I don't like SOS because they vary week to week. A lot of those AFC teams are hovering around 5-7, 6-6, 4-8, so one week, if they win, your SOS is better, then it drops the next week if they lose. The Chiefs are skewing the scale too by choking and letting all of these bum teams get a win too.
I maintain that the current iteration of SoS is meaningless.
Case in point: Beating the Chiefs in week 1-5 would have meant MUCH more than beating them in week 10.
I maintain that a meaningful SoS would be like this.
Week 1: no SoS.
Week 2: Week 1 SoS only. No other games including Week 2 are to be considered. (to consider one's own games is to skew the record. Winning teams would be penalized and losing teams would be given more benefit)
Week 3: Week 1-2 only. No other games including Week 3 are to be considered.
And so on...
Thus, it wouldn't matter if you played an opponent in week 2 that went 15-1 if they'd lost their first game. Your SoS for that opponent would be 0-1. That's it.
Teams would get NO credit for teams that win after they played, nor would they be penalized for teams that lose after they played.
The Eagles and Pats are being penalized because the Chiefs have lost their way after looking unstoppable in the first 5 weeks.
Thus instead of having 192 games that count (12 opponents x 16 games), it would be the sum of 1-15, or 120 games total and those would ONLY precede the contests for each opponent. Then we have to remove divisional games. That makes it 114. The reason it's 114 (or 113 if the last game is a divisional game) is that divisonal opponents are still counted twice with this new metric (since their improvement or decline matters), but the actual games a team played against that divisional opponent would NOT be counted.
Such a system would be far more impartial and relevant to how a team is performing week to week across a season.
Moreover, a further improvement would be to weight up to 4 immediately preceding games (I say up to because obviously, that can't happen before week 5). That would better illustrate how a team is performing in that moment and better capture sustained success or failure.
Moreover, when it comes to using SoS for conference schedules, I think yet another improvement would be to adjust for how a conference performs in non-divisional games. For example, the Pats SoS would be adjusted downward slightly most years because the AFC East in many years has been pretty bad.
The current system is so basic and creates such bias...
What kills me is that a programmer could put together the algorithm for ALL of these parameters probably in less than a day (the UI and other issues would or could take longer)
I won't speculate on why they continue to use such a biased and basic system. If I had money or was a programmer, I'd just have the tool made and then promote the crap out of it because unlike DVR... my "Adjusted Strength of Schedule" ...well... I'd have to figure another name because no-one wants to have the toughest "ASS" or be known to beat the softest "ASS"... and we all know fans and pundits wouldn't stick with ASOS... maybe Modified Strength of Schedule... yeah... that's better...
Anyway, the current stat that is SoS has so many flaws that to me, it's barely a useful tool. It COULD be useful, but it isn't very much at this time the way it's compiled and used.