They can work IF the players know there will be a twist before hand. They don’t need to know what the twist is, but they should be aware of it.
I once ran a game with a co-DM that had a twist game happen within the first two game sessions. The players new something was coming but didn’t now what it was. I fully believe that had they not known something was coming I would have lost at least one player after the reveal.

I did once throw in a major twist, which was only meant to last for a certain phase of the campaign, and it totally put the players off. They all quit after two session of the twist phase even though I told them it was just temporary. So a big twist can easily end the entire campaign.

I had an established game where the characters accepted cybers so they could use guns for the defense of the state. As the technology advanced and the new kids out of the academy had ever more subtle and more effective cybers, they were like dinosaurs, “early adopters” with all the clunkiness that entails.

The government offered to remove their cybers if they would become shadow assassins, still using guns but with no cybers. They agreed.

They noticed that not only the scars from their cybers were gone, but also all their scars from their whole lives. The plot unfolded, getting less and less comfortable, until they infiltrated the government fortress–and found their bodies in cryo! They were clones!

Then they got the choice of playing their original characters (clunky cybers and all) or the clones of their original characters. It was a brain-bender.

The core of the question is whether the players trust the DM and whether they are more invested in the setting and story, or in their precious snowflake characters. If things affect their characters and they just want to walk away because they don’t get the control they feel they deserve, then a twist game will not work. If they want to see how things turn out and they are happy just to get to play in the game, then twists are likely to work better.

I used to have a game group where I could do twists. My current game group feels too entitled, adversarial, and risk-averse.

I suspect a group of players that could shrug off a TPK or high mortality game could handle a twist. Those who could not, would not.

Very early in the 3e era, within a few months of the players book release, I started a campaign with my then-game-group. I had originally wanted to do a 3e Gamma World game, but 3e GW was not going to come out soon. But the psionics book HAD just come out. So, I just said “you can play any character class, but you wont get any of the magical/spell/spell-like abilities”. We wound up with mostly rogues and a ranger… and one person who pulled me aside and asked to play a psionic character (being a slight mutation).

The FIRST twist was: the world was actually magical … and they were fighting not just mutants, but undead. At first, they weren’t sure if it was a plague type undead, or magical, but it only took a few runs to make it clear it was magical. Also, they COULD get spell/spell-like abilities if they did certain things. One character multi-classed into sorcerer as a result of that (his family had a secret past in the arts, and in a secret society). Another had a spiritual awakening (one that _I_ initiated, not the player, but he went along with it), and became a reluctant paladin (called to a mission, but not at all interested in the glory and preaching). The folklore gods of the setting were complete myth… but it turns out that there had been a deeper truth behind certain secret societies and such, that was an actual pantheon of Gods, and the apocalypse had been part of a war between two factions of the “Builders”. (I played it vaguely enough, at first, that you could sort of go back and forth between Clarke’s Law (they’re just really advanced beings) and something more like 3e Deities and Demigods; but over time, the latter became more and more prevalent).

That turned out to be the _smaller_ twist. The bigger twist was that they were actually inside a dyson sphere. Built by the Gods of that setting to keep out a bigger bad. That also brought up one of the small inconsistencies they had found. The ranger got to spell casting level, and actually had been communing with a nature god. The nature god didn’t the pattern of the other gods. Turns out the god of nature wasn’t “one of the builders”, but was the AI that embodied the sphere’s maintenance systems.

So, yeah, I had a very deep techno-fantasy system going. The campaign broke up because two of the key players/characters (the one playing the ranger, and the one playing the rogue/sorcerer) moved away. The last episode was them being stuck between two of the warring builders as they fought over the “vault” (as in Fallout’s vault-tech vaults) that had been near the village where the players all grew up. They all escaped into the lowest levels, and an usual elevator was found, that had been excavated at the lowest caves under the vault. The last scene was them all getting into the elevator, and then hearing a massive earth-quake inducing explosion above/around them as the elevator was moving further down.

That was _supposed_ to lead to adventures in the sphere’s infrastructure, and starting to deal with the “outer dark”. But that set of adventures never came up.

A year or two later, I tried to start a “thousand years later” game. The mountain where the vault had been had completely imploded, and the mountain-lake had spilled out into a larger, less deep, lake around that area (the mountain lake was actually a pump outlet for the dyson sphere, not a “naturally formed lake”). There was also some swampy areas around it. The memory of the entirely non-religious paladin had spawned … a very complex and structured and dogmatic religion.

We only ran one or two installments of that game, though.

