I despise the QB Rating as a statistic.  The minimum and maximum contribution for each of the attributes seems somewhat random.  The fact that there can be a maximum is just sort of stupid, especially since that maximum isn't "100", but rather "158.3".  

Let's break it down, shall we?

Here are the components that go into the QB rating (wikipedia link: http://en.wikipedia.org/wiki/QB_rating ):

Completion %
Yds/att.
TD %
INT %

Each statistic is normalized so there is a minimum and maximum possible contribution.  Here are the mins/maxes:

30%/77.5% comp %
0/12.5 yd/att.
0/11.875%
0/9.5%

Is there anything intuitive about any of those numbers?  I mean, someone that completed 90% of their passes for a game/season clearly played better than someone who completed 77.5%.  Likewise, 15 yd/att is better than 12.5.  15 TDs in 100 attempts is better than 11 and 15 INTs in 100 is worse than 9(.5).

The statistic sucks because there are fictitious limits.  Why don't they just make it one big continuum?  Who the hell came up with this?  

All of that being said, I still think it is a useful stat when talking about someone's career or season, but game to game numbers don't really need this stat.  Just put the damn line up and let people figure it out themselves.  I'm absolutely bored to death reading about the 0.0, 1.3, and 10.2 games Grossman had last year.  The chicagobears website keeps listing those in article after article and it means absolutely nothing.  You'll have a terrible line if you throw 2+ INTs and have no TD passes.  It's that simple.  If receivers drop the passes you actually get to them it will be even worse.  Big deal.  Oh, and listing the QB rating for Griese's 7 passes last week is beyond asinine.  http://www.chicagobears.com/news/NewsStory.asp?story_id=3697