[BostonCard]

I actually follow his logic.  He's saying Oregon is the 4th best team in the country, in his eyes, as they played his #1 team in the country very tough.  I don't agree with his conclusion, but I get it.

I don't; it's circular logic.  The reason he made MSU his #1 team is that they beat Oregon.  It would be one thing if he elevated Oregon for losing a tight game against a team he already had ranked #1, but that's not what happened; he increased the ranking of both teams.

By definition, only one of the two teams could have exceeded his expectation.  If he expected MSU to win big and Oregon kept it closer than he expected, fine, raise Oregon, but then he shouldn't have made Michigan State #1.  If he expected Oregon to win, then by winning MSU exceeded expectations and it's perfectly fine for them to be #1, but then don't also raise Oregon's ranking (unless teams ahead of Oregon embarrassed themselves, which they didn't).

It's exactly this sort of circular logic that lets him rank the SEC so highly when they beat up on each other.  If SEC team A wins, then he raises their ranking because they beat a tough SEC opponent.  But then he doesn't penalize the losing team, because they kept it close against a highly ranked SEC team.

Roughly speaking, if one team exceeds expectations, then the other one had to have underperformed.  That is the logic behind the ELO system in chess (and that 538 applies to the NFL).  Points given to the winner of a match are deducted from the loser, in rough proportion to the degree of surprise of the result; thus a win over a good team gives the winner more points than a win over a bad team.

BC

[BostonCard]

I actually follow his logic.  He's saying Oregon is the 4th best team in the country, in his eyes, as they played his #1 team in the country very tough.  I don't agree with his conclusion, but I get it.

I don't; it's circular logic.  The reason he made MSU his #1 team is that they beat Oregon.  It would be one thing if he elevated Oregon for losing a tight game against a team he already had ranked #1, but that's not what happened; he increased the ranking of both teams.

By definition, only one of the two teams could have exceeded his expectation.  If he expected MSU to win big and Oregon kept it closer than he expected, fine, raise Oregon, but then he shouldn't have made Michigan State #1.  If he expected Oregon to win, then by winning MSU exceeded expectations and it's perfectly fine for them to be #1, but then don't also raise Oregon's ranking (unless teams ahead of Oregon embarrassed themselves, which they didn't).

It's exactly this sort of circular logic that lets him rank the SEC so highly when they beat up on each other.  If SEC team A wins, then he raises their ranking because they beat a tough SEC opponent.  But then he doesn't penalize the losing team, because they kept it close against a highly ranked SEC team.

Roughly speaking, if one team exceeds expectations, then the other one had to have underperformed.  That is the logic behind the ELO system in chess (and that 538 applies to the NFL).  Points given to the winner of a match are deducted from the loser, in rough proportion to the degree of surprise of the result; thus a win over a good team gives the winner more points than a win over a bad team.

BC