Creating a campaign that ends with a TPK by design

Question

I'm creating a campaign that takes players from level 3 to level 15+. The premise is that the party is granted a special role by the reigning good King to protect and serve (essentially they are special ops for the Kingdom). Over time, the players will investigate and protect the nation from threats by going on special missions and such. Hints will be dropped along the way that a major event is going to happen and the party will actively be trying to stop it for the final levels. Ultimately the castle+city they are trying to protect will be attacked and they find that there is no stopping the destruction of the city and castle. They will fight to protect people as they evacuate, but eventually "go down with the ship," so to speak.

The narrative is meant to be crafted much like that of Halo Reach, Rogue One, or similar. Each character will have an important backstory that ties them to the kingdom in a special way. My question is:

Do I tell the players up front their characters are going to die in the end, spoiling the story? Or do I try to craft a narrative where they slowly come to learn their fate?

I fear that if I tell the players their characters will die upfront, it ruins a lot of the suspense and hidden story, but also they may choose to invest less in their characters. On the contrary, if I don't tell them, they might be pissed at the end of the story if somehow I can't bring them down the "self sacrifice lane."

I was reading this related question: How can I plan a TPK finale that doesn't look planned?. I like some of the answers there, except that the question there is for a mini-campaign on the side. My campaign is meant to take well over a year - and players will hopefully be more invested in their characters.

Answer 1

Surprising your players with this will subvert the expectations the game itself is built upon.

D&D is designed as a fundamentally cooperative story-telling game, and not just in a "the designers intended it to be this way but didn't say it had to be this way", I mean that the idea you've presented here runs against the fabric of the game. We see in the introduction to the Dungeon Master's Guide:

You’re the DM, and you are in charge of the game. That said, your goal isn’t to slaughter the adventurers but to create a campaign world that revolves around their actions and decisions, and to keep your players coming back for more! If you’re lucky, the events of your campaign will echo in the memories of your players long after the final game session is concluded.

[...]

The success of a D&D game hinges on your ability to entertain the other players at the game table. Whereas their role is to create characters (the protagonists of the campaign), breathe life into them, and help steer the campaign through their characters’ actions, your role is to keep the players (and yourself) interested and immersed in the world you’ve created, and to let their characters do awesome things.

Knowing what your players enjoy most about the D&D game helps you create and run adventures that they will enjoy and remember. Once you know which of the following activities each player in your group enjoys the most, you can tailor adventures that satisfy your players’ preferences as much as possible, thus keeping them engaged.

By writing the end of the story before you've gotten there, you have undermined the agency of your players and removed them from the basic formula of the game. You are all supposed to be telling the story together, and each player's choice are supposed to matter. Your plan says to your players that their choices don't matter, and that you only care about telling your own story. This will be a very unpleasant surprise, because this is not the game the players thought they were playing. Your players have come to the table expecting their choices to matter. They expect to have some measure of control of their characters' destinies. And you are planning to deny them these things. You are trying to play a completely different game from the one the players think they are playing.

Tim C has some good thoughts about player agency in this answer, I'll reproduce here the definition and Tim's brief analysis of your idea:

What is Agency?

I personally define agency by three criteria:

1. The player has control over their own character's decisions.
2. Those decisions have consequences within the game world.
3. The player has enough information to anticipate what those consequences might be before making them.

[...]

A defeat in which the players had no agency will always feel arbitrary - the players will feel as if the GM cheated them. A defeat that follows from Agency is one that the players can feel responsible for - because they knew the risk (condition 3) and did it anyway (condition 2) by their own free will (condition 1).

This is a bad plan. Just don't do it. Don't remove the cooperative part of cooperative storytelling. Let your player's decide the fate of their characters. Show them that their choices matter and that you care about what they are bringing to the story; not your story, but your story, and her story, and his story, and their story. You should be working together, but this plan says "No, you are working for me."

To make it work, the players must buy in up front.

The question you linked, How can I plan a TPK finale that doesn't look planned?, has a concise solution for making this work. SSD writes:

Tell them that the side campaign ends with an epic, glorious, TPK. Making your players co-conspirators in the shape of the finale means that they will help you drive it to that epic conclusion.

There's no value, in the scenario you describe, to making a TPK a surprise or look unintentional. Save the energy that would be spent on smoke and mirrors designed to trick your friends, and invest it instead in making it awesome with your friends. Make those deaths glorious and inspire song, but do it together.

