, there is considerably less of that. Jordan Bley works with scouting data that includes training and psychological information, along with Kenny Wade. Then you have Caleb Peiffer doing a great deal of arbitration work and punching the numbers on contracts.
Though all of these individuals, and others, have their particular specialties, various individuals can be called in at any time to work on specific projects.
"Whether that’s from [General Manager] Jack Zdueriencik or [Assistant General Manager] Jeff Kingston or [Amateur Scouting Director] Tom McNamara or whoever—lots of people have requests for anyone who’s available and has free time to get pulled onto anything."
Those projects can include particular players, park factors or whatever else.
"There's always something, because the game is constantly changing," Smith said.
But while there's always work to be done, data to work with—" I could never get through it all."—Smith said he's granted a great deal of freedom in his role, depending on the time of year, to provide value in any way he can.
And various portrayals of teams' sabermetric efforts, Moneyball especially, weren't wrong on this: it is all about about value. That much is obvious. Smith said the search for undervalued assets "is a very large focus of our job."
So what tools are at play? What are they using?
As could be expected, there is the holy grail of fans' analytical work—we know it's out there, we know teams are using it, and we wish so much we had access: it is, of course, HIT f/x.
"That allows you to have a much more granular look at the game of baseball, and its applications are immense," Smith said. "It is the coolest thing. You learn a lot about players."
But while HIT f/x is an exceptionally valuable tool, it's a tool at the disposal of every franchise—though, as Smith noted, just because every team is using the tool, it doesn't mean that every team can get value out of it. A key component of extracting value from various analytics tools, aside from the skill of finding the meaning in the noise, is staying up-to-date with the latest developments as there are constantly new technologies available.
The team is frequently pitched on new datasets. Though, Smith didn’t want to get into specifics with regards to which services the teams subscribes to as not all teams subscribe to every service, and he didn’t want to reveal which the Mariners prioritized.
So, we have the numbers, or a sense for what the Mariners gather and what they’re working with, but what’s the interplay between the analytics team and other departments? Before we go up, let’s first go sideways.
The relationship between scouting and analytics has been viewed by many, particularly in the Moneyball-past of a decade ago, as an adversarial one. With every decision, you could go one of two routes: you could trust your eyes, or you could trust the numbers.
But as Smith explained, that is not the case. Though it may have been that way at a time—and it’s worth noting the data was much worse back then—this ideology no longer applies.
The goal of scouting, Smith said, is to paint a complete picture of a player—and you can’t do that without the numbers. Often times, a scout will lean on the analytics department to confirm or deny what they’re seeing with their own eyes. And, naturally, every now and then there will be some disagreements. The numbers and the eyes don’t always line up.
As Smith explained it, "it’s really an opportunity to probe deeper." With advanced tools already in place and proprietary software moving in this direction, teams can pair the numbers with the scouting by leaning on video tools capable of displaying every play a player has ever made.
And, while the scouting team can glean a great deal of information from the analytics' department, so too can the analytics professionals from the scouting department. For example, at the amateur levels, where there's significantly less data—particularly, less good data—scouting is paramount. But it happens at the professional level as well, as Jesse notes he's "learned so much in getting access to scouting."
So that's how the information travels laterally, but what about vertically? How does the research Smith and other members of his team do reach the decision-makers, the coaches and then the players? It isn't quite as glamorous as you think.
"There are lots of meetings," Smith said.
But, of course, that could be expected. How does it all work, what happens in said meetings? It goes down as one might assume.
The decision-makers in the Mariners' organization have any number of people providing them information on a player, and there is bound to be divergence on everyone but the most established big leaguers. And, as Smith explained, the majority of the time, decisions are made on the players that aren't established, that aren't superstars.
And while some have the vision of the analytics department standing up and fighting for their prized undervalued asset, it can sometimes be the opposite.
"Players are very unique—you have to ask yourself why would this player not conform to a model, and why should we really trust the data here," said Smith. "You always have to look for flaws."
But for those who have built the belief that General Manager Jack Zduriencik has turned his back on the analytics department, that doesn't seem to be the case. As is standard across baseball, information from the ops team does travel up through Zduriencik and down to the coaching staff and the players, but Zduriencik frequently solicits the analytics department's guidance.
"[He] has always explicitly encouraged us to point out anything, including on-field strategy, that could improve team performance," Smith said.
And for those of us out here wishing we could influence this organization to do things differently, you might be surprised to know that you can.
Smith and the rest of the team read analysis from third parties, from fans and more established writers every day, including Looking Landing and other blogs. Most of them got into baseball because they were reading that stuff and they consider it part of their jobs to say up-to-date.
"If people post original ideas, it’ll get forwarded around if it’s good," Smith said. "It gets read."
So if you want to make a difference, if you see something you want changed—research it and write it up. You never know whose ears it might reach.