College Football Playoffs: Simulation Nation

Buckeye CommentarySenior Analyst IDecember 17, 2008

Let me preface this post by saying that I am not a huge fan of computer rankings and/or formulas playing a role in determining the MNC. It is a part of our favorite sport, however, so I cannot help but be intrigued when a different sets of computers predict what will happen this bowl season, or in a hypothetical playoff. I did not start this post with a conclusion in mind and this is piece is more observation than analysis. As the fine print, says “your results may vary.”

Within the past few days I have stumbled upon four separate online, simulations of the bowl games and/or a hypothetical playoff. Three are computer-generated, and one is voter-based. As a way of background, here are the four simulations I “studied:”

I checked out Pick Center on ESPN (Insider status necessary), which, according to the site, “simulates the odds on whether the winning team will cover the point spread [and provides] fans with college football predictions and predicted stats, scores and player performances.”

This is easily the most complicated, in-depth, and fun of the simulations. If you choose the “simulate it” link you are able to simulate the game while controlling multiple variables. You can designate player statistics and alter the game conditions. It takes some time to adjust everything but it makes the results even more compelling. Generators beware: you may not want to know the results even after tailoring them favorably to your favorite team.

You can also click on “more info” link in the Team Rankings column and get very cool statistical analysis powered by For my money, the coolest feature is the “Similar Historical Matchups.” The description of the origination of these matchups is vague, but the chart is intriguing if nothing else. The chart for Ohio State v. Texas is thus:

2) This next one is also on ESPN and powered by AccuScore technology, but this time it is a playoff format. You can select one of four ranking systems and play out a 16-team tournament. It does not much matter which rankings system you choose (unless you totally gerrymander the custom version), you will almost always end up with a bracket that looks like this:

Interestingly, the very first time I ran this simulation I got this bracket:

Horrifyingly, I must have run this sim a dozen times before Ohio State advanced out of the first round and I eventually gave up trying to get them to the second round after 20 or so run-throughs.

The next two playoff simulations can be found on Sports Illustrated’s site. Basically, they are variations on the same theme. In one version, readers vote. In the other, the games are played out by EA Sports NCAA Football ’09. Unlike the ESPN phony tourney, the SI playoffs only involve 12 teams. The good folks over at SI do not offer an explanation for the 12-team format, and we can only assume it is to mirror the NFL playoffs. One interesting note (or perhaps not), is that Oklahoma and Florida met in the EA Sports final, with Florida winning a shootout.

After browsing these interpretations of the “most wonderful time of the year” I had two distinct thought streams: (1) what do these simulations say about the Buckeyes (if anything), and (2) what can we take from these simulations and apply to the BCS v. playoff debate?

The short answer to #1: Microchips hate Ohio State. The three computer-generated playoffs treated the Buckeyes like a rented mule. The most forgiving was the EA Sports sim in which the Buckeyes only came up a touchdown short to Texas Tech, but the AccuScore system gives Texas has an 80 percent chance of winning the Fiesta Bowl by a score of 32-22. When you use the BCS or ESPN Power Rankings in the bracket format, OSU’s best odds of advancing out of the first round are 47-percent (versus Utah).


The more startling conclusion is for #2: A tournament style playoff is what most college fans pray for. Judging by the results of these simulations, however, most of those fans would be (severely) disappointed. Looking at the percentages of the CG results, most of the 15 games in a 16-team tourney would not be any more exciting than the average game, and this is season with several teams having a legitimate claims at #1. Looking at the first bracket above, only the semifinals and final had games within 20 percentage points. That does not sound like an entertaining playoff.

I know you thinking that computer projections do not have anything to do with real games and I agree. So, fans should still be excited about the concept of a college football playoff, right? Then why do the fans vote in such overwhelming fashion to the contrary?

If you look at the Sports Illustrated bracket determined by reader vote, it depicts a tournament without a bit of suspense through two rounds. The closest match in the first round was Texas Tech’s “victory” over Ohio State by a margin of 72-28 percent. In the second round, the readers had barely any additional faith with only one matchup voted within 20 percentage points (USC over Alabama 57-33). I assume the voting will be closer for the semis and finals, but what this tells me is that most of America only wants a 4-team playoff (at least subconsciously).

I know you can argue the geo-analytic data may reveal voting skews causing these disparate percentages but I am not buying that. These sites are too heavily trafficked. Plus, you know how college fans are, they will go out of their way to vote as many times as possible for their team.

You could also point out that the voters do not give a margin of victory so, in theory, they could all believe it will be a close game, but that a certain team will win 72-percent of the time. To me, the overwhelming percentage point leads indicates a great deal of confidence by the collective voters. I think it is safe to surmise that that confidence is accompanied by a belief that these games would not be nail biters.

Like I said, I did not start out to come to any specific conclusion and I am pretty certain that I have not. This may be a bunch of rambling, but the upshot for me is that the BCS may not be as bad as a 16-team tournament.