Can anyone explain math behind win and loss points.
I play against lvl 59. I win I get 1.4 -3.6 ponts for map. When I loose I also loose 10 points for map. This is not I discovered recently. 1 day I play 10 games, 8 win and 2 loss. And I drop best points list 10 places.
I try to figure out math behind it and I dont see any logic.
He's talking about QM. Level of opponent doesn't matter. QM rating of opponent matters. Your rating is 802.47, mine for example is 210.
If you beat me, you won't gain very much because your rating is much higher than mine. If I beat you I will gain 10. If I lose to you I won't lose very much because the system expects me to lose to you.
The problem is, that to a large degree ratings are based on activity rather than skill. It expects me to lose to you because my rating is much lower than yours. But the only reason my rating is lower is because I have played fewer games.
I'm 21-0 on quickmatch, max gain of 10 per game, thus 210 rating. Despite this, someone who wins 50% of their games, but has a lot more total games, and thus a much higher rating, will get punished severely for losing to me.
So I am 50th place. I have 600 pints. First place has 1200 points. No big deal if its reachable, but..... its autodis. If auto dis. gives "nice" start and opponent aint very stupit, then u probably loose points. Therefor the first 2 places that are around 1200 points, they are untouchable. U cant reach if win gives 2.3 pint and loss takes 10 points.
Of course it is reachable. In fact, it was arguably even harder before as the avg rating was lower. However, yes it is quite a grind and not one I would recommend :) At the top, QM rating is more about being active and playing a lot (tons!) of games, then being unbeatable.
EDIT: There is also a different leaderboard for each of the ratings. If you want a skill based rating, then there are the ladders to focus on.
lmao at recommending the ladder in a topic about confusing rating systems. is there's some reason we can't just use Elo or Glicko? these are standard for a reason...