I don't see anything wrong with those rankings, and I don't think it's "just popularity." Power rankings are an attempt to organize how good people think different teams are, taking into account the fact that we know more about the teams than just their records in the first two games of the season. If you want perfect consistency with season-to-date W-L records, then just rank by that.
In the 2022 season, the 49ers lost their season opener to the Bears. In that situation, it seems like it would have been more than reasonable to put the 0-1 49ers ahead of the 1-0 Bears. By the way, when the season ended, the 49ers were 13-4 and the Bears 3-14.
"Power rankings" are just about all made-up nonsense, but a few people have tried to make some more-objective rankings of teams based on things other than raw W-L record, and the 49ers-Bears example shows that it makes sense to talk about how good we think teams are without going 100% Big Tuna ("you are what your record says you are").
Back in the aughts,
Baseball Prospectus (or somebody) had a thing that might not have been called a "power ranking," but it based them on how the teams were performing on the field and not their season-to-date W-L records.
One thing they used was expected wins based on runs scored and runs allowed ("first-order wins") via formulas like Bill James's "Pythagorean" estimate of expected wins given runs scored and runs allowed and the modifications of it by Clay Davenport (the "Pythagenport" formula) and another one developed by some Patriots fan (and hence known as "Pythagenpat") that Davenport ended up agreeing was better.
The thing is that just like there's luck involved in W-L records, there's also some luck involved in run scoring and run prevention. To reduce the effects of
that luck on
BP's rankings (which I think they called something like "adjusted standings"), they took some of the formulas to predict run scoring based on stats measuring what players were doing individually (hits, walks, outs, extra bases, stolen bases, fielding balls in one's area of responsibility and getting them where they needed to go, pitcher strikeouts,
etc.) and plugged those into Pythagenport or Pythagenpat to get "second-order wins."
But wait... at any point in the season, some teams will have had schedule luck in facing weaker opponents (think of a strong team in a weak division) and some have had bad schedule luck in facing stronger opponents. So the folks at
BP adjusted the components of second-order wins for the quality of the offenses and run-prevention units (defenses and pitching) each team had faced, yielding "third-order wins."
Does that seem like a lot of nonsense to you? Does it seem less so when you learn that "second-order and third-order wins" are better predictors of future team winning percentages than "first-order wins" and especially actual season-to-date winning percentages?
EDITED to add:
@DarkVictory23's methodology
described above is based on the same kind of thinking as
BP's "adjusted standings."