Me again... well for the last month while I've been on a forced employment renaissance (aka laid off) I've picked back up on my Fantasy Football Analytics. If you remember, the last time we took this on, we began looking at the problem (how to win at fantasy football better than anyone else) I put together a concept that blended Computer Vision with Advanced Statistics. Here's the proposition in a nutshell: Can we create a programmatic method of analyzing football video to give us an edge in fantasy football play (betting)?
The answer is complicated, and I set out to describe the Epic (An epic captures a large body of work. It is essentially is a large user story that can be broken down into a number of smaller stories. It may take several sprints to complete an epic. An epic can span more than one project, if multiple projects are included in the board to which the epic belongs.) The epic looks somewhat like this:
As I work on this effort, I'm drilling down on each story (represented as a cloud). For example, the first cloud directly in front of our thinker (salmon shade), is related to the subject area of "Stats". Its format isn't unstructured, quite the contrary, its very structured. Each team (NFL or Collegiate) will be represented in the Master, their entire history, singular and most up to date, represented in some structured tables. I'll need to continue doing my refinement, as I'm no expert in NFL structures, however the idea is, to have some (if not all) stats on anyone who is or has played the game. For my concept in analysis, the content of the stats is very essence of what we're after, a numerical representation of the player, team, and the plays. But it doesn't end there, I plan on capturing, scraping, crawling every source of sports info out there, and by the time I'm done, we'll have a huge storage of both structured and unstructured information. When you're looking for an edge, you look under every rock, leave no stone unturned.
More to follow... blog on!