2018 - A New Agenda

Time to start over again. My team-building approach last year (building a fantasy baseball WAR that was adaptable to conditions where I didn’t have empirical weekly results) was successful in that my team made the playoffs despite some terrible luck from injuries.

My blogging approach was a complete failure. I let worrying about having my ideas be “done” block me from putting my thoughts out there.

I’m going to endeavor to write more often and more consistently this year, even if it’s just to keep notes about my avenues of research and software learning. Here’s what I’m hoping to accomplish this winter before Opening Day:

• Generate “synthetic weeks” (simulate a player’s performance many times over a weekly span using a set of annual projections) so that I can calculate the weekly team H2H mean and standard deviation for any plausible league configuration. I’ll be using Python tools to do this, having had success with SciPy and some of the distributions available in PyMC3. Ruby’s existing tools just aren’t fleshed out enough to do the things I want to do with statistical distributions.
• Improve my data crosswalk so that it’s easier to link data from different sources.
• Start making headway on risk from injury or a collapse in performance into my player pricing model. I thought this Saber Seminar talk might be an interesting place to start.