College Football -- Better Than Ever

[Guynemer]

What's the relative value? For example, if two teams play schedules like so:
Team 1: A,B,C,C,C
Team 2: A,A,B,C,C

Both teams are (one) 'A' on each others' schedule. Team 1 beat Team 2, and lost to a 'C'. Both have one win against 'A', one against 'B', two against 'C' teams. Who is ranked higher?

Now:
Team 3: A,C,C,C,C
Team 4: A,A,A,B,C
Same thing, team 3 (A) beat team 4 (A); otherwise, team 4 has victories against 2 A, 1B, 1C, while team 3 has three victories and one loss to C teams.

Finally:
Team 5: A,B,C,C,C
Team 6: A,A,A,C,C
Ditto, etc.

How much does H2H matter versus SOS? How many equivalent-level wins cancel a H2H loss? Ignore differences due to other issues [conference, injuries, etc.] and solely focus on H2H here.

The head-to-head victor should be ranked higher, every time.

The only time that shouldn't apply is when one runs into a situation like the Texas-Oklahoma-Texas Tech troika last year, ie. rock-paper-scissors.

[Imran Siddiqui]

It depends. On how each team performs against those opponents first of all. Close wins are less impressive than blow out wins (and vice versa for losses).

However, if you'd like me to make a hypothetical decision ignoring for quality of wins or losses (which is impossible in the real world) - Team 1, Team 4, Team 6

[ColdWizard]

The head-to-head victor should be ranked higher, every time.

The only time that shouldn't apply is when one runs into a situation like the Texas-Oklahoma-Texas Tech troika last year, ie. rock-paper-scissors.

Fluke losses shouldn't necessarily outweigh a season of success.

[snoopy369]

Guy,
So if a team (say Oklahoma) beat Ohio State, Texas, Texas Tech, Florida, Cincinnati, USC, and Va Tech, and lost to (say) BYU, while BYU played BYU's current schedule and had one loss (as present), you'd rank BYU ahead of them?

That I have a hard time with (at that extreme).

[DanS]

At the beginning of the season, I agree that head-to-heads should determine rankings, with a 3 point modifier for home field advantage. F.e., Washington and USC would end up a tie and therefore you would be fine ranking USC ahead of Washington.

But once you have a half-dozen weeks played, the head-to-heads should have much less impact. The outcome of all games, weighted for degree of difficulty, should be the determining factor. Sagarin's rankings do this, so I like to follow his rankings most closely.

[Drake Tungsten]

Sagarin

Computer polls

[snoopy369]

How do we properly take into account smaller conference schools with that factor, though, when it's impossible for them to have as difficult of a schedule? It's certainly possible that BYU could beat the rest of the teams in the B12 [well, Texas might be hard, but who knows?], but now that they've lost to FSU, they're out of contention regardless...

[Imran Siddiqui]

It is almost impossible to do so (distinguish between conferences and how each team could do in another conference), except completely subjectively, especially considering the nature of the college football season (few games and a very large number of teams).

[Guynemer]

Fluke losses shouldn't necessarily outweigh a season of success.

But who is to say what is a fluke loss?

Is Washington beating USC a fluke? How do we know for sure? We assume USC is a better team, given last year's records. But last year is last year. These are new teams.

If Washington and USC finish 11-1, with Washington's only loss being to LSU, and USC's only loss being to Washington, how on Earth can one justify ranking Washington behind USC? And how on Earth can one justify ranking Washington behind USC now?

The same situation applies to BYU and Oklahoma. And BYU and Florida State, for that matter--the coaches have BYU 7 spots ahead of the Noles! What the hell kind of sense does that make?

The same situation applies to Oklahoma State and Houston. And to USC and Ohio State.



The problem is preseason polling. We get a set ranked list of where teams start, when in actuality every team starts the same: 0-0. This makes it near-impossible for teams ranked low on the list (or not at all ranked to begin with) to pass schools that started ahead of them for no reason other than reputation.

[DanS]

If Washington and USC finish 11-1, with Washington's only loss being to LSU, and USC's only loss being to Washington, how on Earth can one justify ranking Washington behind USC? And how on Earth can one justify ranking Washington behind USC now?

Pretty simple really. You have to consider that Washington won at home by 3. This should be viewed as a tie. If over the course of the season, USC wins better against a more challenging slate, it should be ranked ahead of Washington.

[Drake Tungsten]

Also, Nebraska should not have dropped in the polls after losing at Virginia Tech by one point. They should, in fact, have moved ahead of VT.

[Drake Tungsten]

Speaking of which, my Top Ten for the week...

1. Oklahoma
2. Florida State
3. Nebraska
4. Miami
5. USC
6. BYU
7. Alabama
8. Cincinnati
9. Virginia Tech
10. Boise State

[Imran Siddiqui]

It shall be funny to see who has forgotten last season and goes off on Drake's ranking

[Guynemer]

Pretty simple really. You have to consider that Washington won at home by 3. This should be viewed as a tie.

How should that be viewed as a tie? Silly me, I view it as a 3-point victory.

[SlowwHand]

So you are seriously saying that tougher schedules shouldn't matter?! Why play tough out of conference games then? Just schedule patsies.

You sat there previously, gave your Top 10, chattered your head off, in argument to my stance against scheduling weak names. Now this.


Ok, snoopy, I tried. He's a slut. He's a "gone to meet the ship at the dock" sized slut.