Agora is currently in playtest and you can download the full playtest document if you like.

Short Version: Each player controls a character who leads a community of individuals that shares the primary character's ideals.  Play goes around the table, with each player setting up a scene for their character/community.  Another player runs the opposition for that scene, either their own character/community or an NPCish obstacle (which can be an NPC faction, a monster, or even the geography); the opposition is rewarded for goading the player into rolling/activating as many ideals as possible. All the other players run lieutenants within that player's community.  If the player beats the obstacle, they win their stakes ("forge an alliance with the Ardents") but also harvest some dice off of the obstacle for their own use later.  The obstacle is then put into the obstacle library.  So when it's your turn to run the opposition, you can elect to create a new obstacle (with starting dice) or use an obstacle from before (augmenting the dice already on that sheet).  As play goes on, characters will be shaped by the obstacles that they face and the obstacles will be shaped (and beaten and tamed) by the player characters.

My Current Dilemma: Presently this creates a group of parallel accounts, with each player doing their own story, accomplishing their own goals, and sort of meandering their way through the world.  I want to sharpen this in two ways: (a) providing some systemic encouragement to have the character/communities interact, and (b) somehow creating a more coherent larger structure of play, instead of a long sequence of semirelated scenes.

Acts (the "Hamfisted Solution"): The most straightforward solution might be to divide play into Acts as well as Scenes, so that the first two rounds of Scenes are part of Act One, and the next two rounds of Scenes are Act Two.  Each Act could have a basic situation or open question that is addressed through the act's scenes.  So Act One could be "Surviving on the Harsh Surface" and Act Two could be "Discovering the Hidden Threat" and Act Three could be "Strange Bedfellows" and so on.  The Acts could be hardwired in the game text (ick) or created by the players by consensus (loads of OOC discussion, ick).  This is a sort of top-down solution that I'd actually like to avoid; I would much rather see the larger structure created via emergent effects rather than being strapped into an inflexible outline.

Making Obstacles Recur:  One way to accomplish both goals is to make obstacles recur; different characters can be connected by being pitted against the same obstacle, and a recurring obstacle/badguy can create a sort of multi-scene structure.  As it stands, a player running opposition is faced with three options: (a) run their own character, which has some risk, (b) create a new obstacle at starting obstacle dice, or (c) recycle an old obstacle, adding new dice to its traits.  Since a defeated obstacle loses one trait and you only add dice to existing traits, they cannot be recycled forever -- they'll run out of traits.  However, especially for the first couple uses, a recycled obstacle will reliably have more dice than a new obstacle.  So obstacles should enjoy a certain bell curve of potency and then taper off.  The opposition player wants those extra dice on the obstacle because it gives them a better chance of tempting more rolling-in of ideals.

Now, I'm not sure if that's enough there in order to get recurrent obstacles.  A further step would be to add a soft restriction to the total number of obstacles, so that for every three obstacles in the library, you have to pay 1d10 to create a new one to run.  Players running obstacles get 1d10-4d10 for their trouble, so this could be a sort of gamble -- do you think your idea for a new obstacle will be good enough to provoke more ideals to roll in, or can you use something that's already established?  However, there's no real mechanic for removing obstacles from the library.  As it stands they'll hit a certain decrepit stage where they aren't very useful, so they won't get used enough to get their last traits stripped off of them.  So if I was to implement this soft restriction, I'd also need to implement a trash-collection mechanic that removed unused obstacles from the library.  Which is a lot of added mechanics for encouraging a behavior that may already be provided for with the die power difference between new and recycled obstacles.

The Futility of Stakes: Much to my chagrin, stakes seem to be pretty irrelevant to the actual play of the game.  Each scene has stakes and counterstakes, but watching your Fallout, Burnout, and Spoils seems to get more attention (although that may be me obsessing on the numbers and players getting used to the dice).  Further, players don't really get much by winning their stakes -- sure, I recruited the indigenous yeti-folk to my side, but that's pretty irrelevant and does not have a heavy impact on later scenes; really, it just becomes color ("Oh, George who's carrying the platter is a yeti -- but anyway, I tell the commissioner...").  I can pair the stakes that I win with the spoils that I take -- so if I win over the yeti-folk, I can put Yeti-folk Mountaineers 3d8 on my sheet -- but that's not a hard cause-and-effect relationship and I don't really feel like I earned those mountaineers; I feel like I earned those dice and named them in honor of the scene.

The thing of it is, stakes should allow players to address their situation and afford them a sense of accomplishment (or failure) by having a tangible effect.  It should allow the players to guide and shape that larger structure that I'm after.  However, since winning the stakes "I chop down the entire fucking Black Woods" has no real effect on the Black Woods' dice, the stakes feel empty.  Stakes have no effect on agency, and I'm not sure if this is a flaw or not.

