One of the things that gets mentioned when regarding "strength of schedule" is the combined opponents record. For example in 2012 our FBS opponents went a combined 87-66 (~ 57%), but 10 of those combined losses were due to losing to KSU (which seems silly to include when determining SOS). On the flip side you can imagine a team having a poor year (let's say 1-11 against FBS teams) and their opponents went a combined 94-60; this will inflate the SOS because 11 of those combined wins are due to playing the shitty team.
Back to 2012, if you remove the games against KSU in the opponents combined W/L record then it becomes: 87-66 (~61%) which I think is a more reasonable reflection of who KSU played. Applying this back to 1996 (I just picked when the Big 12 started) you get:
What I though was interesting was that the SOS was pretty consistent. For example, the 2004 disaster was not really a product of playing especially good teams, but just being a pile of poop. I then though you could imagine a "quality win %" which was the average of your win % with the adjusted opp win %. For example, in an extreme case if a team went 12-0 and its opponents went a combined 132-0 (remember we removed the loss due to playing the awesome team) its "quality win %" would be 100% If a team when 12-0, but the teams it played when a combined 0-132 its q win % would be 50%. For KSU the q win % would be:
It goes to show how consistently "average" KSU was between 2004 and 2010 and how good it was between 1996 and 2003 (with the exception of 2001 season that in retrospect was still better than any year between 2004 and 2010).
Now the next natural thing to do is extend this to other Big 12 teams. I will post the bar charts in a separate reply, but below is three plots with big12 (and former as well) teams:
It looks like the best teams were 1997 Nubb, 2000 OU, and 2005 UT (man that team was great).