Tie-breakers are always something that cause controversy when they’re used but get little attention elsewhere. Nevertheless, I was looking at the Pac-12 division title tie-breakers and I think they’ve got them in a bad order:
- Head to head (in case of multiple teams, record against all the tied teams)
- Record in the division
- Record against the next highest team in the division (so 3rd place in 2-way tie, 4th place in 3-way, etc.) and then iterate down the list from the highest to lowest until the tie is broken
- Record in common conference games
- BCS standing
All of it seems reasonable except #3 and #4, which should be swapped. The problem with #3 is that for it to break the tie you have to ask yourself a difficult question.
See, for it to matter all the tied teams must have the same division record (see tie-breaker #2) and must have lost to someone outside the tied teams. A common example would be teams A, B and C each are 7-2 in conference. They have a circle in head to head (A beat B, B beat C, C beat A). But they each lost one other divisional game. So let’s say A lost to the team in 4th place, B lost to 5th place and C lost to 6th place.
So what’s “worse” or said another way, what should eliminate you from the tie-breaker? Is team C beating #4 a big positive, but their losing to team #6 not much a negative… or is it the other way around? Is the fact that you lost to a really bad team worse than the fact that you beat a higher team? It’s a difficult question without an objective answer. The Pac 12 decided beating the higher team is more important, which is fine, but arguable as to whether it’s the right choice. Thus it should be as low a tie-breaker as possible.
Tie-breaker #4 however is a great/important one in our divisional, highly unbalanced schedules. A new scenario: Teams A, B and C are 7-2 in conference and 5-1 in division. The only division loss is the circle of loses to each other. But team B has two conference losses because they’re the only ones who played Oregon and the other two got lucky, schedule wise. They instead lost to Washington State. Assuming team B also played WSU (and won), they should win the tie-breaker, right?
So, the way to make that happen is by looking at all common opponents, since that would eliminate a games against tough teams only one team had to play. It doesn’t always work (the other teams could have lost to non-common opponents as well), but when it does, it’s the most even judge and a very important factor in each team’s conference record. Eliminate the games that aren’t in common and then see how it turns out. It’s an objectively good way to judge the teams.
Thus, in my opinion, the conference needs to swap tie-breakers #3 and #4.
(BTW, unlike this will come into play this year. Any potential tie-breakers should be solved by head-to-head this time out (3-ways are unlikely (most likely: ASU loses to both UCLA and Oregon State, UCLA loses to USC, USC loses to Stanford and they all have a 6-3 conference record. Tie breaker would be BCS standing unless ASU loses to Arizona instead of Oregon State in which case it would be divisional record and USC goes to title game.)).)