Here are some examples from the Big12 where the best teams didn't play eachother for the championship game because of geography. (Note, if you don't think the point of a championship game is for the two best teams to play, this doesn't matter because we have fundamental differences in what we think a CCG should be.)
1996: 6-2 UT played in CCG instead of 7-1 Colorado despite losing to Colorado. Also got in over a 6-2 KSU team they didn't play.
1997: 6-2 Texas A&M played in CCG instead of 7-1 KSU. KSU beat A&M in regular season.
1999: 6-2 Texas played in CCG instead of 7-1 KSU. KSU beat Texas in regular season.
2000: 6-2 KSU plays in CCG instead of 7-1 Texas. KSU/Texas didn't play eachother, however Texas beat A&M and A&M beat KSU.
2003: 6-2 KSU plays in CCG instead if 7-1 Texas. Texas beat KSU in regular season.
2004: 4-4 Colorado plays in CCG instead of 7-1 Texas, 5-3 Texas Tech, 5-3 Texas A&M and 4-4 OSU. Colorado lost in regular season to Texas, Texad A&M, and OSU.
2005: 5-3 Colorado plays in CCG instead of 6-2 Texas Tech and 6-2 OU.
2006: 6-2 Nebraska plays in CCG instead of 6-2 Texas. Texas beat Nebraska in regular season.
2008: 5-3 Mizzou plays in CCG instead on 7-1 Texas, 7-1 Texas Tech, and 5-3 OSU. Mizzou lost to Texas and OSU in regular season.
2010: 6-2 Nebraska plays in CCG over 6-2 OSU and 6-2 Texas A&M. Nebraska beat OSU and lost to A&M. A&M beat OSU. Mizzou and OU also went 6-2. OU went to CCG despite losing to Mizzou.
In 10 of 15 years of Big12 CCGs, one of the participants was choosen based off geography/divisions. The only difference is divisions were the original method and thus accepted way of choosing CCG participants. That is no better than any method of choosing CCG participants without divisions.
Edit: To answer your other question. No, we shouldn't have been there in 2000 or 2003. I have no problem saying that. However, we should have been there in 1997 and 1999.