How the BCS Conferences Select their Out-of-Conference Schedules

What began as simple curiosity as to how many FCS teams were played by SEC schools turned into something larger.  By looking through this year's schedule and compiling the out-of-conference numbers, I found some interesting trends in OOC scheduling among the major conferences.  (Ok, it's interesting to me.  Your mileage may vary.)

It first began by looking at Colley's compilation of conference records, which is one of the more convenient tables on the net for this sort of thing.  Ignoring his scoring of conference strength (which is only as good as his actual rankings), that table lets you see the OOC results at a glance.  The first thing that struck me:

WOW!  The SEC has only played 6 FCS teams and 5 Sun Belt teams.  That's less than every conference but the PAC-10!

Then I realized that a part of this disparity stems from the way schedules are set up.  SEC teams have at least one OOC game late in the season (occasionally two), while conferences like the Big 10 and PAC-10 tend to play all of their OOC games prior to playing any conference games.  Having narrowly avoided a misinterpretation, I then looked at the number of scheduled OOC games by major conference.

* - The Sun Belt numbers are also included in the non-BCS numbers, here and onward.  I show the Sun Belt in particular because the conference is traditionally bad enough that scheduling a Sun Belt team is not much different than scheduling a FCS team.  You know they're going to be outmanned.

(Yes, I included the MWC, mostly since they're generally seen as on par-ish with the ACC and Big East as of late.)

What first stands out is that playing Sun Belt teams seems to be a function of distance from the Sun Belt.  The Big 12 is actually closer to the Sun Belt than the ACC.  Here's a FBS map to help you see this.  Question: why would the PAC-10 even bother scheduling a Sun Belt team?  The travel is more expensive for the Sun Belt team, and there's no benefit to the strength of schedule for the PAC-10 team.  That goose egg should be of absolutely no surprise when viewed from a logistics perspective.

But now we have to consider that not all conferences play the same number of OOC games.  The PAC 10 only plays 3 OOC games per team, while the Big East plays 5.  So let's look at the OOC games as a percentage of the total.

Huh.  With the exception of the Big 12, all BCS conferences play roughly 1/4 of their OOC games against FCS opponents.  (That 25% for the ACC isn't sufficiently larger than the rest to worry about.)  The MWC also plays a smaller number of FCS games, but that's partly because they're a mid-major who is still trying to their way up to BCS status.

But it's also partly due to geography (which will be a theme).  Here's a map of FCS teams.  Notice how the FCS is heavily concentrated toward the east. There are 91 teams East of the Mississippi and 34 to the West.  If you ignore FCS schools in states that border the Mississippi (and Texas, where all FCS schools are in the southeast of the state), that number shrinks to 17.  Four of those - the Dakota schools - are on the Minnesota border, meaning that the PAC-10's convenient pool of FCS teams is, at a maximum, 13.

This goes a long way toward explaining why the Big 12 and MWC don't play as many FCS teams: there just aren't many FCS schools close to these conferences.  In fact, it's a little surprising that the PAC-10 plays as many FCS teams (rate-wise) as most other BCS conferences when there are so few FCS teams available to play.  In that sense, it's also surprising that the PAC-10 plays so many BCS opponents.  The PAC-10 motus appears to be to either play BCS games or to skate low whenever possible - at least relative to other conferences.  Considering that the PAC-10 has an extra bye week, thanks to a lack of a championship game and the tendency to schedule their last game on championship Saturday, this strategy effectively equates to getting a FCS team for a third bye week (of sorts, if you can rest your starters after the first half) and scheduling a BCS game to bring up the strength of schedule.

Let's look at the SEC one more time.  Their scheduling of BCS teams and of FCS teams is right about on par with the other major conferences.  They do schedule more Sun Belt teams than everybody else, but that's simply a function of distance - especially when the public schools of the SEC get a chance to double up on in-state revenue by scheduling an in-state opponent (e.g. Troy and Alabama, or LSU and Louisiana-Monroe/Lafayette).  But their 33% of BCS opponents is right about on average across all BCS conferences.

Well, enough of that for the moment.  Let's now look at that geography thing.  It's a little harder to fairly evaluate proximity.  The ideal approach would be to calculate the distances that each school resides from potential opposition (e.g. how far away are SEC teams from FCS teams, and how many are within an easy commute?).  But that's a mountain of work, so I instead took a look at the number of OOC schools that reside within the same states as a conference's member schools, and I looked at the number of OOC schools that are within the same states, plus the adjacent states.  This gives a rough measure of the number of convenient opponents, even if it's not nominal.

It's good to be the SEC.  If you ever need to defend the distance traveled by SEC schools, not that there are 127 other schools in the same state or within 1 state of an SEC team.  That's a sizeable pool for a conference that plays 48 total OOC games.  In contrast, the PAC-10 has a total of 22 schools in such proximity, while playing 30 OOC games (31 this year, thanks to USC's Hawaii game).

Going back to the Big 12 and MWC, their dearth goes further than the numbers in the table - especially for the MWC.  Were it not for TCU, the MWC would only have 14 FCS schools in common+1 states (and 4 fewer were it not for Wyoming, who almost never plays the Montana or South Dakota schools).  The same applies to a smaller degree to the Big 12; even though the South is close to the Texas, Louisiana, and Arkansas FCS schools, the North isn't.  That tends to depress the FCS numbers.

Another surprise is that the ACC only has 29 FBS teams in their common+1.  This is mostly a fault of North Carolina and Virginia, who not only have a majority of the in-state teams in the ACC, but also share a border, reducing the number of available +1 teams for consideration.  Still, there are a lot of FBS schools in the common+2 region, which is not really a far drive for a lot of the ACC teams.  This is more a function of geographical wonkiness than actual limitation.

All in all, it seems that BCS teams would prefer to schedule about 1 FCS opponent per year to theoretically gain a pseudo-bye week.  Deviations are mostly a function of geography except for the PAC-10, who seems to operate on a more committed strategy of playing FCS teams.  That said, the PAC-10 makes a pointed effort to schedule more BCS teams than most - something they can afford to do thanks to that extra bye week.  It'll be interesting to see how their scheduling changes once they have 12 teams in the fold.

The final conclusion appear to be that, for all we like to make of OOC scheduling as a function of conference personality, it's really a lot blander than that.  Scheduling appears to be governed more by proximity, the potential for championship games, and the number of in-conference games than by any grand scheme.

The one remaining question is whether the OOC games scheduled are intended to be easier or harder for the different conferences.  Not all BCS teams are equal, for example.  The SEC may schedule the different categories of teams at the same rate, but are they scheduling easier teams within the categories?  That's not an easy question to answer; with the exception of some perennially bad teams, it's very difficult to assume much about a future opponent's quality when scheduling them.  Even if a team is scheduled only three years in advance, you really can't predict their strength when you do play them.  While I admit that this is an intriguing question, I leave it unanswered because I really don't know of a good way to evaluate it.