September 23, 2011
Adventures in Sabermetrics 101
Believe it or not, most of our writers didn't enter the world sporting an @baseballprospectus.com address; with a few exceptions, they started out somewhere else. In an effort to up your reading pleasure while tipping our caps to some of the most illuminating work being done elsewhere on the internet, we'll be yielding the stage once a week to the best and brightest baseball writers, researchers and thinkers from outside of the BP umbrella. If you'd like to nominate a guest contributor (including yourself), please drop us a line.
During the academic year, Andy Andres is a Senior Lecturer of Natural Sciences and Mathematics at Boston University, where he teaches biology, physics, and natural sciences. Andres otherwise pursues his baseball passions as instructor of one of the first ever college courses in Sabermetrics offered at Tufts University; as a Datacaster/Stringer for mlb.com at Fenway Park; as a Data Analyst for Ron Shandler’s BaseballHQ.com; and as the Head Coach and Senior Instructor for the MIT Science of Baseball Program.
I have been privileged enough to teach a course in Sabermetrics at Tufts College since 2004, when Dr. David Tybor, Morgan Melchiorre and I sat on my back porch in Cambridge MA and designed a course in baseball analysis that we would love to both take as students and teach to others. David Laurila has covered some of the history of this class previously on the pages of BP. Bottom line, Tufts has a very cool place for its undergraduates to explore some university courses that are outside the box, The Experimental College. Since 2004, both Tybor and Melchiore have gone on to bigger and better things (convincing smokin’ hot wives to marry them, great careers, toddlers and babies), leaving me behind to try to teach this course by myself, something I have never done as well as when the three of us did this together. Seven semesters’ worth of Tufts students have suffered through a university course on baseball and sabermetrics—I know, I know, we all have our burdens to bear.
There has been an interesting shift in the students taking the class over the years, which reflects the popularity of the study of sabermetrics. In the Fall of 2004, the group of 20 students hardly had any background in sabermetrics; most had not seen the word. Only some had heard of Bill James, and fewer had ever read any of his work. Moneyball was a new book for the vast majority of them. And when we covered the early work of Phil Birnbaum and Jay Bennett on Game State Transitions, Win Expectancy, and Win Probability Added, the students had never seen anything like this before—they acted as if we had handed them the keys to the baseball analysis kingdom.
Now Moneyball is a hit movie starring Brad Pitt, OPS is reported widely on TV, radio and print, and some students entering my class know as much about sabermetrics as I do. Bill James is now an iconic figure in many students’ eyes, Baseball Prospectus and FanGraphs are well known my students, and “Wins Above Replacement” is thrown around in early class discussions as if it were invented by Henry Chadwick, Peter Gammons, or FC Lane. It is clearly a different world in Sabermetrics 101 at Tufts than it was seven years ago.
Because of the shift in the knowledge of the students entering the course since 2004, I have had to change how the class is taught. Currently I place more of an emphasis on original research, thinking empirically from the start of the semester much earlier than in 2004. But one holdover from the beginning has been the exercise of predicting the future major awards in baseball. This exercise allows students to demonstrate their skills in rhetoric, data analysis, and sabermetrics. In addition, it allows them to look at various projections systems in baseball and determine how much they trust them!
So in late January this year I asked my students to predict the 2011 AL and NL MVP. We had not really covered the sabermetrics of hitting yet, so this was about seeing how far students would take the relatively open-ended assignment. I gave no minimum length of pages or words, just said they had to make and support their arguments.
Some students did further research about park effects that would help Gonzalez, mentioning his fly ball tendencies to left field leading to increased doubles and home runs (not nearly as well done as Jeremy Lundblad’s multiple looks at this). I was hoping for this kind of further analysis—this got the students top grades on this assignment!
Two other AL players had more than one Sabermetrics 101 student arguing for their 2011 AL MVP candidacy:
But there were student arguments for NL MVP, and the only player to get more than one argument was Albert “The Machine” Pujols. Among the real contenders for the 2011 award, recent voting history probably kept them off my student’s radar:
Among the single “votes” in my class for NL MVP were Carlos Gonzalez, Ryan Howard (my Phillies fan), and Hanley Ramirez (my Marlins fan). The arguments in my 2010 class were understandably centered on Pujols. Students noted his consistent and persistent greatness—in nine out of the 10 years he has been eligible for the award, he never finished lower than 4th in the NL MVP race (a very interesting analysis would be to better understand what happened to the writers who voted in 2007—how could Pujols finish 9th with an OPS of .997?), won it three times, and was second 4 times.
Student arguments for MVP were written before the finish of the Pujols pre-season contract deadlines, but his contract situation led the class to analyze performance of players in contract years. As expected, students started out arguing that there was some extra motivation and effort in a player’s walk year, leading to better performance, and therefore more money. They were convinced that this solidified their Pujols-for-MVP argument—since Pujols was about to walk, he would be even more motivated and try harder than usual. Note that my Tufts students had just lived through another convincing season of the “Poster Child for the Walk Year Phenomena,” the 2010 of Adrian Beltre. But through the work of BP and others, my budding Tufts sabermagicians now see the contract walk year effect for what it is, a small effect if any at all.
We also ask our students to predict the AL and NL Cy Young Award winners and AL and NL Gold Glove Award winners. These exercises help to frame discussion of the sabermetrics of pitching and fielding, a topic that has certainly improved since the start of the class in 2004.
As you might imagine, it is fun teaching this class—discussions can get rowdy, passions run high when students argue for their teams and favorite players, and students seem to really enjoy the course. Our guest speakers have especially privileged us over the years, many of whom have been Baseball Prospectus alumni or writing currently here (Jay Jaffe, Joe Sheehan, Jonah Keri, David Laurila). Guests within baseball have included Brian O’Halloran, Bill James, Zack Scott, and Tom Tippett of Red Sox Baseball Operations. Dwight Evans entertained us for many hours with stories of his time playing and coaching baseball. Others who have visited include Pro Scout Joe Bohringer and former Scouting Coordinator Helen Zelman of the Diamondbacks (Helen was a student in our class also!).
Through the years, many legit sabermetricians have been extremely helpful in consulting with students and helping teach the class: a huge baseball analysis shout out to Pat Andriola of FanGraphs (former Sabermetrics 101 student now at Duke Law School), Peter Bendix of the Tampa Bay Rays (former student), Jeremy Greenhouse of Baseball Prospectus and Bloomberg Sports (a former auditing student who really should be teaching the class!), Daniel Rathman (current BP intern), and Tufts Alum Dr. Dan Brooks of BrooksBaseball.net.