Tuesday, November 25, 2003

The deep look at the BCS

NOTE: Despite the current color scheme of this blog, the little "tiger at a press conference" logo at top right and the fact that LSU is currently ranked No. 3 in the BCS, this piece is not about LSU's chances to move past USC into the No. 2 spot of the BCS and play for the national championship. To write about such things would be to look way, way too far ahead. We still have to beat Arkansas the win the SEC Championship game. I understand that, and therefore this piece is a hypothetical examination of the BCS dynamics.


In the office yesterday, me and the one guy I don't hate up here were talking BCS and whether a team in third place near the end of the regular season could possibly pass a team ranked second if the second-place team does not lose its last game of the season.

The only talk about such a circumstance puts it at a longshot. But let's take a closer look at this completely hypothetical BCS situation and I'll try to show how such a flip could be likely.

In this situation, we have Team A (at 2nd in the BCS) and Team B (3rd in the BCS). Both teams are 10-1. Team A is second in both human polls, Team B third in both. Team A's average computer ranking is 2.33, Team B's is 3.00. Team A has a better strength of schedule (SOS), but Team B has -.4 quality win points, leaving Team A at 6.89 in total BCS points and Team B at 9.04 (lower, of course, is better).

The conventional wisdom is that Team A is a lock if it wins its last game. No way Team B can make up a 2.15 spread if both teams win out and have just 1 loss each.

But ... let's look a little closer. Team B not only has a final regular-season game to play this week, but a conference championship game after that. Team A does not play its final game until the same day as Team B's conference championship game.

Team B's opponent this week is currently 8-3, which will help improve Team B's SOS. An educated guess is that this week Team B's SOS points would improve by about .44. And because SOS is a major component in the computer rankings, Team B could very well improve it's BCS points there by .50.

Team B has already passed Team A in 2 of the 7 computer ratings, and if Team B displaces Team A in any more, Team A could lose a few tenths as well. There's a real possibility that computer rankings could be a wash by next Monday at 2.50 each.

And Team A is likely to suffer a slight drop in SOS in their off week because of games played by its opponents.

Therefore, it's quite possible that the 2.15 point gap between A and B this week could narrow to around 1.25 by next Monday.

Then the following week, Team A is back in action against a 7-4 team while Team B would play either a 9-2 or 8-3 team in its conference championship. Team A's SOS would stay more of less the same, while Team B would see another SOS boost, gaining another .2 or so in the BCS.

That boost will also help in the computer rankings, which will be the final component to come together in the final standings Dec. 7. Team B would need to gain an edge in computer rankings over Team A, which is not far fetched given the quality win and improving SOS.

However, as everyone is quick to point out, even then Team B would not have enough juice to pass Team A (especially if Team B's conference-championship opponent is the team currently giving Team B a -.4 quality win bonus).

But - and here's what nobody seems to be taking into consideration - after their games on Dec. 6, Team A will be 11-1 and will have just beaten a mediocre in-conference opponent to finish the season, and Team B will be 12-1 and will have just beaten a team ranked in the BCS top 10.

Why does that matter? One word: Humans.

Between now and Dec. 7, Team B could very well gain .56 points in SOS, a full point in computer rankings, but lose the .4 quality win point, meaning a net gain of 1.16 points. That's still .99 less than Team A.

And that's where the humans come in. Come the wee hours of Dec. 6, when Team B has just beaten a BCS top 10 team to win its conference championship and move to 12-1, 65 sportswriters (AP) and 63 coaches (ESPN) will have to decide if they really think Team A - at only 11-1 and with a season-ending win over its conference's 5th-place team - is the country's second-best team.

If 33 of the sportswriters or 32 of the coaches believe a 12-1 Team B is better than a 11-1 Team A, Team A will lose .5 in the human poll average, and Team B will gain .5. That's 1.00 in BCS rankings if just one poll flips Team A and Team B, and that's enough to move Team B up to BCS No. 2.

How likely is all this to happen? Hard to say. But if Team A and Team B were to both win out, it's a certainty that Team B will have improved its SOS and computer ratings, and it would be up to the humans to declare a team with one fewer wins over lesser opponents to be more worthy of a national title shot.

0 Comments:

Post a Comment

<< Home