Boards›Gameplay

Matchmaking based on winrate

gottmilk101·2/25/2020, 8:21:06 PM·1 votes·2,400 views

I have been playing this game for about five years now and it seems that matchmaking continues to stay in a really bad state especially in low elo. I know this wouldn't be a perfect change, but I think it would also combat smurfing. If you change the matchmaking to fit players winrates instead of only their mmr/rank. It adds a third variable and would likely slow queue timers a bit but I think it would be healthier. I have played close to 100 games this season and almost always a player with a sub 40% winrate is one of the two worst players in the game. I can almost immediately predict the outcome of the game when I see immense winrate differentials on opgg. Team with 52%, 50%, 49%, 52%, 50% vs. 50%, 54%, 38%, 49%, 49% guess which team is going to win. I did the math on my last 25 games and the team that averages a higher winrate has a 72% chance of winning (18/25). That is an immense difference. It feels like there should be a system that accounts for this?

6 Comments

Pumpedhero2/25/2020, 8:54:23 PM1 votes

All this does is make it so stuck players stay stuck, and players that are boosting / boosted are queued with others that do the same or are smurfing etc.

I Have No Future2/25/2020, 9:18:53 PM1 votes

72%? I'd be fascinated to see some more rigorous stats (because it's honestly probably more like 80% ha, with maybe 10% of the rest going to afk's and tilted autofillers, and that last 10%?.... yup that's right there is your actual chance to win when faced with it - gl gg ez). I also liked the suggestion others have had to somehow account for a player's overall experience differential in matchmaking, which is related but perhaps more .... delicate to determine.

Kai Guy2/25/2020, 10:05:20 PM1 votes

If you try to balance teams by WR for each team. You can end up punishing good players. Compensating an 80% WR with a 20% teammate is not going to be pleasant.... then ultimately it implodes because WR are artificial due to Bias in MM.

So that's a no go yea?

WR alone for MM keeps a lot of weight on previous games. Changing WR with 10 played vs 100 for example. Every game after 100 is less then 1% of total WR.

This is a pretty good article to check out .

Subdue2/26/2020, 2:49:47 AM1 votes

Please share your data where you tracked the win rates of each player on each team for 25 games.