This week, an Overwatch developer made a post about how the community - specifically in high elo - doesn't fully understand "power creep," stating their complaints about current balancing issues are off base. And while it may have just been a random "fun fact," most pros and high elo players did not seem to think it was very fun.
Josh was addressing a concern that newer heroes are overtuned and some of the lesser played/older heroes are too weak in the current gamestate. This type of phenomenon is reffered to as power creep, and plagues most, if not all, constantly updated games like Overwatch. New content is often only interesting because it brings a new mechanic or power level that old content doesn't provide. But too much of a good thing can be a bad thing.
In his statement, Josh claimed that the players' frustration was off, and that the general heroes players complained about were balanced fairly. He used Genji and Soldier: 76 as examples, but only highlighted their winrate to back up his claim. The problem is, those heroes are almost never played, so the winrate is inflated from one-tricks, perfect situations, and times where players picked them at the end of the game when they were already going to win anyway.
After a bunch of comments from OWL pros and staff and other high elo players, Josh returned to clarify his statement, including more thought on pickrate and other factors like, "Is this fun to play against?" He also pointed out that "one of the perceived strongest and one of the perceived weakest heroes have nearly identical pick rate and win rate." Finally, he also clarified that the devs would be making changes to some of the heroes people most recently have been complaining about, specifically Baptiste and Mei.
The final point is that while players have a lot of knowledge about the game and what feels strong or weak, they don't have as much data or experience to know what decisions devs need to make to "fix" a game. Josh acknowledges the difference between the ranked ladder and pro play as well, but states that it's likely that off-meta champs have more viability than current GM players give credit.