Teaching Sabermetrics

Earlier this year a small Bay Area nonprofit called Tutorpedia asked me to teach a poetry workshop over the summer. Actually, they asked to write a curriculum for a poetry workshop, and then to teach it later, which may seem like a slight distinction, but is not. Seeing as I clearly didn't have enough to do already – I was also in the midst of a Master's in Education program at Stanford – I petitioned to write and teach not only a poetry workshop, but a sabermetrics workshop as well.

To my surprise, Tutorpedia not only said yes to my nefarious, baseball-teaching plan, they were thrilled.  They even offered me free Oakland Athletics tickets so I could go to a game with my students.  And so I started crafting my curricula in my 'spare time.'

[caption id="attachment_3109" align="alignleft" width="269" caption="Who doesn't love the Oakland A's?"][/caption]

Tutorpedia believes in project-based learning, and I believe in dialogue-driven classrooms, so my model was not, from the outset, stand up and lecture. Rather, the poetry curriculum was to be a series of conversations about famous poems, followed by activities highlighting techniques the poets in question were using, all culminating in the composition (and public recitation) of an original work of poetry. Likewise, the sabermetrics course was designed to give students the skills they needed to perform statistical analysis before sending them off to do independent research projects.

Unfortunately, after carefully putting together my curricula, I found out that no one was signing up for my workshops. And not only me, Tutorpedia was hoping to run dozens of workshops over the summer, and no one was signing up for any of them. Ultimately, three students signed up for my sabermetrics class, and none for my poetry workshop. So be it! If high schoolers would rather learn about wOBA and WAR than enjambment, simile, and pentameter, who am I to argue?

One of my three students ended up having to drop the course due to a family emergency, leaving us at a paltry two – and a brother and sister (and Yankees fans) at that – but nevertheless we trod onwards towards understanding and insight, or at least indoctrination into a different set of beliefs than those you get from watching the Fox Game of The Week. I say that in jest, partially, but it's also a question. Is it possible to really get students to learn to think critically about anything – even something as trivial as baseball statistics – in just 16 hours of class time? And isn't critical thinking what sabermetrics is supposed to be about? That's the real can of worms, I suppose. As sabermetrics have grown more prominent, it's not clear what the study is about anymore, let alone what teaching the subject means.

Certainly the origins of sabermetrics are in critical thought. At a time when no one questioned the power of batting average and the RBI, advocating statistics like OBP or, even worse, creating new ones like Win Shares, was the kind of heresy that required a great deal of level-headed, penetrating insight and at least a little gusto to boot.  Enter Bill James. These days, however, we have Fangraphs, Baseball Prospectus, the Hardball Times, Baseball Analysts, Baseball-Reference, and countless other websites devoted to providing information and analysis about baseball statistics. The old heresy is the new religion.

Which is not to say there's anything wrong with sabermetrics – I wouldn't teach it if I were so unfaithful – but rather it raises the question: which is more important, the thought process behind the study of baseball statistics, or the outcome of that process? Which is more important, knowing how WAR is calculated (as much as they let you know, anyway) or knowing more or less how to use it to evaluate players?

[caption id="" align="alignright" width="413" caption="Another object of study: Win Probability graphs"][/caption]

I, for one, am persuaded that WAR is a better measure of a player's value than almost any other single statistic, and WAR formed the basis of the research project – building an all-time Yankees roster – my students and I did at the end of the course. That said, my inclination as both an educator and as a baseball fan is and always will be towards process over outcome. Without the ability to think through a statistic, the statistic has no meaning to me other than the one that other people give me.

How to convey that in eight two-hour sessions? Well, it's hard. I taught my students how to use excel and basic statistical tools like regression and correlation (which, to their credit more than mine, they understood), and I exposed them to a few of the many sabermetric websites. And in the end I saw that hard work rewarded by rapid – almost immediate – buy-in when we learned about WAR and UZR and ERA+. It remains to be seen whether my students will ever bring the tools of the skeptic's trade - critical questions, excel speadsheets, and statistical techniques – to bear on the new numbers as we did on the old.  I hope I at least planted a skeptical seed, but the narrative of modern baseball statistics, it turns out, is quite compelling.

In retrospect, the course was perhaps a bit too practical. Maybe the better path towards critical thinking runs through a more philosophical approach. Maybe stepping even further back, teaching less and asking the students to explore more on their own, would have netted a more skeptical attitude. But, taking that approach, what do you do if the students come out the other end convinced that pitcher wins and runs batted in are awesome?  What if being hands-off means letting someone else do the brainwashing?

I suppose the answer, on some level, comes down to the students. Because I was working with a brother and sister who were at least familiar with the existence of Moneyball, I knew they had the potential to dive into concepts like WAR, wOBA, and UZR, but when our initial discussion turned up batting average as THE statistic of choice, I also knew we had some work to do before we got there. Considering the slightly younger-than-anticipated enrollment (I expected high schoolers and got middle schoolers), we had to begin at the beginning, building regression tables to show that, in fact, OBP and SLG correlate way better with winning than batting average or homers.

As I dove into these analytical and pedagogical depths, considering and reconsidering both the statistics I was teaching and my approach in teaching them, I was called back to another world as we were working on the final project. The sister wrote a limerick on the board in a moment of whimsy, likely tired of comparing Red Ruffing's ERA+ to that of Roger Clemens.

"How do you like my poem?” She asked me.

“That was supposed to be the other workshop,” I told her.