Happy NFL Kickoff! In Part I of this series, I presented of few personal theories of mine about what makes football teams successful, and then analyzed these theories in Part II using a linear regression on statistics relevant to each theory. Here in Part III, I’ll use the information obtained to introduce this year’s format for “Crunching The Numbers.”
If you’re new to this site (welcome!), Crunching The Numbers is a project I undertake each season in which I use statistics to “grade” each NFL team. It began as a pet project for me in 2011, inspired by the 2009 season, and I started posting it here in the FanPosts in 2012. It was actually how I originally got promoted to contributor at Bleeding Green Nation by former manager JasonB.
The idea behind it was simple. I would analyze some statistics, combine the relevant ones into a formula that made mathematical sense, and then determine a “score” for each team every week. The ultimate goal was to develop a formula that could separate “contenders” from early-season “pretenders” as quickly as possible. Each year I would attempt to tweak the formula to make the predictive model more accurate.
Well, now I’m here to tell you that I won’t be doing that anymore.
After I graduated college, I was sure I would have more free time to work on this since I wouldn’t have the endless activities, exams, and classes. As it turned out, when you get a job and start making money “free time” magically goes out the window. With everything going on I can’t commit the time each week to develop a score for every team. So I’m dropping the “power rankings” format for Crunching The Numbers. Honestly we get enough of those anyway.
Introducing the new “streamlined” Crunching The Numbers
A common feature in my previous versions of this series was I would use the statistics analyzed to preview the Eagles’ upcoming opponent. After a while I noticed that my “innocuous” predictions - look for the Eagles to run the ball early, this game will come down to defending the intermediate passing game, things like that - often held up during the game. Of course, I wasn’t prognosticating wins and losses or anything like that, but in general my subtler “keys to the game” carried more merit than I expected.
This is the direction I’ll be taking Crunching The Numbers this year. Starting in Week 5 (when I feel I’ll have a good sample size), I’ll compare the Eagles to their next opponent using statistics from Part II of this series that were significant. Based on how the teams stack up, I’ll offer some small details to look for beyond “turnovers will be the deciding factor in this one” and “whoever has the ball last will probably win.” I’m hoping to be able to make statements like, “The Eagles should have success on defense if they play Jenkins in the box” or “I’d expect Sproles to be a bigger factor in this game than Smallwood.”
This method for previewing games is a little different than James Keane’s slick simulations or a general “players to watch” post. I’m more going for the big picture here, almost from a coaching perspective. What formations will we see on both sides of the ball? Who will get more snaps? Will the defense be better off in man or zone coverage? In essence I’ll be using the metrics to distill a rough “game plan” that the Eagles should follow.
Okay, so that’s all well and good. But what statistics will I be using to make these claims? Before I even looked at a metric this offseason, I decided I would only use ones that the analysis deemed “significant” (see Part II for more information on this). It doesn’t make sense to do an analysis if you’re going to arbitrarily ignore the results you don’t like. This ended up having interesting side effects, as I will not be looking at anything related to the running game or pass rush here. I look forward to seeing how this affects my thoughts when previewing each matchup.
The statistics that made the cut were (in no particular order):
- Yards per Pass Attempt (YPA)
- Opponent Yards per Completion (Y/CMP)
- Time of Possession (TOP)
- First Half Points (PTS/1HLF)
- Opponent Yards per Point (Y/PT)
These five statistics will form the backbone of my weekly game previews. As much as it bothers me from the outset to ignore large facets of the game (namely the running game and pass rush), I’m willing to trust the data here. I’m also hoping that some of the “forgotten” aspects will be buried in other metrics I am using by virtue of being collinear. For example, TOP also works to assess the running game by extension since running the football is the most effective means of clock control.
Of course, in a blog the most important opinions are in the comments. What do you think of this? Are you intrigued by looking for something in a game you might not have thought of before? Or will this project here be all bark and no bite? Who knows! We’ll find out for sure Week 5.
Football is back! GO EAGLES!