On
this one point, we actually agree. There's huge value to PTRs -- or even to having a select group of players who have access to early builds and a single test world, if doing a full PTR that works for all regions is problematic. (Like the Evocati with Star Citizen, for instance, much as I dislike that system.)
I know some people scoff and are like "oh, that's just outsourcing the testing" and I suppose you can look at it that way, but:
- Software QA in general -- and game development QA in particular -- is not well-suited to really thorough testing, given how it's usually handled.
- Live environments and development environments vary greatly in online gaming; a development server is handling a much lower load, and is thus far less likely to see errors like I/O bottlenecks or database locking race conditions. (See also: housing placards not receiving lottery data.)
- Players will generally try weird things you never imagined anyone would try. (Is it rational for someone to try hitting a wall 50 times in Elden Ring to see if it broke? Probably not really. Did someone do it anyway? Yep.)
Making a test realm that's wiped frequently (but lets you copy your real character over after a wipe, to ensure people can test endgame content and whatnot) and letting players loose on it really, definitively will help a team find bugs triggered in ways they'd
never consider in order to put into a QA plan.
Mind you, even a PTR won't catch 100% of bugs; for instance, a PTR probably would also have a low load compared to a real live world, so even testing the housing lottery
there might not have had the situation arise that prevented placards from receiving data, and still let that bug slip through. But even if it's not 100% of bugs, a PTR will
absolutely catch more bugs than purely internal testing will, and finding any non-zero number of additional bugs -- yes, there can be
huge value in that.