Ray's Rankings Week 14

 Hi folks!

This week I'm going to do something a little different and spend some time on comparing inputs and how they affect the rankings.

Before we dive in, though, let's talk about some football! Lots of fun games yesterday; I tried to watch as many as I could. I just changed from Hulu Live back to YouTube TV because of Hulu's price increase, and it's easier to keep up on scores on YT.

My few takeaways from the weekend:

• Alabama may be unbeatable. The deeper I dig the fewer weaknesses I see. Yes, they've only been punched in the face once by Ole Miss, but that didn't seem to matter.
• Oklahoma has been playing defense. So weird to see. 
• The wheels have fallen off in Eugene. It may walk like a Duck and quack like a Duck, but it plays football like an albatross.
• UW let themselves get into a deep hole and this time couldn't climb out.
• Clemson looks less dominant than last year even with that 45-10 win.
• Ditto OSU, even as well as they played against MSU. Of course, they were missing 23 players and their head coach so maybe I'm wrong about them?
• And every Indiana win makes OSU look better.
• ND is still playing very solid football. They don't look special, but they keep doing their thing.
• I knew that TCU was going to do that! Sleep on Gary Patterson at your own risk. 

The one really fun game was BYU v Coastal Carolina on that high school field under middle school lights. The Chanticleers looked the biz with that read option played to perfection. Zach Wilson just shut up everyone clamoring for his Heisman campaign. Sorry, Cougars, this isn't 1984.

Top 25

Right now my biggest question is this: how good/bad are Group of 5 teams in general? Last year I treated them as basically half as good. If you beat a G5 school, you got 0.5pt. This season I've been making a lot of changes along those lines, but I don't really know what the right answer is. I mean, there's simply no way to argue that beating Cincinnati should be worth less than beating Syracuse this season, but that's just this season. Same with beating Clemson vs beating FIU; they're just not equal. 

My opinion is that an average P5 or G5 team beating a P5 team is generally more difficult than beating a G5 team. Is it half again as difficult? 25% more difficult? Again, I'm looking for an average here. Boise v OU in 2007 clarified a lot of things for us as far as G5 schools go. Yes, they can play with anyone. Yes, we need to consider that they are better than we usually think.

So let's play with the numbers. This first list is the algorithm from last season: 1pt for P5, 0.5pt for G5, -1pt for FCS.

Note that Clemson is hurt most by playing FCS schools. This is the way I will go back to ranking next season. Teams that play FCS schools are benefitting from a bye week being counted as a win, and I find that unacceptable.

How about if I don't subtract points for FCS schools? I'll just count them as 0pts. It's still a penalty, but not as bad.

And there's Clemson back in the top 4. Cincinnati breaks into the top 25. OK, what if we count G5 schools as 75% as good as P5? G5 at 0.75pt, FCS still at 0pt.

Well, Miami really jumps up. Cincy is up to 16, and please welcome Tulsa, SJSU, and BYU. 

This is all well and good, but what if we treat G5 as equal to P5? This part gets interesting. All P5 and G5 wins are 1pt, FSC still 0 pts.

Well, well, well. Cincinnati, Coastal Carolina, Tulsa, SJSU, BYU, UL-Laf, UCF, Nevada all crashing the P5 party. I can hear you now: "San Jose State? They'd lose by 100pts to Oklahoma!" Maybe. Maybe they do a Boise State.

Let's try one more. We'll go ahead and count FCS wins as just a 1pt win.

Gotta admit that it was hard to type for a bit there. Tears of laughter filled my eyes; couldn't see the screen! Can you imagine?

There's a lesson here. Let's go back to the 1984 CFB season when a truly undeserving BYU was crowned national champ by voters who couldn't pull their heads out long enough to check schedules. The Cougars finished 13-0, but they played no teams with less than 4 losses. Their bowl win over a 6-6 UM in the Holiday Bowl didn't help. Meanwhile, UW had one loss to USC and beat OU in the Orange Bowl to finish #2. That particular travesty is what happens when you let people vote for these things, and that is what would happen if my system treated all wins equally. 

Not all wins are equal. Neither are all losses. So let's get back to where we ought to be. Wins are 1pt for P5, 0.75pt for G5, 0pt for FCS; losses are 0pt for P5, -0.25 for G5, and -1 for FCS.

The super surprise to me is Cincinnati, but only until I realize that they did play Austin Peay. 

I kind of like this setup for now. Somewhere in the near future I am going to download the last 20 years of games played, sort them by conferences, and measure the P5-G5 matchups to see how it actually plays out. Until then, enjoy the weirdness!