December 28, 2004
Exclusive Birdhouse Interview with Mark Johnson
By Brian Walton
Several who read my interview series with Jeff Luhnow have asked about Mark Johnson,
who was prominently mentioned there. Johnson was Luhnow’s first hire upon joining the
Cardinals, and was a tremendous addition as the team’s Senior Baseball Analyst.
Johnson is a true talent, with a unique ability to meld the theoretical and practical. Even his
biography makes that point crystal clear. Mark earned a PhD in Applied and Computational
Mathematics from Princeton yet served as Bobby Knight’s student manager for the University of
Indiana basketball team. He is an ironman triathlete and a bike racer. Not your average technogeek.
I would have thought that if Johnson could handle four years of Knight, he was surely well-equipped to be able to deal with anything. Apparently not. Surprisingly, after less than a year, Johnson resigned from his Cardinals post, returning to California.
I caught up with Johnson over the holidays, giving him a break from unpacking moving boxes.
How long did you stay with the Cardinals and what were the factors behind your decision to leave?
I moved to St. Louis in January, 2004, intending to stay 2-3 years, but was fully prepared to stay longer if everything was lining up perfectly. If the experience continued to be as exciting as it was the day I was offered the job, and if I felt like I was having enough impact in the right places, I had no plans on leaving so soon.
There were many factors that led to my resignation, but I can go back to one of my original intentions on why I came in the first place to explain why the role was not ideal.
In the summer of 2003, I was asked to work in collaboration with my brother (Matt S. Johnson, an assistant professor of Statistics at CUNY's Baruch College) on some projects for the Commissioner's Initiative -- I had the pleasure of working with a committee of people put together by the Commissioner's Office (which included Rob Manfred and John Henry, among others). Immediately following that experience, I became driven by a desire to somehow integrate mathematics and sports. My vision quickly became a desire to bring practical, cutting-edge mathematical tools to sports decision makers, and to bring an exciting, real-world application to mathematics students. Links between the academic community and the sporting community are clearly underdeveloped today, and both communities stand to gain.
As a first step, I co-founded a consulting company, SportMetrika, with my brother and a business partner, Rob Beck, and one thing led to another. Almost overnight I was talking to Kevin Weiberg (Commissioner of the Big 12 and current Chairman of the Bowl Championship Series) about the BCS ranking system, speaking with the Commissioner's Office of the NFL about possible rule changes, was involved in an email flurry with Mark Cuban about the collection of statistics in the NBA, and was talking to one of the Directors of the NCAA Men's Basketball Tournament about ranking algorithms.
It was the Cardinals, however, that offered me a full-time position in St. Louis. I could not pass-up the opportunity to learn from the experts inside a front office and immediately put the consulting company on hold before becoming a member of the Cardinals Staff.
Once the amateur draft had passed and I started doing relatively focused research on professional player evaluation, it dawned upon me that the academician inside of me was more interested in sharing the results of my work like this with a larger audience than this role would allow.
More on this later.
You mentioned you are working on independent projects for later re-entry into baseball. What are the projects and how and when do you want to return?
I currently have several projects underway, some of which I can discuss more than others. I have been in discussions with several mathematics and statistics professors across the country toward organizing a baseball research consortium that would serve toward using baseball statistics as the foundation both for the purposes of educating undergraduates of the challenges, and rewards, of a life in mathematical modeling while studying a game that they love.
Additionally, some baseball observations can be used as another test-bed of data toward developing cutting-edge machine learning methodologies. In short, I want to bring "hard science" to baseball as an alternative to the "back-of-the-envelope" calculations used by most baseball fans and analysts today.
Several other projects are in their early stages, but we can save that conversation for when they are a bit more developed.
All of this being said, I would not be surprised if I eventually became affiliated with a ballclub again a couple years down the road.
How much of your leaving the Cardinals was due to these external opportunities and how much was due to how the job evolved?
My resignation was due 100% to how the job evolved. External opportunities only started to come into play only after one foot had already stepped out the door.
Has the team embraced Mitchel Lichtman's philosophies (Super Linear Weights, Ultimate Zone Rating, etc.)? How are they being deployed? Do you agree with them?
Over the course of my short tenure with the Cardinals, I spent a significant amount of my time talking to members of the advisory board about the various player evaluation techniques that had evolved in the baseball analytic world as well as in the domain of fantasy baseball. No two methods were the same, and I quickly learned that each analyst has their own methods and their own reasons why they think their methods are better than the others. This is fair, and is a fact of life in any application of statistical modeling -- it's rare that you will ever find two statisticians agreeing that a model captures perfectly the phenomena being studied. In fact, it is common that different modelers are asking slightly different questions for slightly different reasons.
In that light, I felt that although a lot of existing techniques are decent, nothing truly satisfied the scientist in me -- especially for the typs of questions we would be asking. I felt like there were more robust methods to be developed, and a much more rigorous approach to the modeling could be done. The advisory board, full of exceptional analytical baseball minds, helped me gain exposure to the wide array of methods out there and were very helpful for me to bounce ideas off of, but as for developing the holy grail of player evaluation metrics as it pertained to a baseball front office, I chose to talk instead to some professor friends of mine about cutting edge machine learning approaches. When it became clear to me that a research approach to the problem would not be supported as much as I would have liked, I knew it was time to branch out to solve the problems in a different setting.
