What approach do people take towards playtesting and balancing? RPGs can be long and involved, so there's a lot to consider, and soliciting feedback on such potentially large projects isn't necessarily feasible.
Viewing post in Let's talk about making RPGs
Thank you for reminding me this thread exists, I’d meant to write in it but forgot!
I haven’t gotten to this point in my game yet, but I do have some theoretical ideas for this.
The first is to build the entire game’s combat encounters very early in the process (with placeholders ofc, basically just fighting rectangles with mechanics), essentially bypassing all of the writing and level design to have a ‘complete’ game that’s just a straight corridor of encounters (and then just gradually replacing segments with real game content as the project progresses). The idea I had there was to let me easily work out how to adjust things like stat gains and xp curves on a macro scale, since I’m building a game with a job system where early game choices can have a real impact on later parts.
Another thought was spreadsheets and automation. My game’s assets and formulas are deliberately kept separate from the engine so that I can setup a commandline script that, say, performs every attack in the game on a dummy 1000 times or whatever locally or from my CI/CD pipeline. From there I can probably generate some spreadsheets and graphs to visualize how statistics, balance tweaks, status effects, gear etc. impact how abilities perform in combat more easily.
I think these two points will help me get a feel for balance early on, though I’m definitely interested in seeing other approaches and suggestions for inspiration.
As for playtesting, I don’t have many ideas….. given how my game’s progression gets it’s going to be a bit hellish no matter how I slice it so I’m probably just going to try finding some playtesters and pray :D
Interesting ideas. I use spreadsheets for some crude manual forecasting, but I hadn't thought of generating reports like that. That would give you a lot of data to identify the pain points.
Outside of raw numbers though, I wonder how to get feedback on content and design. For my previous game, I was able to get a good bit of feedback to improve the early game, but relatively little beyond that. And that's fair enough, because 30m of peoples' time is already a lot to ask for, but it does leave me uncertain about what (if anything) I'm doing well and what needs to improve.