Well, that was an eventful two days at the 2013 MIT SSAC.  I blogged about this repeatedly in previous posts.  As a teacher of statistics, what lessons will I take home from this event? Where do I go from here? 

• Take risks to connect with people. I talked up two presenters. First, Tim Chou He works in Los Angeles as a project manager for an engineering firm. He’s also a high-school football coach. I asked him if he’d be interested in being a resource for my students in “Statistics, Sports, and School.”  Before I could get out the first sentence, he said, “I’d love to be part of this.”  We then talked for about 90 minutes about his work and the new course. We made plans to work together and find ways for him to help our students. The second one was Nate Silver, who was getting assaulted by fanboys and gushers at every turn. I kept it class, shook his hand, and handed him my card. It helped that someone noticed where I was from and said “GREAT school.” I asked him if he’d consider spending the day with our students, and he was very receptive. So two GREAT contacts to follow through on.
• You don’t need to be a genius from MIT to ask and answer meaningful questions  in sports analytics. Tim Chou’s experience coaching football was probably the most important in his efforts to find some metrics in football that caught the eyes of the professionals in the field.    The researchers and statisticians repeatedly discussed that GM’s, Coaches, and owners catalyzed research questions.
• My students can do this, and I can help them.  I was initially worried that I would need to become an expert in sports analytics in order to be helpful to my students. But I am better understanding how I can be helpful in moving them forward through the process. Here’s what I can prepare for:

1. Collect and organize as many resources students could use to “play in the garden” of sports and data. Good people, Good databases, good papers, good books, good use of our school LMS (learning management system) to have them collect their  own resources /thoughts,
2. Prepare activities around common issues working with data: Here’s what’s bubbling up: Misconceptions about randomness, how the short term “feels” in the midst of long-term predictions, displaying data so they tell a correct, important story,  Differences between two players / teams could easily be due to chance. Definitely this one: Inputs and outputs of mathematical models need to be clearly defined. Use units that make sense to the math-averse. In nearly every good talk, the presenters made sure that variables were expressed as “Points per game,”  “additional yards per quarter,” “Points above what a replacement player would contribute,” etc.  The best presenters took the time to walk through a simple example so folks could “click in” to what they were talking about.
3. I need to put my “student cap” on, and ask my own questions this spring and summer. I want to know what’s possible for students to do next year.  Can I do my own sports analytics project?  I need to brainstorm a bunch of questions, and zero in on ones that get me jazzed. I need to give myself a deadline, answer them to the best of my ability, and share my results (probably on this blog).

This was a blast, and I highly recommend it. Soon, the real tough work will begin….