On September 10, continuing a trend away from the subscription model, the Tolkien MMOG LOTRO opened the doors of Middle Earth to anyone. I hadn't dropped in to check until tonight; I couldn't sleep so I thought I might kill some orcs for awhile. I was surprised to find, at this hour (3am US EST), a queue on my server. The game normally had 8 servers, now it has 15. There were more people than usual questing at mid-level, and quite a few young hobbits running around the Shire. I don't know about the end-game - as with most MMOGs, its more of a sport than a fantasy RPG at that level, so I don't have characters up there.
In other words, it seems that Turbine's guess that F2P would greatly increase interest was accurate.
The economics of this involve network effects and entry costs. If your product provides exponentially more value when more people use it, you really want to take a loss on the first users. Those early users are the source of value of potential new users, and having them in your game means that more outsiders will join.
There's also an aspect of low entry cost here - why indeed should there be any barrier at all between the new user and the experience? The prior test drive mechanism required people to get the software at the store, sign up to pay, and then quit within a month if they were unhappy. The new method lets people simply download the client for free and click 'Play.' So long as you can eventually get revenue from these more casual folks - through simple in-game points systems and such - this can work.
The question is, will you get enough revenue? By opening the door, will you get enough new people that the nickle-and-dime revenue strategy makes up for the loss of good, solid, almost-guaranteed subscriptions?
The fact is, even though F2P started among casual games, it is the main model of new launches these days. It will be interesting to see whether LOTRO's move further extends this trend to the intense-game space, or conversely provides a counter-example.
So far, it looks like a winner.