The Royals hired Mike Groopman as an analytics intern in 2008 and since then have hired him full-time to head up their new Analytics Department, hiring geophysicist John Williams and computer engineer Daniel Mack. The Royals, once maligned for their dismissal of analytics, have been cited as "believers" more recently with their emphasis on defense and high-contact rates for hitters credited for some of their success last year.
Williams and Mack made a presentation at Saber Seminar over the weekend. Jimmy Wulf of Sons of Sam Horn does a good job covering their presentation here. Jeff Zimmerman was also in attendance (taking time out from presenting!)
The pair detailed the Royals rise to the World Series. They stressed the need to contextualize playing environments when evaluating players, particularly at the amateur level. When developing new statistics, the pair stressed the need to make it measurable, useful, and explainable. The statistic should measure some skill of a player, rather than serve as a badge for a situation like the save. The data should be useful in translating some aspect of the game, and expandable to be used for future work. The statistic should also be explainable, to get buy-in from the front office and personnel.
Getting buy-in from executives, coaches, and players has been something brought up by former player Gabe Kapler as a new frontier in analytics. Williams agreed that packaging the conclusions the right way was important to get everyone on the same page. As Wulf writes:
"A lot of our job is to take this info and find cogent ways of presenting that to the coaches and players." Williams added, "If I’m talking to coaches or GMs but saying something that only makes sense to a small group of academics, I’ve failed the people paying me money."
Williams mentioned how character was a factor in player evaluations, particularly on long-term deals. This is a philosophy long espoused by General Manager Dayton Moore. Wulf writes:
On the outside, he saw the quantitative analysis as definitive. On the inside, now both responsible for that analysis and interacting with the players, he views it as "only the first step for the people making decisions." They even take the same approach when hiring in the front office, and Williams finds that easy definitions of traditional-versus-scouting GMs do not hold up to scrutiny. "The reality is, it’s a long year for everyone, and skillsets don’t turn into results unless they’re bringing the dedication day after day".
The Royals have been knocked for not stressing on-base percentage like more cutting-edge analytic teams like the Athletics and Red Sox. Williams mentioned how intangibles and defense can add value despite a poor OBA, such as the case with Salvador Perez. Tailoring the team to fit Kauffman Stadium has been a very deliberative effort, with the team focusing on athletic outfielders and a willingness to take on fly-ball pitchers.
They also explained that the market has overpriced on-base percentage.
Royals analytics guys point out that OBP is expensive. It doesn't age well and often comes at non-premium positions. #saberseminar— Sky Kalkman (@Sky_Kalkman) August 23, 2015
Even despite last year's run and the emphasis on character on defense, the pair admitted that the club had overachieved this year, according to their projections, and was probably a little bit lucky. The Morales signing has been a big home run for the club, but Williams admitted it was one of the few times when a player they targeted was willing to sign with Kansas City, and that Dayton Moore pushed very hard to land the slugger.
The Royals were remarkably healthy last year, and the pair seemed to suggest they had analytics attempting to predict injury risk based on the number of stressful plays or pitches a player endured. They also revealed that Lorenzo Cain has figured out when he needs to go 100% on a play, keeping him off the disabled list the last two seasons. Even then, injuries can be hard to predict, such as when Jason Vargas was lost for the season despite not showing up as much of an injury risk.