Support my attention-whoring ways by following us on twitter!

Get the SKOdcast imported directly into your brain!

Wednesday, September 23, 2009

The problem with football statistics.

On last week's FOX NFL Sunday pregame show Howie Long happened to say that statistics were for baseball fans. Initially my response was something like, "What an old school meathead" but then I started to think. Every time I get into an argument with Red about a player there always seems to be an argument about how "Good" a player's game actually was. For example:

Frank Gore
Frank Gore rushed for a little over 200 yards on Sunday, which is an astronomical game when you simply look at that number. Even the more respected statistic Yards Per Carry is incredibly high at something like 12 ypc. But almost all his yards came on two runs of 80 yards each. Can we factor out those two long runs? Or is that just part of the game and doesn't diminish his incredible stats? We can't really answer that question, which is the problem with stats in football: there are too many factors to consider.

Baseball statistics are successful in grading a player because of the linearity of the game. A player either gets a hit, walks, or is out. He is either caught stealing or steals a base. Even errors are easily factored in. In football too many things can happen to "screw up" somebody's stats.

For quarterbacks stats get screwed up because of dropped passes, tipped balls, interceptions that weren't their fault, and freak occurrences. Not to mention they can be screwed up in good ways for the QB; a receiver catches a 2 yard pass and runs 70 yards, a pass is almost intercepted but is tipped and caught by a receiver for a TD, etc. For running backs like Gore it's even more difficult to find a relevant statistic. Is it the RB or is it his offensive line doing the work? Should we count it against them if they have a terrible line?

Baseball statistics have matured with newer stats like OBP, WHIP, and Slugging pct. The "relevant" football statistics haven't changed for years. Simply measuring a football player by how many yards they gain, tackles they make, and touchdowns they score will not prove whether or not they are great. I'm not saying I can, because if I could I would have, but football needs new kinds of relevant statistics. Football stats are moving in the right direction by measuring things that aren't readily obvious, like YAC (yards after contact) and "bags" (holding calls that a defensive player forces), but I'd like to see more combined statistics.

As it stands, football statistics simply create argument about their legitimacy. I'd much rather be able to argue about whether Forte is better than Frank Gore using legitimate statistics than having to argue about whether or not Gore's stats are a fluke.

No comments: