Lead designer of Overwatch confirms that some classes are not equally ranked in competitive.
http://us.battle.net/forums/en/overwatch/topic/20748606456?page=16#post-308
-Jeff Kaplan
Thanks for the input, guys. This is really helpful. A lot of what you're saying aligns with what we're seeing.
We've made so many changes since Season 1 that the data from that Season is not relevant anymore. Also, we made significant changes after the placement reset on the PTR. So our sample size is limited in evaluating this. It's also somewhat tainted because much of the feedback we get is either Season 1 holdover feedback or pre-"placement reset" Season 2 feedback. So trying to separate out all of the old issues from possible current new ones is challenging.
What it looks like to us is:
--Zen places fine
--Ana places fine
--Mercy places low (fix in flight)
--Lucio places really low (fix in flight)
There were other heroes (non-support) who were placing too high and too low... and there are fixes in flight for these as well. This should further help to balance things out... hopefully. We'll keep a very close eye on this and continue to make improvements as needed.
On reddit they talk about what stats are the best to work on to improve your ranking most. It turns out that accuracy is highly valued. This means not just healers but spammy characters tend to get low scores. For example Junkrat is also ranking very low right now as well. He typically has very low stats for accuracy after all.
Which is interesting because it turns out that Tobjorn is actually one of the best characters to use for ranking up with. At that point I learn of this guy:
Overbuff: https://www.overbuff.com/players/pc/Torbj%C3%B6rn-11130/heroes/torbjorn
Twitch:
He is a rank 80 Torbjorn main who has never played another character other than Torbjorn. A Torbjorn main who is regularly ranked with the best players in the world. Watching him play makes me seriously wonder if I should be playing more Torbjorn myself. :D
It's hard for them to balance some of these characters, because a primary focus of theirs will always be making the game (and ALL characters, to some extent) accessible to new players. This will always result in certain heroes having abilities which are useful in the hands of a novice and overpowered in the hands of an expert.
Still I think they do a great job overall with the balance of the game, and I look forward to any future updates aimed at improving this. Good post!
The problem is that every character shouldn't be viable for every situation. There should be times where defensive characters are the only viable characters to hold on a map. Just because something works well most of the time, it should have maps that are weak to use it on or counter picks. Pidgeon holing a game with this many diverse characters with drastically different play styles into one or two core meta builds is annoying as hell to me. I am glad pro players are using junkrat and mae and other characters that have negative views from the community. Inherently, any ranking system will have it's fall backs and I still hold to my thoughts that a good junkrat will have fairly low accuracy due to prefire and predicted fire. It's the only class that I feel like gets half of their viability from holding back enemies by angle shots and manipulation of projectile bounce. When playing junkrat I am firing almost constantly if I have any idea where an enemy or group push is coming from or going. Close range burst he is a monster as well.
This isn't really about that, the issue is that if you play Junkrat in comp you lose out on MMR if we win and you lose even more MMR if we lose. Same for mercy and Lucio. Due to the fact that accuracy is a top stat that skews everything right now Torbjorn is getting free MMR right now.
The bottom line is, their system has been unfairly fucking many different characters. Right now the only healers not being fucked by MMR is Zen and Ana. It also means that spamming Reinhardt shields correlates to gaining more MMR if we win a match.
Its pretty messed up lol
Interesting, that would explain some things. I think they are going to have to adjust weighted metrics based on the character.