In the original game, the only “disgruntlement” was that the psionic character became less and less interesting compared to the sorcerer’s emerging abilities. So that player sort of retired that character (he came back a few times, but mainly he was retired) in favor of a new character that fit a gap in the party at that point: a hard-core fighter. It was also around that point that the party had found an enclave of dwarves in the mountain range (not related to the vault). So that player switched over to playing one of those dwarves.

But, no one was turned off by the genre-bending nor the major plot twists. We had played a lot of different genres in previous games, and I had been preparing them for the more system level changes (giving them jumping off points to pick up magic, once that aspect of the setting was apparent, things like that)… and I like to think I had been doing a good job of laying the foundations of the story level twists (and they all seemed to agree). So they were twists, some surprising (the day the dyson sphere was revealed REALLY threw a couple of them), but none that were completely out of line with the game group’s sensibilities.

Also, many of them were fans of comic books that had some elements of modern fantasy (John Constantine stories, the Kevin Matchstick “Mage” series, etc.). I think that was a bigger factor in whether or not they were comfortable with the genre twists than “how much warning they had before the game” (which was “none” — they were initially set up to have a pure post-apoc game that was going to be similar to Fallout 2 — as they were starting out as villagers, not vault-dwellers).

So we were fans of multiple genres, gamers of multiple genres, and at least fans of a few hybrid-genre stories.

I think the keys are:

a) a game group that is comfortable with all of the genres you’re covering — my HS game group had some hard core “fantasy only, preferably Tolkien based” players, and I couldn’t have done this in THAT group. It had to be my post-college game group.

b) a plot line that has the players invested in seeing where you’re going with it. Even if they’re not 100% sure about the genre twist, they might stick with you just to see where the rabbit hole goes, as long as they already care about that rabbit hole. Assuming they don’t like twists in general, then if they don’t care about the plot and story and characters, then they wont have as much tolerance for twists. If you get them invested in the story/plot/characters, they’ll have more tolerance for the twists. I had set a few key A plots in motion before the genre twist (adding magic), and a few key B plots in motion before the setting twist (and that caused some of the A plots to become B plots, and visa versa). So the players were all heavily invested by that point.

But, mainly, as with just about any social activity: know your audience. When you’re playing a storytelling type role, play to their strengths, avoid their absolute taboos*, and pull on the strings to their thematic weaknesses. That’s what good storytellers do.

(* unless you’re trying to be avant-garde … which is its own special form of “known your audience”)

I haven’t run any with a complete twist, but I think I was close to that and I have thought about running such campaigns and ran into the same trap.

Currently I run a campaign where one of the players is the new and too-young king (as all his brothers died unexpectedly when performing the inauguration rituals). The premise was a run-down kingdom and a megadungeon below the capital city. That much the players all knew and I think it was vital to know that, though the megadungeon actually is kept secret from the public and the players and the characters still do not know what it’s actually about. I only gave them some vague hints in session 1. But for them to flesh out their characters it is vital to let them know that it’s about intrigue and dungeon crawling and that they start in a run-down kindgom, so there are shades of grey especially within the elite of the kingdom.

The same, I think, holds true for twisting campaigns. You need at least to foreshadow the twist for the players so they know what they’re facing. First of all they need to build their characters towards that, otherwise some players will be driven out of spotlight and the plot. Second, as a campaign is always a cooperation between gm and players, they need to like the campaign to a certain extend and give their opinions, needs and wants to you as the gm, so they can have fun, too. You can’t have any of the if you keep the twist totally hidden from them. Surely this will lead to some players optimizing their characters story- or rules-wise towards something that has not yet happened in the campaign world, but if that is a problem, talk them or run something else. It’s sadly as simple as that. You should also, I think, make the twist the very first thing that happens in the campaign. Maybe not the pinnacle of the twist, but the first session should evolve around things relating to the twist. Otherwise you’ll have some sessions that don’t actually have any meaning for the campaign and are useless. You are also reinforcing the problem of twist-optimized characters and players not liking the campaign.

What might help are options to leave parts of the characters blank or to be able to change them later. This should be espacially difficult for story-related matters, but maybe you can come up with a solution. I’m sure there will be ideas maybe even mechanics about how to flesh out a character during game, which maybe even more cool.

But anyhow I guess you won’t be able to have the “wow effect” for your players within the campaign. Would be cool to have one, to be able to play it all out with all the emotions the surprised players will have, but I think there is only one chance to have it that way: You’re sure your players will like campaign and they can flesh out or modify their characters (partly) while playing the campaign.