I imagine the thinking here was "We have 50 servers, let's split them evenly, and the playerbase will distribute itself."
Stupid? Well, in hindsight, yes, it probably was stpid to do it like that. But they can't just convert some of those JP servers to NA/EU (because then they wouldn't be NA/EU servers, they'd still be JP servers).
I think, in hindsight, spreading/splitting the servers like this was stupid, as it created this weird unproportional demand of activity.
This has nothing to do with the "paying players" -- it's just network architecture, and hedging bets on where the population will lie. They bet, and they lost. It's not something they can fix with a wink of an eye like most people think, especially if they're thinking long term ("oh hey, we bought all these servers, and now a ton of people quit, and now these servers are empty and people are bitching because no one is one")
That's a THING? That would definitely be an unplanned, large spike in traffic if you suddenly granted a ton of people free access to a hot new MMORPG
Yeah, the Japanese servers are physically located in Japan. You can't just make them, NOT Japanese servers in the blink of an eye. The latency caused by distance would still be there, and the fact that most players are speaking Japanese would remain.
Additionally, while they have some capacity to add more servers, in the short term this wouldn't do much good. Too many players are invested in their characters now, they won't re-roll on a new server easily. That's putting aside the fact that Square has no idea whether or not the player base will remain in place after two or three weeks. A number of players probably won't even stick around after the free period after all.
No, the real solution is do things like implement a proper queuing system, and a mechanism to auto-kick idle players. That'll solve 2/3rds the problems in the short term, and in the long term the gradual reduction in traffic would deal with the rest.