I think the issue is that once it’s implemented and becomes meta in a sense, then it can be exploited in some way.
For instance if you look at how people feel about the “recently timed out of a game tag” "has recently timed out of a game" -- uh .. does years ago count?, it’s clear some people don’t like being labelled, especially if labelled incorrectly. Some people accepting correspondence games probably do look at profiles to see who they’re playing and whether this person is likely to timeout out by checking the label.
One can imagine a similar thing, that you might look at an opponents ‘manner’ label/points/badges to decide whether you want to accept a game with them. The exploit I’m imagining is that if there’s an option to upvote, there’s possibly one to downvote, meaning that even if they were friendly you might just downvote them because you’re annoyed, lost the game maybe, or even if you won but didn’t enjoy the game, the opponents style/opening etc.
Maybe then you don’t give the option to downvote, but you can just as well view a lack of upvotes relative to the number of games played as a similar indicator to downvotes (why didn’t more people upvote this player?)
It’s just something to keep in mind when introducing a new feature like this.