LuxuryJesper wrote: ↑25 May 2020, 23:35
Let me point out, that it was the user Jest Phulin, who tried to compare a game on BGA to a league (which is ridiculous). I didn't bring it up, I only responded to his comment.
It doesn't make any sense, that one player benefits from a draw. Otherwise, it basically isn't a draw. Then (as I wrote previously) you might as well say, that the weaker player won a draw(!). How that makes any sense to you, I have no idea...
As I said, the comments would be about how the team expected to lose did well, or how the team expected to win did poorly.
I know you mentioned you don't like the word "expected" in this situation, but it is appropriate. If Manchester United played my local high school team, I'm pretty sure people would say that one team was expected to win.
League match play is based on all teams playing each other (or at least all teams with a pool playing each other, and the best from the pool advancing to a pool containing members from several pools). ELO is a different system. ELO needs to compare A and B, when they never played each other but they did each play against C, D, and E. If A wins against C, D, and E, where B loses against them, it is reasonable to assume that A would beat B. Therefore A would be ranked higher.
Now, as sprockkitz keeps pointing out, if A and B did play each other and tied, their rankings should move closer to each other. This can be done one of three ways. If B's ranking was raised significantly, they would benefit greatly although they did not win a game. If A's ranking was dropped significantly, they would suffer although they did not lose a game. So, A is dropped moderately and B is raised moderately.