Posted by Adam Jacobi
One of the most odious aspects of the BCS -- and let's be clear, there are very many -- is the fact that the Coaches Poll constitutes one-third of the voting for the standings. If the poll's involvement weren't already accepted as normal, it would sound absurd: the selected coaches (or their selected assistant who actually fills these things out without attribution), given about 20 hours after the conclusion of their games, are tasked with ranking 25 out of the 120 teams in the FBS. The coach will never gameplan for, or have anything more than a cursory opinion about, the vast majority of these teams. The more time the coaches spend researching the poll, the less time they have to do their job (which isn't one with a great deal of spare time to begin with).
Thus, we get the same win-go-up, lose-go-down lazy polling that we can very well get from the AP already. What's the point? Does adding yet another hastily arranged Top 25 to the BCS add any merit? Moreover, isn't it a waste of what the coaches bring to the table for the BCS? Coaches do have exemplary abilities when it comes to evaluating other teams, after all, but that skill is primarily used in the daily rigmarole of their job, which is to say, on teams that they're actually going to play at some point.
So let's embrace that: have every single coach participate in the new coaches poll by ranking only their 12 opponents. As with traditional polls, a no. 1 gets the highest value (in this case 12), a no. 2 gets 11, and so on down the line. You know, like a normal poll. Now, since this is necessarily grading only FBS play (unless fans really want to see Montana come in at no. 8 in the poll or something similar), the teams with an FCS opponent are only going to be ranked by 11 opponents, so the rankings will be by average value instead of total.
Does this unfairly reward good teams in weak conferences (see: Boise State)? Well, maybe when it comes to their rankings relative to their conference pals. But look at who Boise's opponents are playing. Oregon State also plays TCU and Oregon. Wyoming got Boise, TCU, Utah, and Texas for this season (yes, Texas tanked, but that's an anomaly). Lowly San Jose State? The Spartans see Boise State, Utah, Wisconsin, and Alabama. Boise State may have some control over their schedule, but they certainly have little control over who their opponents play, and that's going to matter in this poll. Meanwhile, Ohio State may play in a tougher conference, but does anyone seriously think any of the Big Ten's coaches would rank another conference member over OSU as long as the Buckeyes stay undefeated? Would anybody have put Alabama second in the SEC before South Carolina pulled the upset?
Also, once the season starts to get into its late stages, coaches will be able to rank these teams based on what they saw first-hand in actual gameplay. Will this result in some coaches ranking teams based largely on how they performed against that coach's team? Sure. That's called rewarding wins and punishing losses. In other words, it's the entire point of polling. And if a coach seriously thinks a team that's, say, 19th in the AP played his team better than the 11th-ranked team, well, that's information that absolutely deserves to be integrated into the poll -- and it's much easier to justify making that adjustment in this format instead of the win-go-up/lose-go-down cookie cutter Top 25s.
Is this a perfect poll? No, of course not. There's still some value in a straight Top 25 poll, and the computer rankings have their merit. But if we're including coaches in the BCS process -- and we should! -- we should play to their strengths, not make them play pollster. This is how to do it.