I know that most of you know this, but I will reiterate...
Making games is hard. Making complex RPGs is even harder. Take it from a person that has worked on a few different styles of games, these sprawling RPGs are insanely complex in comparison. Systems are weaved into a bunch of other systems and even the tiniest change can cause a wave of instability throughout the game.
The thing that lots of people forget is that a game is a special piece of software. What I mean is that we can plan out a game completely - all of the design, art, programming, audio, etc. - and it can be executed perfectly... and the game may still be horrible. Proper planning and designing are essential in making great games, but it won't get you there alone. Lots of our ideas sound great on paper, but in practice they fall flat. Making great games require myriad adjustments and changes.
Sometimes these changes are minor and don't require adjustments to a previously implemented system. Those are the best kind of changes. Other times you need to make sweeping changes to a system that have long term, rippling repercussions. We try to avoid those types of changes, but sometimes it is the only way to make something fun. As a real world example, there were some classes in PoE that went through some pretty drastic changes from their original designs. This happened because, once we started using the classes in an actual game environment, we found they weren't as fun or useful as we wanted. We could have left the original designs in place (and maybe it would have given us a more polished experience), but it might have made a game that wasn't as fun.
All of this isn't to say that the project's management team is without blame. Far from it. None of these decisions are made in a vacuum and for every planned feature we change, we should be taking it out of something else - either a time buffer or another feature (or, if you are lucky, you can get more budget/time). That means if we want to revamp one of our classes, the responsible thing to do would be to see how much it will cost (in resources and time) to implement these changes and pull that from something else. It is a constant push and pull and I will be the first to admit that I can always do better in this department.
Overall, I was very pleased with what we accomplished on the limited budget and resources. Nowadays games of this size have budgets anywhere from three to ten times what our team was working with.
The important thing is to take the lessons learned from PoE and move it forward into our future projects. Much easier said than done, though.