Bowl projections: good and bad

I’ve mentioned before that Bowl projections are good filler, but ultimately hot air. The records and revenue determine which team goes where. Here are examples of good projections and sloppy ones. SI.com Stewart Mandel does a pretty solid job (he has BC going to Boise). Sportsline does a crappy job (BC returning to Charlotte?). Why do I say it is crappy when this is merely a guessing game? Because some options -- like poor drawing teams returning to a Bowl site (i.e. BC returning to Charlotte) -- can clearly be eliminated and Sportsline didn’t do that. How hard is it to make a good fake list?