cfp rankings if stats were paramount

[beorach]

I know they have computers do this but I'm pretty sure they're not doing such basic calculations as I am.  If I just take the average of all the average category rankings from the nine categories I track for the undefeated and 1-loss teams in FBS that have a shot at the playoff (not looking at you, UCF), and then adjust for strength of schedule, these are your rankings:

 

 

 

Losses don't matter for this, as I was trying to indicate in my post title.  This is all about stats and how they measure up against the entire sample from games played against P5 teams by P5 teams and Notre Dame.  I only considered games against P5 teams such that Michigan's loss to ND isn't even in the mix (is the only BS part).  Even without counting ND, though, Michigan still had the toughest SOS.  I calculated SOS by averaging the average percentile ratings for all the opponents each of these teams faced (across the nine categories I share details for in the Husker Football forum on a weekly basis).  The simple adjustment I did for SOS was take the average percentile rating for a team's stats times the ratio between their SOS and Michigan's.

 

Notre Dame had the second-best SOS per this method despite Michigan, Stanford, and Pittsburgh being the only teams with an average category percentile rating over 50 (which would be exactly average).  LSU and Mississippi State are the only teams with above average ratings for Alabama.  Maryland, Wisconsin, and Penn State were above average foes for Michigan.

 

The best win is ND's over Michigan.  Second place is Clemson's over Syracuse (avg. category percentile rating of 63.28).  Third is WSU's over Utah (avg. category percentile rating of 62.46).

 

Without any SOS adjustment, the rankings would be as follows:

 

 

Lastly, here are the SOS rankings:

 

 

 

p.s. - I am having trouble getting rid of this extra table I don't recall having inserted to begin with so here it stays...

 

[beorach]

Here's the updated post after last week (which saw West Virginia drop out):

 

No SOS adjustment

 

 

With SOS adjustment

 

 

SOS Calc's/Rankings

 

 

p.s. - I accidentally left WV in there up above.

[jaws]

I understand this is stats based but I think the selection committee should take the SEC to task for cupcake Saturday. Even if it is for one week, penalize them for that schedule. Everyone else is battling conference opponents and the SEC pulls this crap every year. I love to see Alabama play a conference game the week before they play Auburn. 

[beorach]

On 11/19/2018 at 12:10 PM, jaws said:

I understand this is stats based but I think the selection committee should take the SEC to task for cupcake Saturday. Even if it is for one week, penalize them for that schedule. Everyone else is battling conference opponents and the SEC pulls this crap every year. I love to see Alabama play a conference game the week before they play Auburn. 

 

I agree.  What's worse is that we've made calculating strength of schedule too complicated.  One of the first sites that comes up, when I search for SOS rankings for college football, has Alabama at #5!

 

https://www.teamrankings.com/college-football/ranking/schedule-strength-by-other

 

I haven't crunched the numbers yet this week but would be surprised to find that Auburn has better stats than LSU and Mississippi State (to list the top two wins Alabama had before last week's games).  Michigan played one really good team all year, before last weekend, and they (Notre Dame) ran them off the field in the first half before going conservative in the second.  Not every P5 conference has even two really good teams annually but we act like we can tell which P5 champ is better than another despite having virtually no games between P5 conference teams?

[beorach]

I'm updating these lists again to include the six teams in the mix from P5 conferences and ND.  In the interest of defending my SOS rankings, please also consider the following...

 

Alabama has played no great teams (average category percentile rating of at least 68), three above-average teams (average category percentile rating between 50 and 68), two below-average teams (average category percentile rating between 32 and 50), and four bad teams (average category percentile rating no greater than 32).

 

Clemson has played no great teams, three above-average teams, five below-average teams, and two bad teams.

 

Georgia has played no great teams, five above-average teams, three below-average teams, and one bad team.

 

Notre Dame has played one great team, four above-average teams, four below-average teams, and one bad team.

 

Ohio State has played one great team, four above-average teams, three below-average teams, and three bad teams.

 

Oklahoma has played no great teams, three above-average teams, seven below-average teams, and no bad team.

 

No SOS considered

 

SOS considered

 

SOS Rankings

 

[Moiraine]

On 11/19/2018 at 12:10 PM, jaws said:

I understand this is stats based but I think the selection committee should take the SEC to task for cupcake Saturday. Even if it is for one week, penalize them for that schedule. Everyone else is battling conference opponents and the SEC pulls this crap every year. I love to see Alabama play a conference game the week before they play Auburn. 

 

 

Or I guess we could play their game and the other 4 conferences play a cupcake before their final conference game.