What SSD suggests here is exactly how the Dungeon Master's Guide describes the way the game is best played. If you want to do this, you have to pitch it to the players beforehand and get everyone on the same page about how the story will ultimately go down.

Answer 2

I have had this done to me before, it is a BAD idea

The basic idea of D&D is that you tell a story with the players, something I see too often on here is a DM who wants to tell a story to their players.

In a normal campaign, although nobody tends to spell it out, it is generally accepted that the story is 'heroes who win and defeat some great evil'. If you think about it, the ending there is still known by all, the real story is uncovering what that evil is, and the agency is in how they stop it. Every generally assumes the good guys (players) will win in the end.

All you need to do here is specifically state to the players that the end condition is different, this isn't a story about players overcoming evil, this is a story about the player trying to achieve some other goal. Maybe that goal is to save as many as they can, maybe it achieving a life ambition etc.

Work with the players to determine what the new end goals of this campaign are, and let them know that in advance.

The story and agency still plays out as normal, the end is known, the manner, methods and characters within the story are where the agency lies.

Why is is a bad idea to hide it

As I said, the normal unspoken premise is that the players win. I recently was involved in a level 20 one shot where my character was effectively forced into sacrificing themselves to save the world. I hadn't known that when I created my character, and realistically my character would not have made that choice. It left such a bad taste in my mouth that it was on the list of reasons why I no longer enjoy that person being a DM, the trust was broken and once broken trust is hard to regain.

Warning for running a game like this

I haven't tried it, but having watched EXU where they did this kind of thing, you could see the players actively trying to stop the end of the world, as their characters would be expected to do. The players knew they would fail, but the DM had to adapt to situations where the original plan might have actually been foiled. It takes a lot of creativity to be able to let the players try and destroy your plans and if you don't have enough layers to those plans, it has a higher chance of feeling forced than in a normal campaign.

The players know it is coming, but seeing it forced on them might still be a little disappointing. I don't know I would have the DM skills to pull this off well, but good luck!

Answer 3

You need to write a novel, not run an RPG

Since you tag this dnd-5e, I'll reference the rules of that game - PHB p.6:

1. The DM describes the environment.
2. The players describe what they want to do.
3. The DM narrates the results of the characters' actions.

I can't see step 2 in your plan.

For example, what happens if given the die heroes option one player says "Bugger this for a game of soldiers, I've got a Teleport spell and I'm out of here. Who's coming with?" (Actual real game experience right there - I went with and we left the fighter to sort out his issues with the lich).

Or, what if the cleric gets lucky with their divine intervention feature. Are you saying this threat is beyond the power of a deity? If so, option 1 above should have happened a long time ago.

The only thing the players control in the world is their characters. You have no right to subvert their control. If that's what you want, then, as I said, write a novel - then you can control everything.

If you like, you can put a very difficult situation in front of the players but players are inventive and if they can find a way out, you have to let them.

Answer 4

You don't need to tell them, but...

You don't need to tell them, and there are definitely advantages to not giving away the ending, but you'll need to soften some of your expectations and do a lot of work building the expectations of your players. As others have emphasized, this can go very wrong. But that is contingent on a few things, namely your players expecting a "happy" ending (or perhaps no ending at all and a never-ending adventure), and you as the DM being dead-set on a complete TPK no matter what.

You need to set expectations for your players

You'll need to set expectations about what kind of story you're telling (grim, finite, etc.) that sets the tone; this can't just be something you say offhand in a session 0, it has to have sincere buy-in from your players and be something that's reinforced. If they forget that expectation, and most of the campaign up until the end feels like a more generic adventure, the ending will still come across as unexpected and likely leave a bad taste in their mouths.

Some advice on achieving the appropriate expectations without giving away the ending:

• If the campaign is going to have a largely tragic ending, that needs to be built up throughout the campaign--bad things need to happen, beloved NPCs need to die, etc. Nothing good can be won without cost, which will foreshadow the kind of ending you're aiming for. It's a difficult balance to walk, but it needs to feel like the story is not going to have a purely good ending.

• In my experience, defining the campaign as a fixed, finite adventure goes a long way. Lay out that this is not a sprawling ongoing campaign of adventure, but a specific story with narrative and character arcs that are going to reach a conclusion.

