Taken from this reddit post: https://www.reddit.com/r/gamedesign/s/bNQIC5H2TG

Short answer: Yes, it’s impossible for perfect balance, but you strive for “best in the moment” and ideally back it up with tools to make testing for that a lot easier

Long answer: soooooo, I’m a very old school gamer and my main multiplayer game genre was MUDs – these were the text-based precursors to MMOs, and these games still have some concentrated player bases in a few corporate companies (IRE and Simultronics). I was also one of the “big” players, which basically translated to free dev work on my behalf.

I work in QA and so I would run the initial QA pass for balance changes. I would assess to find the potential OP combos, edge cases, unexpected interactions. This was all done on a test server which mirrored attributes of live and also let us custom create characters for the specific thing we were testing.

From there, changes went to the “liaison group” which was an IG role where the top fighters from every class basically had a combat council to discuss the game meta. We had a pocket realm we could use in game where we could teleport in, adjust stats, equip everything in the game, etc. This let us easily setup situations for testing the crazy combos we dreamed up – instead of just mental/pen and paper we could tweak stuff to test specific circumstances.

Every few months, we’d also suggest combat changes. Normal players could also submit ideas, which we could support for higher level discussion and review. We’d review all the submitted ideas and comment on them. We used both asynchronous communication (comments on submissions) and synchronous communication (chatting in a chat channel in game).

This system had its good/bad.

Good:

– the people discussing balance knew the actual gameplay impacts incredibly well.

– the suggestions for potential solutions were made by people taking into account combat balance, lore, theme, etc

– crazy math at the drop of a hat to prove points

– directed and edited pipeline to devs

Bad:

– selection process for these players wasn’t always great. In one game, for example, it was a guild role, so I became liaison for a dead guild simply because I was the only person who fought…even though I had been playing less than a month

– personal bias: players were always loathe to nerf their own class, even if they knew it was OP and nerf/buff selection sometimes ended up quite political. We’ll agree to x nerf if you agree to x buff. I became briefly both hated (by allies) and beloved (by enemies) because I submitted an idea to fix a way my class was OP

– admin tension: This role/communication channel was really in touch with admins and if an admin didn’t like a player (hella common in MUDs) everything broke down

– in-game harassment of liaisons: this was common enough that I need to call it out. People would harass to try to engineer the results they wanted.

Sorry, rambled a ton, but I think there are maybe some helpful takeaways there!

I can talk more about this if you have any questions! When I worked on Age of Empires, for example, we had one of the top ranked players in the world as part of the test team – his role was just to try to APM break shit and find basically skill ceilings to help the devs balance downwards.

The concept of balancing downwards is something that I think can really help game dev – it establishes a maximum QA boundary and then all testing never has to deal with that max limit! If the best player at your game says xyz is max potential combo, you don’t have to edge case all these weird potentials!