Not sure why I have been thinking of how I'd run a MUD so much lately, though my repeated failures at being content with MUDs must have something to do with it. In addition to my thoughts about how combat might work and some thoughts on a setting, I've been thinking about the question of player-driven plots.
Those of us used to pencil-and-paper roleplaying games take for granted the idea that player actions drive the story almost completely. In MUDs, the tiniest crumb of player-driven plot, which even the most railroaded face-to-face RPG would have hundreds of, is taken as a wondrous new innovation and accomplishment. It's partially because of the limitations of coding. A MUD by its nature is more asynchronous -- admins and players log on and off at varying times, which is one of its strengths, you can always go play when you want -- which means things have to be built and coded ahead of time. There's a lot less room for improvisation when the admin can't just say "sure, that might work" without having to stop everything and go recode something or build a bunch of new objects whose need wasn't anticipated. The admins can't spend the time to build things that might not be needed that much because of limited time and resources. And it's even more non-viable to have an event able to adapt to unexpected player actions if it has to keep happening when the admins aren't around!
Ultimately it's a question of resources. Imagine a MUD like Lusternia which has a very active set of admins, builders, and coders, who churn out new stuff at an unprecedented rate. Yet the players complain because events are handed down from on high, rather than bubbling up from player actions, and are essentially linear, with only one possible outcome. As well they should; you don't get a feeling that you can impact the world you're in if it's impossible to do so.
Now imagine if they came down one day and said, "From now on, you will get only about 1/4 as many events, and everything else will slow down too -- new areas, new skillsets, bug fixes, etc. But in exchange for that, we will have player-driven events every month." Would people be happy or not? Hard to say -- people are always willing to say "we need more of X" but have a remarkable capability to ignore that the only way to get it is to have less of Y and Z.
So imagine how this might work in a new MUD where this went in from day one. I imagine a monthly sweepstakes; anyone in the game could buy a single ticket per month for a very low cost (say 5 credits). Once a month, 10 people would "win" the lottery. The prize would be 50 credits plus a chance to submit an idea for an event, in the form of a writeup of no more than one page. Admins would review these ideas and choose the best one, considering implementation difficulties, the event's "neutrality" (events that take the form of "my city has an idea for how to beat up the other cities" would be less likely to win), how well it'd fit into the overall story, etc. The selected proposal would then be developed, possibly with the player who proposed it taking an active hand in helping make it happen, both by roleplaying the stuff related to it in-game, and maybe even by doing some of the building if that's possible (though I think that part wouldn't likely work -- getting someone else to write room descs is often more work than doing it yourself, for instance; not everyone has the talent). The winner would also get another credit prize, say 200 credits, plus get to see the event play out -- probably going in directions even he didn't anticipate!
And these events might well lead to changes in the world around them. In fact, after the initial setup, the events might be the primary method of introducing change into the world; admins wouldn't even make their own events except in the rare case where they needed to do so for some kind of coding or balance reasons.
Naturally I see all kinds of problems that could crop up with this, so consider it brainstorming. I think a lot of those problems might be addressable, at least assuming that the MUD was otherwise able to succeed so you'd have enough players and admins to do it. But I like the idea in broad strokes. Wonder if it could be made to work.