  • This can be reinforced by structuring the campaign more explicitly as something with a beginning, middle, and end. Structure it like a tv series with episode previews, make references to mid-season cliffhangers and talk about entering the final stretch of "episodes;" this keeps the expectation going that things are going to wrap up and that things are gradually heading towards a (tragic) conclusion.

• References to specific media with similar tones is helpful. You mentioned Halo Reach and Rogue One, which is a good start, but if you don't want to make a TPK explicit, more general stories with tragic conclusions that feature large amounts of character death might be good touchstones (e.g. Gundam series by "Kill 'em All Tomino").

  • Make sure that this isn't just something that comes up in a session 0; this isn't a checkbox for you to mark to justify a surprise TPK. The campaign should feel like its borrowing from those kinds of stories throughout.

• One thing that I've found is a useful clarification is to indicate that the intent of the campaign is to play into genre conventions not subvert them. The meta goal is not to find a way out of whatever tragic situation presents itself, but to play into the tragedy to make it as satisfying and meaningful as possible.

And temper your own expectations

On the flip side, you need to temper your own expectations. If you insist on a TPK no matter what, you run a serious risk of either shoe-horning in your ending via fiat and contrivances that make for a disappointing narrative climax, or you'll be disappointed when things don't go your way. Maybe one character survives, maybe half of them do, but you can still maintain the tone of a grim conclusion and a victory only won at great cost--characters that survive will have lost friends and loved ones, they'll be stricken with guilt over the people they couldn't save, etc.

You want to create an ending that is satisfying

You can have a narratively satisfying ending, built with your players, while still keeping secrets. What you want is to build situations, narratively and mechanically, where your players want to sacrifice their characters. Where everything is on the line, where the characters realize that there are things worth the incredible cost, etc. This means building the narrative around their stories, make sure they have the relationships to people, places, and ideas that would be worth sacrificing for. And then mechanically set up the dichotomies that will enable those scenarios to play out--someone needs to face impossible odds to cover an escape, they need to blow up the demon-engine from inside, etc.

Again, some tips for achieving these ends:

• When it comes down to the end, that's when it's time to speak more explicitly with your players--is your character willing to go out in a blaze of glory next session? What kind of heroic/tragic ending feels right for them? etc.

  • Maybe they don't want the character to die, and you'll need to be ok with that. Ideally you've built things up until that point that this is a considered, mutual decision though. "I want my character to falter at the last moment, trying to save themselves at the expense of others, unable to overcome their own selfishness" is something you can work with and maintain the tone of the ending; "I want my character to save the city and ride off into the sunset to a happy life" is more of a problem and a sign that expectations were not mutually understood.
  • Keeping these conversations one-on-one can keep the suspense going for the other players.

• Use carrots not sticks. Fudge the narrative and rules in favor of making the character deaths more meaningful, not more likely.

• Make sure you have dynamic end scenarios with many hooks for compelling character death. A static fight against a big bad has more limited opportunities than a chaotic escape (which it sounds like you're going for), but also keep multiple objectives rolling so characters can split off for their own heroic ends. Enemies need to be held at bay while obstacles need to be cleared and someone needs to stage a distraction and someone needs to run back into the burning building to save that beloved NPC etc.

• Pay some thought to the mechanics of character death. In something like D&D 5e, this can actually be a little tricky, as combat mechanics lean towards rigidity and death is a pretty binary thing (contrast with something like Genesys where GM fiat is more common and it's easier mechanically for characters to be on the brink of death while still having agency to do things). This is the climactic ending of your campaign, so you should consider bending the rules, or setting them aside for purely narrative moments--when a character says they'll go back and hold off the horde of demons, you don't need to make a bunch of attack rolls, it's a wall of demons, the outcome is clear, just narrate it for maximum drama (this is to taste for different groups though).

  • Take advantage of situations where players say there's something their character wants to do to set up possibilities of heroic sacrifices. "Yes, you can stop the demon-engine, but you'd need to physically throw yourself into its gears to jam them up." This of course requires care--you want to be offering players options to achieve something cool and heroic that they want, not shoehorning them into a character death in a scenario where it's not really warranted.
  • Don't be completely afraid to put things in the hands of the dice and mechanics. "Can I hold off the enemies long enough for my friends to get across the bridge?" "Maybe, it looks like it'll take two rounds of combat for everyone to get across, do you want to try it?" and then play out the combat.
  • Much of the specifics depend on the group and system in question, but the basic idea is to use the system rules when they play to the strengths of the situation, move away from them if your group is cool with that and it enhances the excitement.

