Would you have a problem with keeping the BCS system, but removing all human polls from the calculation? Base the computer rankings on efficiency stats (points per play, etc to account for jet tempo vs grind-it-out teams) and strength of schedule (more credit for wins or close losses to good teams, less when playing bad teams), and include some measure for margin of victory against good opponents. Make it important to beat GOOD teams by a lot of points rather than scheduling cupcakes. Give schools a choice - schedule patsies and play in an upper level bowl at best, or schedule upper-half BCS opponents in the non-con for an opportunity to play in a BCS bowl or national championship game. Every conference has it's Kansas/Vanderbuilt/Indiana, so not all conference games are killers.

The AP and coaches have an agenda to generate money and public interest, and even though the final vote is made public there are no repercussions. The computers have no such agenda. The reason people want a playoff is to get an objective national champion. If you just let the computers make the choice, you can still remain objective without drastically changing the post-season. It will still generally be the traditional powers in the big bowl games. But when a team like K-State has a great year in a tough conference, they will get rewarded appropriately.

You could alo argue that the human polls should be excluded because they start prior to the season, and pre-season ranking has proven to be critical to final season ranking. There's a reason the computer polls say "the numbers aren't good until Week 7."

Using just the cpu rankings would have given the following BCS bowl games: LSU/OSU in the championship, Alabama/KSU, Stanford/Arkansas, Oklahoma/Oregon and Boise State/South Carolina.