Sure, but that’s non issue, as someone with 2000 average rating will not be confused skill-wise with him, but will have also higher peak (let’s say 2200), thus there is still distinction.
Well if those examples are silly, then we can go off of peak rating no problem.
And again, I talk about players’ perception rather than what it’s supposed to be. Players mostly just check highest achievments.
Except for cases where a player got, let’s say the 2700 achievement, but 500 games played at, let’s bring back the example of 1800 rating, with 2.2k as the peak in that season.
That’ll make people not only look at that highest score of a 2700 achievement, but rather start looking at the date of that 2700 achievement, and then ask if the person has been spending their time boosting that season, or if there’s some other explanation as to why the peak is 2.2k while having only 1.8k current rating.
That’s what the ability to lose rating accomplishes. And that’s why it circles back to why asymmetrical checkpoints fits better than symmetrical.
The problem with lower bounds in rating is that if I’m 2.1k say as resto druid and start playing feral (that I really suck at), then my cr will drop to 1.8 but my mmr can drop even lower. This means that if I play resto and want to climb back, I’ll be matched against 1.6k cr people and won’t get any rating the first games. It’s also gonna make their experience really crap, like they’re fighting against a 2.1k resto but are 1.6k cr/mmr.
I strongly feel that a rating per role is the way to go. Some classes might have just one role but it is what it is. Another advantage is that it will reduce queue time (particularly for RBGs where people are just looking for tanks and heals) as well as arenas (where most people are just waiting for healers)
This isn’t solvable by anything however. Unless you come up with something completely different like rating checkpoints combined with hard boundaries (like “you cannot fight people with 200+ lower rating than you”)
You can simply lose as your main spec just to lower your rating (and boost people), drop very low and then 1600 players fight 2100 player anyway even without him changing specs.
It doesn’t solve the issue though.
Dota 2 has a similar system but with a catch. If you lose to much then you can still fall out of the bracket.
There are two sides that have to be looked into under the proposed circumstances
1)win trading
2)boosting
They both fall under the category of MMR manipulation, which is the key issue to it. But that can’t be solved without other consequences, and it isn’t any better in the current iteration of the rating system either, when compared to asymmetrical checkpoints.
But it’d just be made worse with the OP’s suggestions, so that’s why it’s pointless to do it.
Regarding win trading, what about introducing a diminishing return on the trading that is won if you play the same players consecutively or within a time range?
To sum up, the main argument that has been presented against adding “rating per role” is wintrading. Else all is good
Because the top of the ladder got so few players with their matchmaking range without needing to wait a very long time for each match for the parameters to expand over time to include the lower rated ones, which would lead to very little or no rating points lost but losing a huge chunk if lost.
Putting in a decay of rating point changes vs. the same players got too many consequences to be a practical solution to the problem.
That isn’t what was said, they won’t be trading the wins with a “rating per role”-system. They’ll manipulate the MMR to farm wins below their CR. That isn’t wintrading.
Oh gotcha! Thanks! What if the mmr is split for each role? There wont be this problem surely?
Because everyone will want one. Don’t you think every rogue, hunter, mage, warlock, and so on, will look at those different ratings and come crying out for their own?
But that’s not the only part of it. Hybrid teams for example can do the MMR manipulation to farm wins below their CR just as easy like that.
It’s not practical.
I’m not sure they will be farming to be honest: playing DPS or healer are two different worlds. Let’s take Ret/Warrior/Rdrood. If the team reaches 2.2 say, and they want to farm win at a lower mmr, the ret has to play holy and rdrood balance. I don’t think they’re gonna be farming that much. Maybe the warrior is good but playing at 2.4 as DPS or healer will not imply that this person can play at 2.4 as healer, respectively DPS.
From what I’ve observed so far in my surrounding, main healers usually play other healers as alts (look at Chas for example, he plays all the healer but never plays DPS) and DPS play other DPS.
TLDR: a lot of skill is lost when swapping roles, so games are not gonna be that easier.
I’ve played both for as long as I’ve played this game. Don’t be so quick to assume there aren’t others.
I’ve played at 2.4 as both healer and dps. Again, don’t be so quick to assume there aren’t others.
People playing on that level are very different compared to the 2.4 crowd.
Don’t get me wrong: I’m not saying there are no other people! I’m saying that if we look at the stats, it’s a minority, that’s all! 
Furthermore, for people who only played healer or only played DPS at 2.4, them upping their game in the other role to get it to 2.4 is gonna take some time.
Are you saying that the people who are gonna win-trade are 3k+ people?
But why should they get the advantage of those different role queues, when they’re able to handle the different roles?
What are you talking about now? Wintrading is done in the upper part of the MMR, when not a lot of people are playing, to make sure they end up facing each other, then trade wins to each other, and sometimes boosting someone like that.
The people who can play both well are a minority, because of the barriers of the current system.
If we have this cr per role thing, more people are gonna be able to play multiple roles as getting out of their comfort zone is not gonna punish them on their main spec.
And again, why should they get an advantage like that?
This isn’t about what good you think it can do. It’s about the consequences from it.
I think it’s not because a problem (here “not able to play other roles competitively without killing your main rating”) is only experienced by a subset of the whole userbase that it shouldn’t be tackled.
Oh ok, so you don’t care if that “minority” abuses it with MMR manipulation, ok. Gotcha.
If it’s abused, it’s just going to be for a couple weeks (their alt-spec MMR will increase) and by a very small portion of the total playerbase of players who play hybrid classes (the ones who can play very well both). That is my point. I’m not for MMR manipulation.
For most of the players, MMR drop from swapping into the alt spec correlates a lot with a big loss in skill, so they’re still matched to opponants matching their skill level
That’s mistaken. Since people only need one rating to be high, they can dump losses in pugs as much as they want inbetween those “MMR manipulation”-sessions with their other role, and then secure that low MMR for their real team over and over and over again, as much as they want.