I read half of this post and checked fastly all of it.
About the elo and the % of victory with the difference of elo, I find the k-factor great for games with 2/5 luck.
For games with high luck, I think the k-factor is too high compared to the reality (I proposed 1.5 years ago to change it there:
https://boardgamearena.com/forum/viewtopic.php?p=186553 ). If you was reducing the k-factor for those games, the swing of elo would be lowers, so you would find great players close to the top more often, and the top players are intencive to play against random players instead of top players.
.
About the initial post and cooperatives games, I wroted around 25% of the guides on them.
You get two type of players: players who plays a specific game a lot that I will name 'hardcore', and players more 'casual' which play one game of that game from time to time.
The problem is that the amount of games played by casual players are much lower than the amount of games played by hardcore players (it's what I think, I have zero proof of that line).
But so, to find casual games, you will need to wait a lot more.
Hardcore players will have done 100+ games, sometimes 1000+ depending of the coop game. They have played multiple time with new players, and are in average better than the average player.
At the same time, myself included, I like going deep in games, I play way less versus games because doing things against others use more of my energy or make me more angry. And so I play mostly advanced difficulties/advanced tables which usually are restricted to at least 100+ elo.
To take the crew as an example, games which are in deep sea and missions 28+ (close to hardest settings) I usually is much more talkative than games with lower elo.
And I could be wrong but I don't think I would be the only one.
Strangly enough, I think people which are more toxic don't play those settings, because they could be angry in chat, get some strong logical answers, and would double down which lead to a very bad experience for everyone even themselves.(it's linked with IQ too, the more IQ you have, the better and the more friendly you should be in average, even if it's extremely simplified).
And so they would goes in lower/medium tables where they can feel great, and be toxic against others players.
I insist, those last sentences are what I think, I could be wrong.
You can see one short post about it here:
https://boardgamearena.com/forum/viewtopic.php?t=43310 .
Hanabi has the most conventions by a large margin.
The crew have some, which are explained in 15 lines here:
https://en.doc.boardgamearena.com/Tips_thecrew .
Outside of those two, there is some games which can be toxic but that's mostly because they are hard, and a single players can totally make the game loss (some players wouldn't care, but some others would).
Personnally Yokai and Burgle bros got that feeling on me, where playing with bad players make me not enjoy the game at all. While most others are still manageable, and I still find fun playing them/reparing mistakes somehow.
.
Looking at your profile, you talked about 'coop games', but only played Hanabi which in fact is extremely weird, you should have said Hanabi directly
About Hanabi, if you play without conventions but have played the game a lot, after someone give an hint, the question became 'does that player make a mistake? Is it the commun interpretation? Is it the uncommon one that 3% of players does?'. And answering those questions just is a roll of dice.
The thing with IRL is that you play with the same group of players multiples times, so you know how everyone react, you can create conventions conscientiously (by telling, from now on we will do X in that situation) or unconsciously (the last two times we played player X did that, so from now on, I will assume in that spot we will do that). And it will work well.
When you play with random players, you will not know what happen in the previous games, and so need either:
- to restart from scratch all your predictions,
- or start to play with uneficient conventions and work the best you can with it (that has been done, and that play pretty differently)(but still after a while, the more you play, the more you optimise, and the conventions arises. You either add them, or you still split the differents cases, to the point where you got two players with two strong reasonnings but which lead with differents interpretations (per example in situation X do you saves 3 & 4 in chop?) which are both good, but which lead to bomb and losses),
- or note the conventions the community agrees on, and everyone agrees to play with those 'rules'.
You can't recreate your unique experience IRL except if you play a lot with the same group of people. And even still if you try to win, after 50 or 100 games, you will have problems arising that I tried to explain before.
. You keep making up a lot of things.