I remember a while back reading one or several pieces suggesting that FLOSS games would never take off because the process of creating a game doesn't follow the same dynamics as that of creating some other pieces of software - once you have finished your game you already know all there is to find out and you have little incentive to actually play it to work out all the glitches; and similarly other people have a limited interest in hacking on your game to improve it as it will spoil their gameplay fun.
I don't know how true that is, but assuming it is, I would have thought that FLOSS game development processes could be adjusted to take it into account. There are so many independent components in a game, many of which do repay time invested into improving them. For instance, the world that a game takes place in doesn't have to have been created for the game, but can be one that someone else created for other purposes, possibly adjusted a bit to add elements of surprise (but I am sure there are enough virtual worlds going around that the average player won't know most of them anyway). There are people who enjoy creating animated creatures outside of games, but which could be reused inside of games. Libraries of worlds and characters could be also redone and reused for other games.
There are virtual world engines which can be used for games. And I could imaging competitions between people creating gaming levels to produce those that will be used in a game, with those contributing being those allowed to vote on which will get into the final product, which gives a special incentive for each contributor to polish their piece as much as they can.