Which is all to say that you can build to a satisfying but not entirely expected TPK (or near TPK). And I think it does have serious advantages; I've run multiple campaigns with similar dynamics, and the climactic sessions have been thrilling and character deaths incredibly meaningful. Laying out a TPK ending upfront has its advantages, but it does also foreclose the possibility of that slow-rolling tragedy, the dawning realization of how grim the ending is going to get, the price that's going to need to be paid to salvage any glimmer of hope from the depths of despair.

Answer 5

I agree with all of the other answers that suggest this is a bad idea. It probably is a bad idea, and will likely end poorly. However, if you're set on it, there a few things you can do that might help keep it from being a trainwreck*.

Prophecise their doom early

In the first session have them see a vision or have a prophecy explained to them, detailing basically your plans for the final session. This will be some minor spoilers obviously, but they'll be able able to plan to try and thwart the prophecy.

The Future Isn't Set In Stone

You'll be tempted to force their hands, and make everything turn out exactly as you planned, but PCs never do what the DM expects, and if the party does come up with a legit way to avoid the prophesied doom let them have it. D&D is a game of heroes after all, not a novel. Part of the joy is playing to find out what's going to happen, not to watch the pre-planned plot unfold.

*Note: You're very much railroading here, and your players will want to go off track for sure, which generally isn't great for trains.

Answer 6

Definitely Tell Them

I feel like most of the replies are negative, or saying how you must base a game around players, but I disagree. The campaign idea sounds fun (I'd play it), but as a player I'd also want to be in on it. Planning my own character's glorious last stand would just be part of the game, then. Although I might suggest coming up with a scenario where, indeed, there is no chance for escape or second thoughts. Maybe, in the end, the players can only stop the calamity by making a trip into another plane where there is no way back and long term survival is impossible. Or maybe the people they save become the basis for the characters in their next adventure (which they can only do by saving them, which they can only do by dying).

Players should have agency, but the world can have a story too.

Perhaps, alternatively, you could not quite tell them but say something more like "This campaign is hard-mode, expect to die in the finale." (If there's not a Game of Thrones RPG that kills nearly all the players, I just had a great idea for a new product.) I feel like the warning makes it fair. Kinda like choosing the difficulty on a video game. You died, your game is over, but, ya knew what you were getting into when you selected "Hardest" on the title screen. Same thing here.

Basically I agree with your "Rogue One" kind of premise. It's a fair setting. You're definitely going to die in the end and no amount of player agency is going to prevent that but the question is if you can get those death star plans or not.

Answer 7

Figure out, narratively, what is reminding the players of their mortality.

You have actually asked a very specific question about narrating a certain thing, and you are getting a lot of answers about how the certain thing you're trying to narrate could be very shoehorn-y. I think I agree with those answers in that aspect... but it doesn't, for me, answer the question!

So the question as I understand it is essentially,

How do you narrate that someone should expect that they are about to die, constantly remind them about that, and pursue themes of “what death is meaningful? what is a good death?” within the context of a roleplaying game?

You actually are about 90% of the way there, in that you have discovered the first way—metagaming, telling the players that this is the goal—and you have rejected it as seeming too fussy for you, potentially killing their engagement with the story, etc.

So, what’s left?

Well, you’ve got a narrative reason for their mortality... “off over there,” down the line. You need to bring it “back over here,” so that you can, inside the current narrative, remind them of it. Choose the reminder first. So for example here's five random story-hook-reminders off the top of my head:

• Random physical sensations -- a hacking cough, debilitating physical conditions, just a feeling of ants crawling all over you?
• Nightmarish visions?
• All the magical folks (warlocks, hedge witches, faeries) that they see remind them of it?
• A monster periodically bursts forth and must be escaped, and the players are told by Those Who Know that it will eventually get them?
• All of certain official folks (soldiers, guards) that they see remind them of it?

Just some story prompts to consider, you can make up your own of course.

Then work backwards from the reminder.

If you can figure out what makes the most sense for your campaign, then that suggests a class of narrative explanations why those circumstances would be actively reminding the players. For these particular five examples:

• Physical sensations: Maybe the players are not actually going to be killed, necessarily, by the TPK. Instead they have some sort of wasting disease that they determine to be incurable.