How much of a factor are these statistics in the team’s decisions, for example not to keep Renteria or Womack or whether or not to pick up a free agent such as Polanco?
Unfortunately, I left the organization before any conversations about this off-season's moves had started. As you could imagine, however, there are many factors that are involved in making a final decision about when to sign, or not to sign, a player. Market demand, injury risk, budgeting, and team philosophy all play a role in making the final decision. Any statistical measure, whatever it may be, will only play a part of the final decision.
There are always a couple questions to ask when discussing player evaluations. How *good* a player is he? Does the market think that, too, and is it willing to pay for it? Are *we* willing to pay for it, given all of the other constraints? Unfortunately, a final decision is not at all as simple as looking at one number.
How much of the evaluation process time is spent on the majors versus the minors? Is that affected by lack of data on the minors?
You touch on a very interesting subject here. On the minds of a baseball organization are players from a wide range of levels. On the left side of the spectrum, you have Dominican teenagers and American high school students. On the right, you have the Albert Pujols' and Barry Bonds'.
On the left, there are a LOT of players, about which you know very little. On the right, there are FEW players, about which you know a LOT. Clearly, any mathematical analyses will be more effective on the right-end, since more data means richer conclusions. On the left end, you have little or no data, so such analyses don't get you so far.
But because you know very little about players on the left end, *any* sort of mathematical analyses is more likely to have a larger impact in terms of being able to find a "diamond-in-the-rough". On the right side, however, even having lots of data, and interesting conclusions, it is difficult for the "magic" of the numbers to outweigh an expert scout's (or fan's) perception.
The stock market makes for a decent (though far from perfect) analogy. If you were faced to make a decision between two big-named, well-known companies, you might look at all the numbers you want, but your final decision is more likely to come as a result of a "gut feeling" of where the economy, and the society, is heading.
But if you were asked to choose between 1,000 unknown (to you) companies, it's probably your best bet to look at some very simple sets of numbers for all of these, sort the companies by some metric, then pull the trigger and split your money over the top 5-10 of these, hoping that the upside outweighs the downside.
In baseball, you have the amateur draft, which is similar to trying to maximize your upside while minimizing your losses. In the free agent market, you have to pick between a few big, well-known, names to go after. There certainly are some similarities.
During my season with the Cardinals, I worked closely with John Mozeliak on the amateur draft, I worked with Bruce Manno with minor league player evaluation, and I spent time developing player evaluation systems for major league players for Walt Jocketty's sake, as well. My efforts spanned the spectrum, and were generally focused on a particular area at different points throughout the season, depending on what the next big deadline was. In the spring, the team is focused on setting the roster, then quickly moves into focusing on the amateur draft. The summer brings one's attention to the trade deadline while the late summer you start thinking about the minor league, then clearly the major league, free agent market and the Winter Meetings.
How is the Cardinals Advisory Board working out among the team, BaseballHQ and Lichtman? What changes might be anticipated for 2005?
As for the future of the advisory board, I really cannot speculate. I did often act as the liaison between them and the rest of the front office, so that gap will need to be absorbed by someone else if the relationship were to continue.
What is the level of acceptance of sabermetics within the Cards organization and is that improving, declining or unchanged?
The Cardinals organization is very open-minded about listening to new ideas, and to work towards incorporating them into their existing philosophies. Where they go now is not for me to speculate.
What was your biggest surprise and biggest disappointment while working for the team?
Hands down, the biggest surprise was how incredibly welcoming Walt Jocketty, Bill DeWitt, and the rest of the Cardinals Front Office, scouting staff and ownership group were to my arrival. I was expecting the worst, and I found a lot of very open-minded people willing to listen, and genuinely curious about learning new things. Never did I feel unwelcome.
Obviously, the biggest disappointment was not winning the World Series! Seriously, though, the biggest disappointment is that I am not affiliated in some fashion with the club right now. Several possible solutions toward keeping me affiliated had been discussed, but none were ideal for both sides, so I decided to move on.
What is the most interesting thing that happened behind the scenes during your time with the Cardinals?
There really isn't one specific memory that stands out -- my first season in baseball will always come with incredible memories, especially with the nearly perfect season that we were all able to enjoy.
But hands down, the opportunities I had to watch and talk baseball with some of the greats will be unforgettable. From watching a few innings with Bob Gibson this season and watching many, many innings sitting next to Red Schoendienst, I will never watch a baseball game the same way again. Learning the GM ropes from Walt Jocketty and Bob Gebhard; learning about a life in baseball from Bruce Manno, John Mozeliak, and Jerry Walker; learning about in-game management from Tony La Russa and Jim Leyland and catching a few innings with Stan Musial and Mark McGwire. How many hall-of-famers did I just mention? St. Louis has such an incredible history that I am very fortunate to have been a part of.