Just one game remaining, the traditional Army-Navy tilt in Philadelphia and the 2015 regular season will be in the books. All year we've been taking a look at one, well I suppose, two different ways of looking at national rankings. For good measure, let's throw in a third and see if that method gets us any closer to how the various computer, human and College Football Playoff polls look.
For the past fourteen weeks, we've been using similar design and metrics using the Gameplan pre-season annual (the Book) and Phil Steele's College Football Preview to create the weekly rankings. Each magazine assigns a power rating to teams in the pre-season and then between some update formulas the magazines provide and somethings I have added over the years, we are able to create updated rankings each week as well as point spreads for every game played.
As I have often said, the fact that something really so simple can be so reasonably accurate when looking at rankings and spreads is pretty amazing in and of itself, but now let's add one more piece to the puzzle.
What we do with the Book and Steele weekly are really power ratings. They aren't meant to be reflective of what the AP poll or the CFP committee produces, but remove wins and losses and stack rank teams based on how they would do against each other. This is why a team like an Arkansas or Washington right now (although the Huskies really confuse me) can be in our top 25 despite mediocre records.
One other measurement I have added in the past few years though is a bonus component for who you beat or lose to, in essence a strength of record component. This then takes into account who you play, when you play them and whether you win or lose and really removes direct margin of victory as a factor, although that does contribute to where those teams are ranked and how much of a bump you get for playing them.
One of the things I debated is whether strength of record, which I also use to determine strength of schedule needs to be weighted in some way. There is the fact that the CFP committee definitely puts more weight on games later in the season and allows teams to overcome early hurdles. It can also be argued that teams change over the season and wins that may look good in week 1, don't look as good in week 14. I have stayed away from that at least for now on the basis that teams do morph and that a win or loss for that matter, against Baylor was a bigger deal when the Bears were healthy as opposed to later in the season when they weren't.
Therefore, what I will refer to as strength of record, takes into account wins and losses and who they were against and gives you a more human poll like look.
The results were very interesting and are listed below, at least for the top 25. As you can see, although they don't precisely mimic the CFP rankings, they aren't far off. Clemson, who is #7/#8 in the Book and Phil Steele power polls respectively is now #2 in the Strength of Record (SOR) ranking.
What it also said to me is that the championship game "touch points", in other words, the 13th game, are in fact valuable and play just like the SOR does. Oklahoma then, didn't get added points for playing the 13th game and it caused them to be passed by Clemson in the rankings. It also allowed Stanford and Michigan State to pass Iowa. This was something Jeff Long called out in particular as Chairman of the committee
The one noticeable disagreement then between the committee and our SOR, is Michigan State. What we are then saying here is that Stanford's 11-2 record is superior to Michigan State's 12-1 record, although I will tell you the margin is thin.
To confirm that we look at strength of schedule, where Stanford was #1 in the country, with Alabama #2. The others in the top eight. Oklahoma (9), Notre Dame (10), Michigan State (36), Clemson (41), Iowa (51) and Ohio State (61). That difference between playing what this method viewed as the nation's toughest schedule actually overcame the one loss to put Stanford ahead of Sparty.
Teams not listed on our Strength of Record ranking who are in the College Football Playoff Top 25 are: Temple (26th in SOR, 33 Massey, 24 CFP), Navy (28th in SOR, 18 Massey, 21 CFP), Tennessee (39th in SOR, 22 Massey, 23 CFP)
Teams in our Book/Steele top 25 not in the SOR rankings, Arkansas (37), Washington (48), Tennessee (39), Mississippi State (31), West Virginia (47), Bowling Green (36).
If you are curious, this format didn't do BC any favors. The Eagles wind up the regular season 99th in the country. That puts them ahead of only Oregon State (102), Purdue (104) and Kansas (118) among Power five schools. BC played the 64th ranked schedule.
This week's lone game has Navy a 38.8/39.0 favorite over Army. Once again a large spread and that didn't work out too well for us last week in rivalry week with both magazines having losing weekends. Heading into Army/Navy and the bowls, Phil Steele holds a 6 1/2 game lead.
Back next week with the bowl predictions.
Enjoy your football weekend and Go Eagles!
Legend - SOR (Strength of record ranking), GP/PS (Gameplan/Phil Steele), Massey/CFP (Massey composite ranking, College Football Playoff ranking)
:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/47814983/usa-today-8978072.0.jpg)