College Football: How a 16-Team FBS Playoff Might Work

What follows is a simulation run on “ESPN College Football Playoffs Simulation (powered by AccuScore)”: http://proxy.espn.go.com/ncf/bowls08/bracket

The simulation was run only once, and for those who question the results, note that the percentages given after the score are the probabilities of each team winning based on 10,000 (!) simulations of every possible matchup among the teams allowed in the simulation. Note that only the Top 25 teams in the final pre-bowl BCS rankings are available for the simulation.

Where the probabilities (percentages) of a team's chances of winning do not match the result, an upset has occurred, and I've labeled it as such.

I could run the simulation with these particular 1-16 seeds over and over, and get different results each time. If I were to run it 10,000 times, then the "average" results of the 10,000 times would more closely match the probabilities, and the upsets would disappear into those "average" results. Of course, the games are played only once.

FIRST ROUND (Sweet Sixteen):

Group A:
Oklahoma 46, Georgia 29 (79% - 20%)
Boise State 41, Texas 25 (18% - 81%, upset)

Group B:
Ohio State 28, Cincinnati 17 (69% - 31%)
Penn State 27, TCU 20 (59% - 40%)

Group C:
Oklahoma State 35, USC 29 (41% - 58%, upset)
Texas Tech 39, Virginia Tech 27 (67% - 32%)

Group D:
Utah 32, Alabama 19 (21% - 78%, upset)
Florida 50, Georgia Tech 24 (87% - 12%)

QUARTERFINALS (Elite Eight):

Group A: Boise State 53, Oklahoma 31 (16% - 83%, upset)
Group B: Ohio State 34, Penn State 28 (38% - 61%, upset)
Group C: Texas Tech 43, Oklahoma State 39 (52% - 47%)
Group D: Florida 48, Utah 25 (81% - 18%)

SEMIFINALS AND FINAL (Final Four):

Semifinal 1: Ohio State 31, Boise State 22 (61% - 38%)
Semifinal 2: Florida 46, Texas Tech 32 (67% - 32%)

Final: Ohio State 42, Florida 27 (24% - 75%, upset)

***IMPORTANT*** Note that the locations of the games are not specified in the simulation and thus do not have an effect. We can assume that, in the simulation, all games take place at a “neutral” site (i.e. in cyberspace, or on the ESPN or AccuScore servers). In my playoff model on the previous page, I assign the games to specific locations, but do not project the results.

So... about this idea for a sixteen-team playoff...

Is it perfect? No. Nothing on this Earth is perfect.

Is it better than the "Broken Championship System?" I'd like to think so. Let me know what you think, both of the prospective matchups created by the seedings and of the playoff model in general. And feel free to try running simulations of this playoff model or another one of your choice.

You're currently logged into Facebook

Do you want to share your Bleacher Report activity on Facebook?