Cliffhangers: Right now each scene encapsulates an entire conflict, so the scene is you petitioning to be admitted into the Council of Small States (or whatever), and when the scene is over, you're either in or you're not.  Which is nice and neat and simple -- so nice and neat and simple that the scenes become hermatically sealed nuggets that do not relate to each other in a larger narrative structure.  I'm considering a "cliffhanger" or "entailment" mechanic where, at the end of the scene, whoever loses their stakes gets to add some detail that gets revealed in the fiction.  You're accepted into the Council, but that night, the president of the Council comes to you in secret asking for your cooperation against conspirators on the council.  Dun dun dunnnn!  However, I'd want that cliffhanger to have some sort of mechanical weight to it, or to point to some other obstacle (perhaps not yet created), or something similar, and I'm not sure how to implement that.  Giving the opposition a die-bonus when incorporating cliffhangers from other obstacles is close, but kludgy (and requires the opposition to keep track of lots of bits of information on lots of bits of paper that aren't in front of him).  Or perhaps this doesn't need mechanics to be engaging.  Thoughts?

Increasing the Allure of Multiple-Player Scenes:  Presently whoever is running the opposition has a vested interest in getting another player to throw their character into the scene instead of playing a lieutenant.  The more character/communities they're playing against, the more ideals will be rolled in.  So if it's Adam's scene, I'm going to try and pick an obstacle and frame the scene in such a way that Bob will want to jump in, too.  On the other side, whoever is playing the opposition may elect to bring in their character as the opposition, in which case they're switching their reward structure from getting 1d10-4d10 for goading the player to harvesting a complication off of the player (which may be a lot more than one to four dice, or be counted in a completely different die size).

Now, presently there are no "helping other characters" rules -- if two players are in a scene, they both batter away at a common obstacle but they can't support each other any more than that.  Adding assist rules ("Here, use this extra two I rolled" and then handing over the die for the duration of the conflict) might give players a good reason to ask for help from the other players ("Help me out here and I'll give you half the spoils!").  Said help can turn into alliances, relationships, and plot developments.  I'm not sure how much that might unbalance things, however -- twice the available dice (roughly) to create two blocs of dice would, I suspect, be a lot more reliable and effective, and trounce Obstacles without any effort.

As you can tell, I'm very much up in the air on how to resolve my problems; I'm in fact not even positive that the problems really exist -- they may just be me overthinking things.  I'd love to hear some different points of view and assessments.  Thanks in advance.

Hi!
  I missed playtesting, sorry, but maybe that will give me the objective view needed for a solution?

Acts
  Well, I think "Acts" might make the activity of narrating easier, since there is an established theme for the Act. But I do not think it really addresses your main concern.

Making Obstacles Recur
  If by making, you mean forcing, I do not think this is wise. I think if an obstacle was fun and interesting, then the players will be more likely to narrate open endings to conflicts in order to allow for it to recur. And other players will re-use it just because it is interesting.

The Futility of Stakes
  Maybe the trick is to have a rule/mechanic/whatever that encourages the players to write the Stakes as it relates to their Goal/Theme. In other words, say player 1 wins Yeti/3d8 and their goal is political, then the should write, Won the political support of Yetis/3d8, while another player who has a military goal wold write Yeti Alpine Soldiers/3d8.
  I think if you adopt a style similar to this, then you are encouraging the stakes to be relevant and useful!

Cliffhangers
  Cliffhangers should be part of fallout. I can't give you any solid suggestion in how to implement them though. But if you can implement them, in a reasonable way, it should add real depth to your game.
  You might want to consider weightless cliffhangers. You know? Just let them throw out the idea and see if it sticks. That way you won't have 80 crazy subplots and none of them getting resolved because someone makes a new one after each conflict.

Increasing the Allure of Multiple-Player Scenes
  I feel like the trick of this is relationship mapping. But instead of mapping, the strength of the relationship or using some crazy numbers scheme. The connection is a resource that one side has in abundance and another needs.
  Say my char is the leader of the insect people, I might have two relationships with the  Cyborg Nation
Insects -> Cyborgs - Propoganda tactics (We help them spread the bull around)
Cyborgs -> Insects - Fertilizer (And we get a cut of all the bull, lol)
  In this way the players already have a built in connection to launch from. But it is not so overpowering or mechanical that it stifles the creativity of narrations...

  Just a few thoughts, I hope they help in some small way!

About five rules iterations ago, I had a "Loyalty" ideal instead of a "Descent" ideal, and it always referenced one or more other player characters.  In the end, though, it was too kludgy to make that stay a relationship what with Fallout rewriting the ideals throughout play.

As far as a hard relation between stakes and spoils, I'm wary of this, if only because it would reduce all stakes into resource-shopping.  My stakes are to get some raw materials.  My stakes are to raise an army.  My stakes are to build a superweaponthingy.  It would leave out tons of good stakes like forming a planetary council, convincing other factions of your noble blood, or even reconciling with your father.

Hi!
  Well, I did suggest that the relationship NOT be about trust or loyalty. But that it is about needs. This need puts the narrative in context, you know?

  And you have a conundrum, you want to have relevance for spoils, but you don't want those spoils to directly impact the next Stakes.
  So, where does that relevence come from? I feel like relating the spoils to your ultimate goal gives it the wanted relevance, and I think we should trust the players not to just win spoils A and Spoils B and then combine them to achieve their goal.
  Maybe I am over-simplifying the issue, but I hope I, at least, can give you a jump-off point to fix this issue.
  Good luck man, sounds likea cool game!

dindenver wrote: you want to have relevance for spoils, but you don't want those spoils to directly impact the next Stakes.

