Wintrading in comp needs to be fixed

General Discussion
1

So let’s say I’m 4000SR. I win 1 game, lose 1. Where should I be at afterwards? I should still be 4000 right? Well that’s not the case in most scenarios, normally you will be lower than you were before. And this is wrong.

Performance based SR was removed a long time ago for Diamond rank and above, so it only stands to reason that my SR should remain unchanged.

However this is not the case, and what normally happens is your SR will slowly decrease if your wins are going back and forth in a play session.

2

Going 1 for 1 does not keep you at 4k. You have to have over a 50% win rate. And it makes sense, if they dont want ppl to be sitting at 5k cap. Even when performance based SR was still around, the only ppl really sitting at 4k+ with 50% win rate or less played off meta heroes or mercy.

3

The number of people who have actually reached 5k before can be counted on one hand. That’s not a problem here at all.

4

It’s like any elo system. If your team has a higher average mmr you gain less off a win then you lose off a loss. That said you should have a roughly equal amount of games where your teams mmr is lower than the enemy team’s and you gain more SR from them then you would lose.

At GM (4200+) you would expect to lose more on average from a loss than you would gain from a win. This is the same reason gm hero’s average win rate is a decent chunk above 50%. Because you’re high enough that the match maker can no longer be perfectly matched and you more frequently end up on the team with the average higher SR. An extreme version of this can be seen in top 100 players. Watch any of them stream and they typically drop far harder in position than they get back when they win. This is because frequently their side has higher average mmr.

If you really think you get less from a win than a loss and you aren’t like 3800+ you can do the following:
Record SR of each win and record SR of each loss. Should be every game you play for some period of time. 30+ games would be better.
Average the sr gain and the sr loss.
I’d be interested/surprised to hear if the averages aren’t within one point. If you do this hit us with an update pls :slight_smile:

1 Like
5

Im saying it would be if a 51% winrate meant they continue to climb.