*DSFL S(22) Power Rankings Edition 2

[iStegosauruz]

Performance is just yards and points put into a little bit math. The top team of each category (yards, or points) gets 1.5 points. the max you can have from performance itself is 3 points. every team below the max is given a percentage of 1.5 based on their own performance.

Whats the reasoning behind choosing to use yards as a factor into the metric? I think using a sum of yards is really faulty and you end up with a lot of noise in the calculation. For example, its a really easy metric to influence based on in strategy decisions in the sim. Teams that run a higher tempo have the potential for putting up more yards. Its not always a 1-1 to conversion between more yards on offense with a higher tempo and giving up more yards on defense because they're on the field more often as well. I also don't think it fully captures the differences in strategies you can run. In general, most DSFL teams do run Power 30; however, in theory you're going to get more passing yards per attempt than you are rushing yards per attempt. That would mean teams that pass more are going to scale a total number yards more often than teams that run the ball more.

I'd also be interested in knowing whether a yardage total - i.e. pass yards, rush yards, total yards, etc. - is better for producing an accurate result of a good team than a stat that delves a bit deeper into overall efficiency and normalizes a bit better for tempo. For example, I'd generally look at rush yards and pass yards allowed per attempt as a better metric for assuming team quality than a total number of yards put up. The same can be calculated on the defensive side of the ball and generally feels like it would be more accurate. I haven't done any direct work that determines the weight each of the metrics factors into points with but I do know from some tangential work that when it comes to point differential or total points put up in a game that yards per attempt have a much better relationship with point differential and total points in a game than a number of yards per game does. The relationship having been calculated with a regression. I'd assume that relationship would also hold for determining individual efficiency and effectiveness.

Just to be transparent as well I think one of the major reasons I'm interested in the underlying methodology behind your formulations here is because I think we see some really wonky results in this current iteration. For example, they're only one slot in front of Dallas but the Pythons lost at home 34-0 against Myrtle Beach and 34-18 on the road against Minnesota.

Dallas for comparison beat Minnesota 27-15 at home and then lost on the road 26-6 to Myrtle Beach. They performed better against the same opponents that Portland played and arguably drew the harder home/away split with having to play Myrtle Beach on the road (better offense and better defense in your rankings). Yet the formulation somehow weights them as a better team?

Now I understand you also use TPE as a factor in the overall calculation but Dallas had the 5th TPE offense this week compared to Portland at 6th. Portland had the 4th best defense compared to Dallas at 6th. So the margins there are fairly close. The difference is in basically the same matchups portland put up 13 points and gave up 68 for a -55 point differential. Dallas put up 33 points and gave up 41 for a -8 point differential.

Portland is one net slot better in TPE rankings than Dallas but had a much worse performance against equivalent opponent and gets ranked higher? And the gap between them in your overarching number is the third biggest gap between teams? Behind Myrtle Beach -> London and Minnesota -> Kansas City. It feels like there is a lot of noise in the rankings or that they aren't utilizing the metrics that have the best relationships to points/team quality.