Statistics revisited

In response to my post yesterday Code Red (who loves his advanced metrics) pointed me to footballoutsiders.com, a website that attempts to satisfy my need for better statistics in football. It's not a bad website for statistics, and the DVOA and DYAR are good parameters for running backs. The defensive and offensive line statistics are relatively sound too, but there is still the all-encompassing problem of separating line play and running back play from one another. I think the running back statistics are pretty well formed. The aforementioned Frank Gore sits at the right spot.

However, the statistics at that website don't work for wide receivers and only semi-work for quarterbacks. To me, a quarterback makes a wide receiver look good. You can have Randy Moss and Andre Johnson on your team, but if Jonathan Quinn is your starting QB their stats are going to reflect that. Because of this, wide receiver statistics are consistently misleading. Even on footballoutsiders they make note that their catch rate statistic only involves the percentage of catches made when that receiver has been thrown to, and does not only include drops. So essentially, If Cutler throws a pass over Johnny Knox's head, that counts against him, which to me makes their wide receiver statistics largely irrelevant.

I use the example of Johnny Knox because his catch rate is currently a 63%, even though I only remember him dropping one pass. It also doesn't factor in wide receivers who aren't given many opportunities. Another Chicago example would be Devin Hester. He's made some truly great catches this year and has done everything very well, but he hasn't had many opportunities to show off his abilities yet, making it difficult to rank him compared to other receivers. Also, if dropped passes aren't factored in, I can't truly get a number on how poorly Greg Olsen has been playing.

Also, there should be a way to account for the "one bad game". Cutler had a terrible game against Green Bay, but let's say that was the exception, not the rule (so I can keep my sanity). How do we account for that? If by the end of the year that was his one bad game, should those statistics be weighted less?

This is just food for thought. Honestly I think football is too complicated a game to get very many true statistical measures. Basically, I encourage people to watch the football games and then make an assessment of a player based on what you saw. Statistics can often be hollow and misleading in the great sport of football.