who plays in the BCS title game?

[BobJoeJim]

[ QUOTE ]
I'd also like to see the chances of the one loss teams winning out.

[/ QUOTE ]
So would I [img]/images/graemlins/wink.gif[/img]

So far I have Oklahoma at 54.7% and Oregon at 10.1% I'm not wrangling any more numbers into this spreadsheet right now, but I'll probably post later with numbers on South Carolina, USC, West Virginia, and maybe a few others. Let me know if you have any specific requests [img]/images/graemlins/smile.gif[/img]

[MyTurn2Raise]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
That's part of the ridiculousness, there should NOT be 119 teams in the same division, period.

D-IA should really be limited to at most 50-60 or so teams.

[/ QUOTE ]

I think I'll always be outside of this because I don't understand the obsession with National Champ

the goal is the Rose Bowl...everything else is a big whatever

[/ QUOTE ]

Perhaps you should post some stats showing how many more Rose Bowl victories the Big 10 has compared to the SEC.

[/ QUOTE ]

Pac10 has owned the series

I think alot of it has to do with Big11Ten teams having the goal of just getting to the Rose Bowl as much higher than winning the Rose Bowl

Playing at UCLA has factored in a few times

curious calls a few other times-- don't ever play West Coast teams on the West Coast

[/ QUOTE ]

So did you not read my question, or did you just miss the joke?

[/ QUOTE ]

hah...wtf....funny the tricks the mind can play
I saw Rose Bowl ---> Big11Ten ----> then just filled in with Pac10....lollerskates

I see it now

[MyTurn2Raise]

along with rwperu's comment
sagarin predictor is still bayesian meaning preseason ratings and previous season results are still weighing in

Here's how I do it these days for my 'most reliable'....I pull out the FBS scores from sagarin and put them in descending order

I then use the massey rating's composite rankings from 1 to 120 and line them up next to the ratings

Basically, I think sagarin does a good job quantifying the difference between #5 and #20, #12 and #70, #55 and #63, etc. However, I don't trust his ordering of the teams given the bayesian influence. Instead, I trust the massey computer composite more as it finds a way to weight all the legitimate computer rankings.



For homefield, a basic 3 or 3.5 isn't that great. Different venues are worth more. Different crowd enthusiasms are worth more.

Something that should be incorporated is that home teams get more of an advantage if teh two teams are near equal or the home team is slightly worse on a nuetral field. Home field should be bumped up another half point.

Home field doesn't play as much of a role in expected blowouts. For spreads over 21, I subtract a half point. However, that might not matter as much when looking at win pcts instead of spread picking.

[BobJoeJim]

[ QUOTE ]
along with rwperu's comment
sagarin predictor is still bayesian meaning preseason ratings and previous season results are still weighing in

Here's how I do it these days for my 'most reliable'....I pull out the FBS scores from sagarin and put them in descending order

I then use the massey rating's composite rankings from 1 to 120 and line them up next to the ratings

Basically, I think sagarin does a good job quantifying the difference between #5 and #20, #12 and #70, #55 and #63, etc. However, I don't trust his ordering of the teams given the bayesian influence. Instead, I trust the massey computer composite more as it finds a way to weight all the legitimate computer rankings.



For homefield, a basic 3 or 3.5 isn't that great. Different venues are worth more. Different crowd enthusiasms are worth more.

Something that should be incorporated is that home teams get more of an advantage if teh two teams are near equal or the home team is slightly worse on a nuetral field. Home field should be bumped up another half point.

Home field doesn't play as much of a role in expected blowouts. For spreads over 21, I subtract a half point. However, that might not matter as much when looking at win pcts instead of spread picking.

[/ QUOTE ]
All valid points. Which is why the numbers I'm listing should be taken with a big grain of salt. Still, For a really rough measure of the odds of going undefeated at this point, I think the system I used is as good as any other equally simple method would've been. As for more complicated methods, well, I'm not gonna put THAT much time into it when the only use I'm getting out of it is that I think it's kind of cool and that I can post it here :P

Even using Sagarin's ratings was actually an afterthought, I originally was just using my own computer rankings, but I figured if I was gonna' post it publicly I should use something a little more reputable, and Sagarin enjoys a good reputation (we'll dodge the debate of whether or not it's deserved) so it fit my purposes well.

[Gregatron]

pvn: Do you have any idea how many win the SEC has in the Sugar Bowl COMBINED over the Pac10, Big10, and all others?

HA! Take that!

