Josh Stein, Assistant General Manager of the San Diego Padres started with the team as a law student in 2003. A graduate of La Jolla High School and UC Berkeley, the long-tenured member of the Padres front office has worked in many roles.
I had the awesome opportunity to meet with Josh last week at Petco Park. We chatted for about half an hour below a framed, white 1984 Tony Gwynn jersey hanging on his office wall. He talked about how the Padres combine statistics and scouting information in their decision making, how analytics have changed the sport, and the state of public sabermetrics.
Could you talk about your current role with the team?
Josh Stein: I’m one of the assistant general managers. What that entails is helping out AJ Preller with the major league roster, the 40 man roster, contracts, the decisions on who we’re going to trade, who we’re going to promote, and how we’re going to allocate our resources.
I then get involved in a lot of our systems that we build from a baseball operations standpoint. So our scouting system, our player development system, our information hub that houses all biographical information, all statistics and data, all medical information, scouting reports, player development plans and things of that nature. I work a lot on technology and trying to find data that potentially could give us a competitive advantage if we use it, analyze it, and ultimately implement it in our evaluations, preparation, advanced scouting, and so on.
It depends on the time of year what I’m specifically focused on. We’re heading into the offseason now and it becomes a lot more about the current roster, where we need to improve it, and what the plans are heading into next season. There’s a salary arbitration component to my job as well that takes up some time. Soon, we’ll look up and it will be February and we’ll be into the amateur scouting season, spring training, and supporting our amateur scouting staff. It really is a season by season type of job.
You mentioned a database of player information, do you build the database, contribute information to the database, or both?
JS: We built it internally a long time ago, about 2005. We realized early on we were going to a bunch of sources and websites to gather information on players. We would go to one website to get his contractual information, one site for his statistics, another piece of software for his scouting reports, etc. It was inefficient, so we built our own system for accessing this information. We have a good group here that works on a lot of that technology and integration of new information into this system.
I remember Moneyball talked about the A’s digging through binders of other teams’ players trying to look up information in a hurry. Seems like a system like yours would help with efficiency there.
JS: That’s exactly the way it works. At the end of the day player evaluation is about the best information. That could be the best scouting, the best statistical information, or the best makeup or medical information. The best information is going to be what drives good decision making, but how you access and share that information just from a pragmatic standpoint is also important. That’s how you have the ability for different scouts, front office personnel, and player development people to align in working towards common goals.
I understand you’re an analytics guy. Actually that’s not the best way to phrase what I’m getting at…
JS: That’s OK, I’m used to being pigeon-holed!
Yeah, that’s what I was trying to avoid.
JS: Well definitely when I evaluate a player I am looking at certain statistical measures that I know can help in that evaluation.
Ok. So like you said I don’t want to pigeon-hole anybody, but it seems that scouting is traditionally the other side of that type of evaluation. It sounds like your database is about integrating those two things to create something cohesive.
Player evaluation ultimately is a decision. Like any decision you make, you’re going to have a number of pieces of information to take into account. For instance if you were looking to buy a house, you’re going to walk into that house and have an immediate feeling about it. “Man, this feels exactly what I want.” It might be open in the right places, private in the right places, a great backyard, etc. You’re also going to want to go out and study things like what other houses in the neighborhood sold for, how close you would be to a great public school, what the neighbors are like, how much noise will there be, etc. Some of those pieces of information will be easier to explain with words based on an experience and some will be easier to quantify with numbers.
I don’t look at evaluating players as being all that different. You’re trying to get as much good information on that player as you can. Certainly your scouts are providing vital information. And if you are looking at the right statistics, those are going to provide vital information as well. We’re trying to marry all that information into one accessible place. We built the system originally for Kevin Towers, later Jed Hoyer, Josh Byrnes, and now AJ Preller, so that when we sit down to make an evaluation about players on our team or other teams, everything we have is at our fingertips.
Do you work with scouts much directly?
JS: I work directly with our scouts. A little about my background, I spent my first 4 years full time with the Padres (I was an intern for 3 summers) as an advance scout, and I did that primarily through video. But I supplemented that with talking with our pro scouts who had seen the teams I was advancing. I would constantly be in contact with them saying, “Hey, here’s what I’m seeing with Player X’s swing. I just read your report looks like we’re either on the same page or we might have a bit of a difference of opinion. Lets talk through this.” I don’t want to base what I’m taking to the coaching staff on just the data, I also want to base it on what the scout saw last week. So from an early point in my career I was talking to scouts to get as much information from them as I could.
And then now with some changes in the front office, I’ve been coordinating our pro scouting efforts for the past couple months. Just trying to make sure that as we got into the end of the minor league season and the playoffs, the end of the major league season, and as we roll into the instructional league, the Arizona Fall League, and winter ball, that we’re aligned with the people in the right ballparks, seeing prospects we need to see.
I don’t know how much you’d be able to say about stuff outside the organization, but it seems more and more teams are embracing data and analytics as the years go on. How would you say that’s changed the game?
Well in some ways it has diminished the ability to gain a competitive advantage just by doing analysis. I would say that a lot of the advantages–certaintly not all–but a lot of the insights that were gleaned from studying basic game data and looking at statistical information, some of those inefficiencies in the player market are now non-existent. The one everyone talks about is on-base percentage. Back in 2003 when I was first getting started in baseball, that was something that the Oakland A’s and a number of teams identified as being vital to run scoring, which they focused on to get a competitive advantage over a number of teams. That doesn’t exist anymore. Everyone understands the relationship between on-base percentage and run scoring.
It almost seems like it’s swung the other direction. For example a number of years ago the A’s traded for Kevin Kouzmanoff. His OBP was fairly low at the time, it didn’t seem like a traditional move Billy Beane would make.
Yeah. One of the things to realize is that any metric needs to be taken in context. Kevin Kouzmanoff as a .300 on-base guy in an on-base universe where the major league average is .330, that’s a real low on-base. But if Kevin Kouzmanoff can still maintain that .300 on base where the on-base universe moves back down to a .310, suddenly he’s not as far off the league average. And if he still hits for power he can be a more attractive asset.
The nature of the game has certainly changed in the last 10 full years that relates to hitters’ abilities. We’re seeing more strikeouts, lower averages, fewer runs, and lower on-base percentages.
At this point it might be more of a matter of finding those edges by looking in other places.
You’re always trying to get edges on the competition. It’s a very competitive landscape. Certainly some of it can be in the analysis itself, I think a lot of it frankly is in the implementation of analysis, the scouting process, or the player development philosophy. That’s not new school or old school. Frankly that’s just business–having philosophies or beliefs and being systematic in how you apply them across an entire organization.
Would you say some of that implementation involves being on the same page as on-field coaches?
I think that’s crucial, yeah.
Do you have any thoughts on the state of public sabermetrics in general? What’s interesting to you?
It’s always improving. It’s always getting better. I saw a quote the other day from the an NBA team that said “I treat blogs as my unpaid interns.”
You can see it in team hiring practices as well. Teams hire people who are writing for websites because they’re creative thinkers and are putting ideas out there that are changing the way people think about the game. Now, there is still a lot of data that’s not publicly accessible, obviously the scouting reports, but also camera and radar-based game data that is helpful in predicting what’s going to happen next. That’s not to say if public sabermetricians had access to the same data they would be significantly behind in any type of analysis.
Last question, for some reason this has become my thing. What’s your favorite beer?
I do like Sculpin. That’s an easy one, a local one.
That’s mine too!
Really yours? Solid.
I walked out of Josh’s office at the end of the meeting in a bit of a haze after such a fun conversation. Huge thanks to Josh for taking time out of his schedule to meet with me.