So hard to climb arenas...

[NoelfromNY]

This is just a whine, but due to poor matchmaking, when you get toward the top of the arena, I find myself constantly matching with people far underneath me. I win, and I go up 4 pts in elo, but lose and lose 35? I went 18-7 for 72% wins, but my elo is lower than before I played those games.

Elo games should only happen when people are within 100 elo pts of each other. You could argue this is the same for everyone but is it really? This is why so many people just stop playing once they are toward the top of the arena.

[Tisaac]

This is just a whine, but due to poor matchmaking, when you get toward the top of the arena, I find myself constantly matching with people far underneath me. I win, and I go up 4 pts in elo, but lose and lose 35? I went 18-7 for 72% wins, but my elo is lower than before I played those games.

Elo games should only happen when people are within 100 elo pts of each other. You could argue this is the same for everyone but is it really? This is why so many people just stop playing once they are toward the top of the arena.

Well maybe that just means that...you know...you are not supposed to be at the top of arena for this game ?

[Romain672]

There is no good solution for that.

Your solution would in some games make you wait 1+ hours if you are top 2.

[Jellby]

Suppose you and your opponent have both your "true deserved" ELO. In the long run, it should not change (unless your abilities change, but suppose they don't). Suppose you win 90% of the matches. After 9 matches, you've won them all and got +4 ELO for each, that's 36 ELO. You lose the 10th match. That kind of "evens out" the streak, so indeed you've won 90% of the last 10 matches. Since your ELO was "right" and your abilities don't change, you should get the same ELO you had before, that means -36 ELO for that loss.

It looks unfair, but it is not. At least not from the point of view of ELO, which is measuring what it's supposed to measure. Another story is whether that's what you want it to measure. Keep in mind that your opponent has also lost -4 ELO 9 times and won +36 ELO once, for a net result of 0

[ufm]

Well, because of this astute top Arena players just sit out and wait until someone climbs up. Rating protection again.
But yeah, as others mentioned your winrate is lower than other top 4 players, hence the rating flows but does not change ranking eventually.

[NoelfromNY]

ELO is taken from chess. Which makes sense because there is no luck in chess. However, board games have luck factors. Some board games have very high ones.

Getting unlucky twice in a row is not uncommon, but means for higher ELO players, who happened to get unlucky against a much lower ELO they lost 35 each time unlike simply getting unlucky against a better player in which case you'd only lose 10. If you only matched against people closer to you, the variability would be greatly reduced. Those are just my thoughts. Again I realize the argument could be made that it affects all players the same. But that is again why people just stop playing once they've reached their spots.

[ufm]

ELO is taken from chess. Which makes sense because there is no luck in chess. However, board games have luck factors. Some board games have very high ones.

Perhaps you'd want to take a look at this... Unlikely to be used in BGA though.
https://arxiv.org/pdf/2104.05422.pdf

Getting unlucky twice in a row is not uncommon, but means for higher ELO players, who happened to get unlucky against a much lower ELO they lost 35 each time unlike simply getting unlucky against a better player in which case you'd only lose 10. If you only matched against people closer to you, the variability would be greatly reduced. Those are just my thoughts. Again I realize the argument could be made that it affects all players the same. But that is again why people just stop playing once they've reached their spots.

That's why some sports and games abandoned ELO altogether and adopted something else (leagues, tournaments, pro points, etc.).