I've Got A New Stat For You

Smart looking statistics image

I have a somewhat statistical mindset* so I am always convinced that there is some magic number, or set of numbers, out there that can tell you if players or teams are performing up to their potential. I am not sure that this stat can exist in any sport except baseball (although I do think +/- gets close on a player level in basketball and hockey), but it will continue to be a lame dream of mine to discover this silver bullet stat that tells you where everyone stands, and have it reach mainstream status in the football world.

*As proof of this statistical mindset..I will admit that the sheer amount of boredom produced by this post will drive our traffic down 30%, and reduce my chances of acquiring a wastedoptimism sex-groupie to 98%.

Well in my quest for stat perfection, I have discovered a new statistic that I would like to share with the world. (At least I think it is new...if it isn't then I am done as this took years for me to formulate...well not really years... but at least an hour. If it isn't new will someone let me know so I'm not looking like so much of an asshole here wearing my letter jacket, brushing back my hair, and leaning on what I think is my new statistic? Thanks.)

Click the jump for the stat (you may want to put a condom on first though... it is that exciting)

So here is a description of the stat. Essentially, I am putting a lot of faith in Vegas to determine how winnable any particular game is, then breaking down a teams record based on how they perform in games that Vegas thinks they should win, toss-up games, and games they should lose. If I were to call it something it would be: "Actual vs Expected Performance" or something less shitty. Basically what it tells you is if a particular team takes care of business when it should, how they handle the pressure of playing an evenly-matched team, and if they are capable of playing above their heads when matched up with a better team. I break this down as a team's record in games where they were favored by more than 5, games that were essentially toss-ups (line of 5 or less either way), and games where they were more than a 5 point underdog.

As an example, here are the Big XII teams records in all three categories from the 2009 season (note: I did not include non D-1 games, or bowl games):

Click here for all teams

Now I realize that you cannot possibly determine how great or terrible a team or coach performs in all these situations based on one season worth of data... however, if taken over a coach's entire span at one school, it could point out that coach's strengths and weaknesses. In this case though, I could only pull one season from vegasinsider.com so we will have to work with that as an example... deal with it.

First example: Texas Tech was a team that was capable of playing with much better teams as they went 2-1 as dogs, however they just could not get fired up for evenly matched games going 0-2 in those games. So what does this mean? You could take it that Tech had the talent to be much better than they were last year if they would have put as much effort into gameplanning games where they weren't outmatched, and as such, as a reflection on the coaches not being consistent week to week.

Second example: O-State was on the other end of the spectrum as they, for the most part, took care of business when they should, performed very well against equally matched teams, and lost to the teams that were better than them. This could be taken as most weeks they played at a level close to their max potential and they were unable to find an extra gear in those games against superior teams...which I think is a pretty accurate summary of the 2009 Cowboys.

So what do you think of this new stat? Pretty exciting no? I still need to figure out a way to 1)get a bunch of historical data, and 2)establish the thresholds at which each record means something. But aside from me not having the data, and even if I did I have no idea what that data means...I think this is a promising new statistic. Once I refine this, look for "Actual vs Expected Performance" (or something less shitty) to start being reported everywhere.

(Usually this would have just been a drunken conversation next weekend, but since I have a blog my absolutely amazing ideas now go in here...you're welcome.)