I lean towards keeping metaplots to the DM. Publishers don't know what the DM and Players are looking to do or want to do.
As a DM I always use metaplots for a campaign. These huge storylines often overlap with other campaigns I do to make it more realistic. Something from a publisher wouldn't work.
That said, a book about how to do huge campaign metaplots with examples and troubleshooting would be something a publisher could do that may interest me.
An example that almost worked was the External War in the Star*Drive Alternity setting. That was awesomely done, as long as you didn't mind losing
Hammer's Star and some other star systems. If you went to a lot of trouble to create NPCs, plot hooks and what not for a system and then the plot destroys it, that's going to create a lot of work for you.
Some examples, such as Dark Sun, Forgotten Realms and Dragonlance are far worse. I don't think the first Dark Sun novel wrecked the setting; IIRC, Tyr was always meant to be free from the start of the setting, and you didn't have to start there. However, by the fifth book previous well-written source books such as the Veiled Alliance were rendered null and void. (VA went into detail about non-Veiled Alliance life in several cities, and much of that was no longer valid after the deaths of so many sorcerer kings.)
Maybe if metaplots expanded the setting, things would work out better. For instance, in the External War, the GM learned much more about the aliens, and since much of that was background info, you didn't have to use it (the changes would be invisible to the players, except what you let them know). It even offered information about new star systems. Too bad there were destructive elements as well; it's pretty difficult to run a war without destroying something.
If a campaign were started after the External War was released, so the GM knew which areas to focus on, things would work much better, IMO. Unfortunately, that only works because Star*Drive has been canceled, whereas most of the examples are still running.
Looking back as someone who didn't participate, I think the coming of the clans in the Battletech universe would count as one that worked in the long haul. Yes, it's only partly an RPG connection. However, the wargame had a lot of roleplaying elements in the campaign sense.
I did stir up a lot of hate and drove some people from the game. I admit any metaplot is likely to do that. However, I think many campaigns thrived because of the change and it certainly is popular in retrospect.
Biggest problem I have with metaplots is that they railroad the campaign setting at least with regards to the published material.
So, BBEG and his horde take on the civilized countries with support from rebels at the imperial court? Good idea! My players will have a field day dealing with them, trying to save the Emperor, prevent a war between allies, and so on.
Next book comes out, and it is placed in a world where the Emperor died, two allies wrecked each other, and the horde occupies half the lands. Sorry, folks, all your deeds are undone!
Alternatively, the DM can adapt the new book to what his setting looks like, but that's more work, and over time it tends to become nigh-impossible since all the metaplot shaped the setting into something unrecognisable.
So, in my opinion, metaplots should either be alternate unvierses, or so small in scope that their results do not change the setting too much. Replacing a king with his brother is one thing, replacing a monarchy with a republic, and dropping to neighboring countries into the sea is another.
Metaplot is one of those amazingly polarizing topics where it seems that people either love it or they hate it with the hate of 10,000 hells.
Me? I adore well written metaplot. Things like the metaplot of oWoD Vampire, or the past and ongoing metaplot from Shadowrun (and its linked Earthdawn stuff), and even the much more subtle metaplot from Planescape (though it did end 1/3 through its planned metaplot and was never picked up when TSR/WotC stopped having a distinct PS product line).
I find a well-written metaplot to be the hallmark of a rich, interesting, and evolving setting. Without a metaplot or some (gradual) timeline evolution of the setting, a setting strongly risks becoming stale, sterile, and of little interest to me. I adore a setting with a metaplot to continually inspire and interest me in seeing how that world develops (even if I don't use all elements of it in my own campaigns).
But when metaplot becomes an openly destructive hammer (see 4e FR) it doesn't work as well, or at all, and risks fracturing the fanbase of a setting pretty harshly.
So, BBEG and his horde take on the civilized countries with support from rebels at the imperial court? Good idea! My players will have a field day dealing with them, trying to save the Emperor, prevent a war between allies, and so on.
Next book comes out, and it is placed in a world where the Emperor died, two allies wrecked each other, and the horde occupies half the lands. Sorry, folks, all your deeds are undone!
