In my opinion, the market sustainable model for MMORPGs will continue to be pay-to-play at launch and subsequent free-to-play with optional paid features (i.e. “freemium”) once the initial excitement has died down and the customer base starts to shrink. Much like real estate rental, freemium produces a modest but solid revenue stream which can support the infrastructure costs inherent to MMORPGs.
World of Warcraft is the major counter-example to this business model; WOW has remained pay-to-play since its initial launch, releasing a series of major expansion packs to the game as a way of keeping customers excited and engaged enough to continue paying for monthly subscriptions. WOW certainly shows that a sufficiently aggressive release of high quality expansion material can keep a game’s content fresh and new, but the relatively underwhelming numbers of its latest expansions also demonstrates that there are limits to how long a game can be kept alive. Shiny expansions or no, World of Warcraft is an incredibly old game which has reached the point of comparing very poorly to new games in terms of graphical richness.
The fundamental limitation here seems to be that expansion packs can add new content, features, and refinements to an existing game, but they can’t effectively upgrade the core of the game engine. In addition, even World of Warcraft with its expansion-centric business model does not seem to have been able to spare the budget to go back and upgrade existing content to the evolved standards of later expansions; just compare any of the original class character models in WOW to the ones for the new races in each new expansion.