I don’t think it’s been the same problem over the years. The original team had the unfortunate habit of implementing sensible sounding ideas that tried to “encourage” the playerbase to play the game how they were “supposed to”.
I think that most of the HISTORICAL balance issues come from an inherent conflict between
-
what the development team WANTED the game to be
-
what the players thought the game was SUPPOSED to be
-
what the game REALLY IS when you sit down and play it
Naturally, these well-intentioned changes were usually the MOST disastrous additions to the game, and led to things like brig (let’s make a support that’s fun to play and breaks the stagnant dive meta!) double-shield (let’s make a new tank that has a barrier you can move around because it standing in one spot holding up a shield isn’t fun) and the slow creep of healspam (let’s make healing your team something you can do quickly so you get to focus on the more fun parts of playing support).
We don’t have that problem so much anymore. Why? No more Jeff Kaplan. Jeff was definitely the most… ambitious… part of the team when it came to adding something completely new and different.
Don’t get me wrong, I think it’s better to have a head guy who wants to keep trying new things, even though some of them end up creating a huge problem(s). You won’t come up with any great ideas if you aren’t willing to try some terrible ones. Why? Because you can’t actually know which ones are great or which are terrible until you do! The best ideas might sound bad, and the WORST ideas always sound reasonable and seem to make sense before you try it.
I think they (the current development team) has a bad track record of relying far too heavily on statistical data mining and also including too much decision making based on short-term trends from top-tier gameplay.
This is what happens when math majors are exposed to the reality of thousands of people (ie a typical game’s playerbase) who complain about EVERYTHING and have absolutely no interest in differentiating between their own personal experience with the game and objective reality. A quantitative thinker exposed to this reality will quickly decide that all opinions are inherently unreliabe and therefore equally useless, thus, the only useful source of information must come from hard data like statistics.
Statistics are great for helping you measure things. Statistics can HELP you determine whether a problem exists, whether evidence is reliable, etc etc
When used as a secondary source in support of observable patterns/phenomena, statistical data is hands down the most reliable and dependable form of information you can have to determine the validity of a claim.
BUT
When used as a PRIMARY source of evidence, or worse, when used to DEVELOP a hypothesis instead of SUPPORT one, pure statistical data is REALLY FREAKIN’ BAD at providing any kind of accurate or useful information. It is entirely unreliable if your goal is to interpret what is ACTUALLY happening on that sheet of data and it is EVEN WORSE at helping you understand the overall system which generates those statistics.
I think until they stop being afraid of causing the next “double-shield meta” that somebody is willing to try a genuinely outside-the-box-type-thinking solution for some of the longstanding issues the game has, they aren’t gonna end up doing much more than tweaking numbers back and forth.