Still, it's not like Ken sits down on his computer at the end of the season and tweaks his #s in a way to make the national champion the best team. 10 of the past 11 years, the national champion has ended up #1 on the list (crazy 2011 Butler/VCU tourney being the exception). Sure, it's possible for a team to "break" the rankings if they run up the score on patsies and play close games against all the better teams, but he's constantly adjusting the effect those have. All the numbers are, really, is an adjusted efficiency margin, used to predict future outcomes in games. They're really successful on the whole, but even if a team has, say, a 90% chance of victory, they'll lose out 10% of the time.