There was an online discussion of advancing the timeline in Eberron. I suggested a series of books that created an event that gave multiple approaches to the event (perhaps a chapter for each). Each future book could include sidebars that would give variations on that books content when you used one of those options.
For example, you could have a "The Next War" book that listed various ways the war could start up in Ebberon. Later adventures that might be affected by those plots would have sidebars with varitions.
Obviously, this approach works best when these books are limited. Either a limited number of the books (so that the books you have to consider are few) or somewhat narrow options that wouldn't affect most future books (maybe a book on the possible causes of the Mournlands).
Greyhawk: From the Ashes
Forgotten Realms: 4e?
Mystara: Wrath of the Immortals/sinking of Alphatia.
Cyberpunk: Cybergeneration, anything after 2020
Dark Sun: 'ruined' right after publication, by release of the 1st novel!
Traveller: The New Era, sunk GDW.
Vampire: ? - I'm not familiar with Vampire: TM, but I hear it's the grand-daddy of metaplot.
Planescape: Faction War pretty much trashed the setting, esp with the removal of factions. The idea was to un-Planescape the Planes and get them back to powerful monster-hunting grounds rather than cosmopolitan planehopping world it had become.
Ravenloft: The Grand Conjunction, which reshaped the domains of dread, removed domains and domain lords, added seas, and pretty much shook everything up.
Forgotten Realms: Two of them. The Avatar Trilogy/Times of Troubles from 1e to 2e, and The Spellplague from 3e to 4e (2e to 3e was handled through a modest advance in timeline, but no major crisis).
Most of White-Wolf's lines (Vampire, Werewolf) dealt with a coming apocalypse at the millennium. After the actual millennium, most of the setting went adrift of plot and White Wolf chose to revamp the system along with reboot the metaplot.
Dragonlance: SAGA. The whole book-line moving from and away from the SAGA iteration of Dragonlance was chock-full of meta-plot.
It should be given special award to Die Vecna, Die: a module that manages to trash no less than THREE different settings (Greyhawk, Ravenloft, Planescape) in the course of one module.
I can grow to like a metaplot (though I will always be initially skeptical to it) as long as it does not mess with what I at least consider the core assumptions of each setting. In some settings, things have changed so dramatically so often (Dragonlance, I am looking at you) that a long running campaign simply is forced to ignore the metaplot. In other settings changes also happen often, and though they may be less dramatic they can easily sidetrack your campaign if they happen in your location (most of FR between Year of Shadows and Year of Blue Fire). In some settings a single event can change several basic aspects of a setting (Faction War and the Prism Pentad changed basic aspects of their settings). Some metaplots can be dramatic and still work, at least for me (the Grand Conjuction in RL for instance). And there are some games where the metaplot is extremely engaging, possibly more so than the game itself (Vampire).
The question is how necessary is a metaplot. Would the FR and DL novels be diminished if they were no longer canon? I for one do not think so. What would ofc happen is that the novels would not be sold as easily, since people would not buy crappy writing just to see where the setting is heading (not that all the novels are crappy but many of them are, especially some of the amazingly tacky and trope-ladden FR trilogies). I don't think that a metaplot is necessary to sell new accessories either. The crunch changes between editions can give you a lot of writing space, and there most often is a lot to write anyway. Not to mention that even in settings thick with metaplot like FR you can find accessories that have entire paragraphs copy-pasted from the past (and not just once).
And then there is the Eberron approach. Instead of a metaplot, tons of plot hooks, with many alternate plot hooks per story. As a previous poster said, you can easily have an entire accessory planning out alternate ways one story could be resolved. The fact that a timeline is static does not mean the setting cannot be vibrant.
So, in my opinion, metaplots should either be alternate unvierses, or so small in scope that their results do not change the setting too much. Replacing a king with his brother is one thing, replacing a monarchy with a republic, and dropping to neighboring countries into the sea is another.
That's how they do it in sitcoms. Lazy dad's actor quits and Lazy Grandpa steps up to the plate. Bright beautiful agent's actress quits and in comes bright beautiful foreign agent and fills her shoes. Happens all the time.