  • Story hook is that maybe the first encounter, we jump into the very last battle, “your team has been infiltrating this dungeon, cultists everywhere, you come down into the final altar and the evil priest has uttered the final word of incantation, a bona fide Lovecraftian demon is coming through the portal!!”. Yes, the players manage to force It back out of the portal before it enters reality... but while It is being banished, It slaps every single person with a tentacle. And then after a year of searching for a cure for the wasting disease they finally come to an actual Angelic Figure, who can heal everything... except they are turned away. “Oh. Oh you're demon-touched. I could have fixed this if you had come to me in the first week after you had been infected but it's been in there for about a year and it's too deep now for me to handle. You just have only a year or two to live. I'm sorry.”
  • But in less-magical campaigns it could be a literal disease, maybe they all met at a support group for a medical condition.

• Nightmarish visions: you can actually ask the players to each describe what their characters see in the cataclysm and then try to puzzle out a way to make those nightmares unfold in the finale. The players are set into the role of prophets who nobody else believes, and they are provided with a sense of impending doom and fatalism from their own prophecy.

• Magical Folks: maybe the players are cursed and are attracting trouble in some sort of cosmic fashion, such that the universe is trying to spit them out. All the magical creatures insist that it's only a matter of time before it does.

  • Maybe eventually when they take this to the Wisest One, they clarify that this isn’t quite what’s going on... instead sometimes a Curse, which spins like a vortex, to be whipped up you have to create a Countercurse, a vortex spinning the opposite way, which will undo it. And they explain that the players are the Countercurse to undo the Curse that is going to destroy the kingdom, and that is what's going to kill either them or the kingdom.

• Monster: This creates an interesting setup for the Cataclysm where maybe the only way to save the kingdom is to lead this monster into the heart of the oncoming enemy so that it unleashes its destructive potential on that enemy—but then you yourself won’t be able to escape it.

• Soldiers Comment: Maybe this is a more mundane sort of being marked. Maybe all of the players have a tattoo on their face that signifies that they were once traitors to The Crown but have now, instead of being Executed, been Redeemed, at the cost that they took a holy vow to die protecting The Crown. Every official soldier calls them “Morituri” when being nice, “Deaders” when being dismissive. They have low status but people accommodate them as they’re on official missions, and those missions are usually life-or-death circumstances that you wouldn’t send “ordinary” soldiers to handle.

So once you have the narrative hook with which you are poking the players, then you can find the leeway to essentially railroad them—more accurately, you are letting them choose whether they wish to be railroaded and accept their mortality and search for a meaningful way to die, or whether they wish to oppose it and reject their mortality and screw everyone else over in the pursuit of their own self-interest, and what sort of character will that make you. (That story can also be interesting! Imagine that those characters become NPC villains for the next campaign!)

Answer 8

Do I tell the players up front their characters are going to die in the end, spoiling the story? Or do I try to craft a narrative where [the characters] slowly come to learn their fate?

Yes, both

Any perceived added value in having the players not know the final outcome does not outweigh the need to clearly communicate with the players the structure and assumptions of the campaign during Session 0 so that they can be properly prepared to play.

I played in a campaign with some similar features. At the outset the DM requested player buy in to the overall theme of the campaign: the characters would need to be interested in collecting the magic artifacts required to banish the BBEG that would arise over the course of the campaign. In this final conflict, character death was to be expected, and even a TPK was possible. This was all sketched out prior to starting play (including character creation); and it worked well because everyone at the table was on the same page.

Incorporating narrative elements so that the characters learn that a) they have a fate, and b) this fate is going to radically change them, is just good story telling.

One thing that makes me suspicious about the question is that it implies that you want/need a TPK to be inevitable. Inevitable player character outcomes is not a good feature in setting up a campaign. Note that in the campaign above, character death was not guaranteed, just very likely.

Answer 9

I would suggest neither telling the players, nor changing your grand scheme, but give them a choice.

I have successfully run several kobayashi maru scenarios, or very close to, both at home games and at conventions, and with the exception of one player, everyone has enjoyed it, even those that fail. I usually present a few clues early on, that if disregarded result in total failure, while also preventing a solution from being deus ex machina. In some, however, I haven't provided any solutions, rather just a choice. Continue to certain death, or face some grave consequence.

At a convention, the characters are pre-generated, so the players have less investment in the player characters, but I've also had success with these types of scenarios in home games.

