Also, to help show how close you are to your opponents, the average skill rating of the match will now be displayed at the start of the game.
The problem with match maker is the one or two players that are drastically under/over the average SR/MMR. Everyone knows SR doesnt mean crap anymore, showing the average skill is not going to fix matches feeling like forced wins/losses.
SR is also being change to match your mmr at the start of the season so the sr dispays should be accurate. We will see. Everyoneâs sr should recalibrate at the start of the season to their current mmr
where did you hear this, do you have a source you can share? All Iâve heard is they stop SR decay from season to season.
SR is already inflated for some people, if they were to truly do this then people would go dropping in SR to match their MMR. Can you imagine the uproar from that?
So theyâre just correcting the seasonal decay for players that did not want to grind. This does not mean it will fix those dps players in master lobbies that go 5 and 20 while spamming the tank. these HANDICAP players were put in that match because of their MMR, not their SR!
As long as the role mirroring is used instead of making sure everyone is within a reasonable amount of hidden MMR, the matchmaking will be bad. There is nothing competitive about placing diamond players with gold players with the reasoning being that both teams have diamond players role mirroring so it is fair.
They are saying this, but I feel its actually to put people straight to where mmr thinks their rank should be.
If this is true, confidence in mm will improve.
I think Iâm in the minority, but Iâve seen a gradual improvement in matchmaker over the course of season 3. Subtle things, not enough to stop the rolls, but movement in the right direction. Season 4 might be enough to move the boosted and troughed characters to where they belong, creating better matches overall.
This is a very succinct statement of the problem with the pathological teams the matchmaker routinely forms. (Though it would also be interesting to know how many of those pathological teams might be mirrors of large groups with imbalanced internal compositions.)
Someone needs to remind those folks one of the first things they teach in introductory statistics: When you summarize data, you should report the mean and the variance.
I know I am awful for saying otherwise, and Iâm not going to tell people it is working great.
I am just more critical of the system, and have been since the start. There are certain predictions of skill equality and inequality Iâm not seeing anymore. Big win/loss streaks arenât significant now, things like that. Iâm hopeful by the end of season 4 this will be more widespread as mmr and mm confidence improves to the new system.
But, I would still prefer no hidden metrics.
I mean in the algorithm being more accurate, not opinion. It has been ages since I touched the mathematical side of anything like this, so I am probably using the wrong word.
If theyâre not posting their code, you shouldnât trust itâŚat all.
Like in what world does code just go live and we canât see it on git and audit/comment? Oh right in the fake/rigged software for cash world. Where they just hand out pixels to hype and gaslight, and the entire experience is tailored.
Thatâs what youâre going to get here with the MM. Fake/ded/rigged ft. patents ⢠for engagement and monetization ft. a game that is entirely centered around engagement and monetization. The kind of world that puts buzzvision first.
Until they show the code,
there isnât a real/fair/explainable/transparent MM that has integrity,
and is legit for gamers and esports.
This far into their devtime? They need to show the codebase.
They had a chance to reveal a working product.
This hasnât been the case since mid-OW1.
They have to PROVE it works, line by line.