[iggymcfly]

Here's a plausible scenario I've been turning over in my head. What if all the Pac-Ten teams end up with at least two losses, the Big East and ACC teams all lose at least once (or once more in the case of WVU), Ohio State finishes 11-1 after losing to Michigan in the finale, Oklahoma finishes 12-1, and Florida runs the table the rest of the way, beating LSU in the SECCG to get to 11-2?

Who would play for the championship? I'm thinking the voters would have something like 1.Okla 2.UF 3.LSU while the computers would have more like 1.LSU 2.UF 3.OSU which could be enough to lead to a rubber match between LSU and Florida in the national championship. It would be about the most bizarre scenario ever, especially with LSU getting a virtual home game in the championship to a team they'd already lost to at a neutral site. Oh, and of course Oklahoma would be playing some random mid-level team for the AP title.

Honestly, I don't see anyone (except possibly LSU) finishing undefeated, and with the Pac-Ten schedules, I'd say it's <30% anyone gets out of there with one loss, so the only real obstacle to this happening is Florida running the table. Their schedule's tough enough that they'll probably slip up somewhere, but they'll be favored in every game until the SEC Title game and even there, I doubt they're more than a one or two point dog.

[Riverman]

[ QUOTE ]
Here's a plausible scenario I've been turning over in my head. What if all the Pac-Ten teams end up with at least two losses, the Big East and ACC teams all lose at least once (or once more in the case of WVU), Ohio State finishes 11-1 after losing to Michigan in the finale, Oklahoma finishes 12-1, and Florida runs the table the rest of the way, beating LSU in the SECCG to get to 11-2?

Who would play for the championship? I'm thinking the voters would have something like 1.Okla 2.UF 3.LSU while the computers would have more like 1.LSU 2.UF 3.OSU which could be enough to lead to a rubber match between LSU and Florida in the national championship. It would be about the most bizarre scenario ever, especially with LSU getting a virtual home game in the championship to a team they'd already lost to at a neutral site. Oh, and of course Oklahoma would be playing some random mid-level team for the AP title.

Honestly, I don't see anyone (except possibly LSU) finishing undefeated, and with the Pac-Ten schedules, I'd say it's <30% anyone gets out of there with one loss, so the only real obstacle to this happening is Florida running the table. Their schedule's tough enough that they'll probably slip up somewhere, but they'll be favored in every game until the SEC Title game and even there, I doubt they're more than a one or two point dog.

[/ QUOTE ]

You really don't think tOSU is a favorite to run the table at this point?

[TomCollins]

[ QUOTE ]
[ QUOTE ]
Here's a plausible scenario I've been turning over in my head. What if all the Pac-Ten teams end up with at least two losses, the Big East and ACC teams all lose at least once (or once more in the case of WVU), Ohio State finishes 11-1 after losing to Michigan in the finale, Oklahoma finishes 12-1, and Florida runs the table the rest of the way, beating LSU in the SECCG to get to 11-2?

Who would play for the championship? I'm thinking the voters would have something like 1.Okla 2.UF 3.LSU while the computers would have more like 1.LSU 2.UF 3.OSU which could be enough to lead to a rubber match between LSU and Florida in the national championship. It would be about the most bizarre scenario ever, especially with LSU getting a virtual home game in the championship to a team they'd already lost to at a neutral site. Oh, and of course Oklahoma would be playing some random mid-level team for the AP title.

Honestly, I don't see anyone (except possibly LSU) finishing undefeated, and with the Pac-Ten schedules, I'd say it's <30% anyone gets out of there with one loss, so the only real obstacle to this happening is Florida running the table. Their schedule's tough enough that they'll probably slip up somewhere, but they'll be favored in every game until the SEC Title game and even there, I doubt they're more than a one or two point dog.

[/ QUOTE ]

You really don't think tOSU is a favorite to run the table at this point?

[/ QUOTE ]

I'm an tOSU homer, but I think its close to 50-50.

Kent State - win 99% of time
Michigan State- win 95% of time
@ Penn State - win 90% of time
Wisconsin - win 85% of time
Illinois - win 75% of time
Michigan - win 70% of time

These are rough guesses. This puts them at 38% chance of making it. It's definitely close to even money.

[iggymcfly]

Yeah, I definitely think OSU's an underdog to win out. They've got 4 very losable games and while I'd have a hard time making them a dog in any one of those games individually, you have to think 1 time out of 4, they're going to have some turnovers and end up on the wrong side.

[capone0]

We haven't seen tOSU ever perform against a decent team yet. Purdue isn't very good at all. The problem is noone on that list is particurely good.