"The same thing goes for berserking games consistently - your rating will likely lower (probably not 200 points IMO), but if it's done consistently, it will represent your true online chess strength, because you do berserk games."
Wrong, wrong, wrong.
Let's keep things simple. Say you have a player A who performs at 2100 level with normal time and at 1900 level with berserking, and this player berserks half the time getting a rating around 2000.
Whenever our player A plays, he will distort the ratings of other players.
Simply put, players using berserking ends up underrated when they actually use their full time but overrated when they click that berserk button. If player B beats player A using full time, player B will receive less rating point than he deserves. If he loses, he will lose more than he deserves. And the other way around, a player which get to play player A when he berserks will either receive too much rating or lose too little.
Thus, the distortion of ratings provoked by regular berserking is unfair (this is a fact, not an opinion).
I play an online game which allows to use handicap in rated games. The rating system includes a built-in compensation to limit such distortion.
Lichess should analyze data to see how berserking actually affects elo performance of players, then add some compensation system to limit distortions.
The rating system could use some other improvements too, variations are usually too large.
Wrong, wrong, wrong.
Let's keep things simple. Say you have a player A who performs at 2100 level with normal time and at 1900 level with berserking, and this player berserks half the time getting a rating around 2000.
Whenever our player A plays, he will distort the ratings of other players.
Simply put, players using berserking ends up underrated when they actually use their full time but overrated when they click that berserk button. If player B beats player A using full time, player B will receive less rating point than he deserves. If he loses, he will lose more than he deserves. And the other way around, a player which get to play player A when he berserks will either receive too much rating or lose too little.
Thus, the distortion of ratings provoked by regular berserking is unfair (this is a fact, not an opinion).
I play an online game which allows to use handicap in rated games. The rating system includes a built-in compensation to limit such distortion.
Lichess should analyze data to see how berserking actually affects elo performance of players, then add some compensation system to limit distortions.
The rating system could use some other improvements too, variations are usually too large.