Dean "Iksar" Ayala Talks About Battlegrounds - Ranked Resets, Party MMR & More
Earlier today, Dean "Iksar" Ayala shared some insights on Reddit about Battleground parties, ranked resets, and more.
There will be a rating reset in a patch 18.4 and they will reset again every major Battlegrounds patch.
We have two core goals directly relating to group queuing. One is to allow you to play with your friends when you want. The other is to create a fun and fair experience for all players involved in a game. Ideally, there is a solution that allows for both of these things to happen.
There are a bunch of potential solutions to solve group queue issues relating to competitive integrity. The TL: DR is there is some data collection involved in understanding the kind of advantage group players have over solo players so we can accurately matchmake and rating adjust. Once we have an accurate baseline of the advantage different skillsets of players get (4k, 6k, 10k, etc) by grouping together, we can adjust the matchmaking rating of those grouped players to create a fair experience. There are many more details about how the end of game rating calculation would work, but the general idea is that whether you queue up into a game with grouped players or solo players your expected rating adjustment end of the game would be the same.
If this doesn't work for whatever reason (actual or perceived fairness), there are many other paths we can go down. We can disable group queue, disable group queue about a certain MMR, only pair groups together, create a separate queue, etc, etc. All of these solutions involve some amount of downside that we'd like to avoid if there is an alternate solution with minimal downside.
Beyond 18.2, we have plans for a rating reset in 18.4 and we will reset again every major battlegrounds patch. Rating reset will also likely involve some iteration based on player feedback, so we'll continue to adjust until players feel like we're in a good place. In the interim, hope you are enjoying BG and you get really big pogos.
Matchmaking Rating: Your rating that we match you with Visual Rating: Your rating you see on your account
When a reset happens your visual rating will go to 0 and your matchmaking rating will stay the same. You will progress from 0 for 30-60 games until your matchmaking rating and your visual rating are roughly aligned, at which point you will stay the same rating unless your skill level improves.
It's very similar to how the star rating reset system works in constructed Hearthstone. We feel like that system was relatively successful because hardcore users zoom through the system pretty quickly and less hardcore users get to play in a system that is mostly progression based. It can feel pretty crappy if you are a new bg player and your experience is just tanking your rating for your first X games.
Having a greater advantage to playing in a group because you do it better than other groups I would guess is true of any game you can queue with friends.
As for the rest, if you are a solo player who queues into a group, the expected result of your game placement should end up being the same. I think that the overwhelming majority of solo players will not know or care if they are playing against grouped players, but ideally in the cases that they are aware they aren't in a disadvantaged position because they are individually much higher skilled than the grouped players they are playing against. You could argue that playing against a group of 4 is unfun regardless of what the outcome of the game is, but that is somewhat subjective. Implemented appropriately, I think it does adjust for the solo player in the 4v1v1v1 scenario so I'm not sure where you are coming from.
And for the last point, we're talking about exploitation. We have some things in place to find and displace players taking advantage of the system through unfair play but I don't like to talk about these things in detail because it makes working around those solutions easier.
There are many 'what about x' scenarios that are hard to cover in one post. In this scenario, the 12k players don't face higher rated players (because they don't exist), but their rating would still be adjusted to account for being grouped.
So 12k players queue in a group of 4, we adjust their rating to (example) 14k because they are grouped. They match with a bunch of other 12k players because those are the best players. If the group of 4 players does well, they gain less rating than normal because they beat players who were 2k worse than them (14k vs 12k). If the solo players lose, they lose less because they lost to players who had a 2k matchmaking rating higher than them (12k vs 14k).
While in terms of fairness I think this math works, it can still sucks to be disadvantaged in terms of win/loss potential at the high level. These are also players who are the most aware of when they are playing against solos or groups. For this reason, we discussed potentially making group queue disabled above X MMR, but I'd like to wait and see before taking that step.
The adjustment for x% more skilled player vs group will be a calculation based on real data. It should be accurate.
There is a lot of data and research to suggest that partying in games does not offer nearly as big of an advantage as players perceive it might give, however, perception matters a lot :).