Try before you buy

However, what I would suggest is instead of playing from level 1-15, then presenting them with an unwinnable scenario, start small. In the very first adventure, build a five room dungeon where the stakes are lower, but still grave, like maybe the lives of a group of villagers. Pit the party against an impossible encounter. If they engage, the villagers can escape, if not they get eaten or something along those lines. If they disregard their own safety and the villagers escape, and you've made it difficult enough to kill at least one, then you can gauge the response.

If they found the sacrifice heroic and enjoyable; a tale to tell, then keep building on this, up to the final encounter.

If they don't have fun and enjoy it, the perhaps consider an alternate ending.

Player death

If they do enjoy it, and you establish player death as a moderate probability, firmly put the choice in the hands of the players so they retain agency and can choose to make the sacrifice.

Bookkeeping

If you do establish player death as a probability in a home game, you will need to alter the experience rules a bit so players can roll up new characters and start at maybe half the party's level but catch up within a few sessions.

Answer 10

You don't have to give away the ending as long as you set the players' expectations correctly. Make sure they know that this is a heroic, but gritty and realistic campaign where they will routinely run into challenges too big to just attack head-on unless they're really looking to make a heroic last stand and die gloriously.

Then, make sure they do regularly come across things that could easily kill them all if they don't have both luck and good strategy on their side, and occasionally come across things that will wipe them out if they do anything but back away quietly. Maybe they can come back when they're stronger, maybe not. Encourage wisdom checks for assessing threat levels of encounters. Consider making it an official skill that somebody can have proficiency in. (Assuming the players aren't seasoned enough to just know off the top of their head how hairy the monsters ahead are and which specific questions to ask about the circumstances.)

Since this is a war, draw out your maps and plot out the campaign strategies of the opposing sides. Give the players a chance to learn what those strategies will be and opportunities to disrupt the enemy, or boost performance of allies. Depending on what they come up with, maybe this band of valiant heroes really can change the course of the war. Probably not... But stranger things have happened.

Don't railroad them into taking part in the last stand. Don't prevent them from starting the evacuation early and escaping once they realise what's going to happen. They need to feel like their decisions matter, at least to their own characters. Even if they're somehow still just low-level grunts with no respect by level 15, they could at least desert if they wanted to.

If they've weakened the enemy enough, and have devious enough defense plans, maybe they can even win. A band of four or five in the middle of a clash between powerful nations probably won't turn the tide by themselves, but it does happen. Make them work for it, but if they deserve it, give it to them.

I'd suggest making sure they know to make full use of minions and sidekicks and hired help. Make them do the bookwork obviously, you're busy enough, but a realistic adventuring party for this kind of thing is probably between twenty and fifty, so if they're not part of the official army, that's about the size group of allies they should be looking to collect up. A lot of them can just be low-level nearly-nameless grunts who schlepp packs and hold the horses and keep watch and dig ditches. They can slowly morph into specialists over time to fill in gaps in the party's abilities and reinforce critical ones.

This can make sacrificing a character here and there over the course of the campaign easier as well since they'll both have a variety to choose from, and replacement candidates to be the player's main avatar already on the scene. The "gritty" campaign I'm currently running everybody has two mains and three or four trainees working their way up. The context is it being a player-run mercenary company, so they have to balance cost of wages against what skills and muscle they'll need for the job they're taking. In your setting they're doing it for king and country rather than cold, hard cash (probably) but a similar arrangement both eases the shock of losing a character and will let them accumulate enough power that it feels like their actions actually matter. If they're charismatic, bold, and successful maybe they really will matter enough. If they're not then, when they get wiped out at the end, you want them saying, "oh, yeah, I should have seen that coming and prepared more thoroughly," not, "Stupid, rigged campaign that's impossible to win."

Which isn't to say that you don't give them plenty of opportunities to indulge the laziness which will lead to your desired outcome. Just that, ideally, they'll blame themselves when it's all over.

And, if you're new to running big story arcs like this, I'd suggest spending quite a bit of time in the early phases where it handles like a typical dungeon crawl, and then slowly turning it more toward letting the players pick their strategy and involvement in the war as you get to know them and they get comfortable with the campaign setting. What they will choose to do is probably more entertaining than the canned ending you had in mind anyway. If they veer off in some other direction, that's fine. They can hear the bards singing the praises of the noble heroes who sacrificed themselves to evacuate the capitol after the fact and experience a glimpse into an alternate timeline where it was them if you want.