There definitely are other metrics available. It's a question of whether or not they are better and what you are even trying to measure. That's kind of where the tricky part is with quantifying achievement. Power ranking is in some ways simpler because the goal (predict future games) is clear and weird **** in pursuit of that goal (i.e. not much difference between a 1 pt loss and a 1 pt win) is fine because your goal is clear. But what do we even mean when we say achievement?

I think a sharper logistic curve might be some of the answer - essentially instead of changing the posterior win probability metric just tweak it so the model sees less of a difference between a 28 point win and a 5 point win.

Trying to quantify achievement makes no sense much of the time. It's basically the foundation of like 90% of CFB ranking arguments (with other 10% being should we use power or achievement and in what proportions).

As in they are only top 20% instead of a top 6 team?

Pedo State has literally played NOBODY so far. They could easily wind up as a three loss team.

You mean figuratively.

You want to know a team that has "literally played NOBODY"? That would be Clemp. They're on their way to a zero win season. Literally.

Might be time for the Italian Cheese Man to take a break from this thread for a little while

2 loss notre dame, 2 loss big 12 champion OU, 2 loss B10 champion oh st, 2 loss UGA all get in over an undefeated UCF. And probably deservingly so.

Also with their higher ranking USF probably still has the inside track

Think I asked about this last year but how do you ypp guys differentiate between a team that averages 6 ypp but gets that average from long plays vs a team that consistently gets 5 ypp? And what if the lower ypp team also goes for it on 4th more often?

Look at NDs remaining schedule - not a chance theyre making it through that gauntlet and making the playoffs.

Vs USC
Vs NC State
Vs Wake
@Miami
Vs Navy
@Stanford

I use this for season YPP stats:

https://www.teamrankings.com/ncf/teams/

The team with ypp of 6 is differentiated from the team with a ypp of 5 by being considered 20% better. I don't understand your other question; it's yards divided by plays run. 4th down plays just go into both sides of the calculation.

I'm saying people who sort by ypp. Wouldn't you rather get 5 yards every play to sustain drives rather than a big chunk then punt?

I sort by ypp. I'm one of those people.

I mean, nobody *actually* gets 5 yards every play and scores a touchdown every drive, just like how "big play" teams sometimes string together several big plays in a row. If you match up these teams the one with 6 ypp is going to win the vast majority of the time.

If you want to imagine a bad side to your consistent 5 ypp team, just watch Georgia Tech the next time they need to run a 2 minute drill.

Here are the teams currently in the top 20 in both ypp on offense and ypp conceded on defense (only counting games between FBS schools, which is all I could find quickly): Alabama, UGA, Ohio State, Wisconsin, Miami. Notre Dame and Oklahoma State would make it if you say top 22 instead of top 20.

As a single (pair of) crude indicator ypp meets the eye test pretty well. Ranking by averaged (offense and defense) ypp ranks appears to give a decent overall ranking.

Simply taking the average of ypp rank and ypp against rank (not the raw numbers, the ranks) yields the following ranking:

Alabama
Ohio State
Wisconsin
Georgia
Oklahoma St
Penn State
Central FL
Miami (FL)
Oklahoma
Notre Dame
Clemson
S Florida
Washington
Arizona
Auburn
USC
VA Tech
S Mississippi
Iowa State
GA Tech
LSU
Texas Tech
Louisville
Michigan St
N Mex State

Obviously S. Miss is some artefact of excluding non FBS opponents, but otherwise it's a pretty reasonable ranking given how apparently little information went into it. As I indicated, I just copied data from some random website, and not sure I pasted everything well, but still the top 10 seems about right. Bama #1. OSU, Wisconsin, UGA in top 5. No PAC teams in the top 10.

This is just asking to take into account second order information from the distribution of yards per play, that is variance. You just need a table of the yards gained/conceded on every play, and you need to do something with the down too. This information surely improves the analysis, but it's a huge amount of work, and probably the difference is not so big, or even washed out completely by the heterogeneity of opposition (in the end you are comparing ypp against different opponents, and that difference is more problematic than ignoring second order information).

This isn't too bad.