The Futility of Stakes: I'm reminded of the explicit advice in Capes that the way to "win" the game mechanically is to throw events into the fiction that the players are themselves invested in, and that seems like a good thing to rephrase into this game. So while there may be no other mechanical representation of what you won from the stakes, what happened in the fiction still matters: for example, even if I have a net loss of dice for pursuing the stakes of "Orphans are not eaten", I still get to say in my society, that I'm creating, this event does not occur.

That said, I want to re-suggest my idea of letting players put their faction and agenda in the context of a larger, more grand dream or idea, their own "Five Year Plan". Perhaps the idea isn't to "win" as such, but it is to create some beautiful vision of the world. So for example, maybe the player describes a large step towards that dream, and the potential opposing players collaboratively bid down (or up?) the scale of that step to determine the stakes for this conflict. This is fuzzy, but maybe it'll give you some ideas.

Another possibility or tying stakes and obstacles: perhaps your stakes should alter the nature of an Obstacle (so, I clearcut the Black Woods, despite the Black Woods dice still being there). By virtue of my stakes, I'm entitled to rewrite that Complication into something else, so long as it is still threatening, to reflect the changes. So I get to narrate that the woods are entirely gone, and on the Obstacle I change "Black woods 2d8" to "Farmland prone to mudslides 2d8".

Acts: I think that's a sweet idea. (Matt Wilson's Galactic has specific phases where the Big Bad eventually enters the situation.) Here's an idea: put each new Obstacle into a basic category, like Natural, Political, Economic, Social, Tech, Psychic, Wierd, or whatever. Then, there are multiple different ways that the Acts can break down. Maybe it's Survive->Discover Threat->Fight Threat, or maybe it's Survive->Renaissance->Hegemony, or maybe it's Survive->Discover New Apocolypse->New Exodus to Stars. Then, based on the kinds of Obstacles the players have chosen, the Act structure is determined.

Cliffhangers: Doesn't feel quite right. It's too easy for the losing party's thrown in detail to not be really relevant to the protagonist's vision without some negotiation, I think. Are we okay with the player just being "uh, whatever" to the newly revealed Cliffhanger?

Making Obstacles Recur: You could offer a +1d10 reward for reusing an older obstacle, but more generally I'd let the players use their judgement at deciding what's the most engaging obstacle for them at the moment. If a player is already dealing with stakes of high commerce and diplomacy, the "Set up a water treatment facility" Obstacle isn't going to feel right.

I suspect your first paragraph on stakes is more true than I'm allowing for.  It probably becomes even more true when the players get into the swing of things in running Obstacles, where their goal is not so much to win as to goad the player into investing all four Ideals into the scene.  Once players really internalize that, and start creating or recycling Obstacles that will engage Ideals, I think the stakes will become more meaningful, not for what they represent, but for what they cost to win and what that says about the importance of what they represent.

Here's the thing about the Five-Year Plan -- while that's totally something that would be appropriate in the fiction, it's not something that I'd want to hardwire from the start of the game since Ideals can change so thoroughly.  If I start off with a pacifist and my ideals warp around until I'm leading Agora's version of the Fedayeen, the goals that I set in character generation will be moot.

In terms of stakes interacting directly with obstacles, I'm beginning to think that the conflict is only superficial.  If my stakes are "Clear-cut the Black Woods" and the Black Woods is an obstacle with complications "Animate Trees," "Treacherous Terrain," and "Poisonous Mosquitos the Size of Baseballs" then, if I run that scene with the Black Woods as my obstacle, I can harvest the Animate Trees complication, and the Black Woods will be clear cut -- later scenes with the Black Woods as an obstacle can depict it as a wasteland of tree stumps and fetid water.  Even if the scene is played opposite a different obstacle -- say some Green Protectors-of-Nature group -- and I win, the Black Woods can still be clear cut and later scenes describing it as stumpland, just with the Animate Trees as refugees lurking in the shadows behind boulders or whatnot.

The real restriction is that you can't frame stakes to take an obstacle out of play, which isn't a bad thing, to my mind.  Taking an obstacle out of play is not an in-fiction consequence, so why should it be accomplished with in-fiction actions?  You can do all sorts of bad things to an obstacle, you can strip its complications if it's run against you, but in terms of story structure, the obstacle from back in the day coming back at you, perhaps in a different way, is never really invalidated (how many times did Murdoc return on MacGuyver?).  It's even possible to subsume or take over an obstacle (say, the Council of Small States or something) and have it run as an obstacle against you later -- after all, there can always be dissidents, or even just conflicts of interest that make it possible for things you 'control' to give you trouble.

Acts seem a little too hard-wired for me, and I'd really rather see structure be a product of game interactions rather than be something that limits game interactions.  Which is why the Cliffhangers intrigue me.  Note, a cliffhanger on an obstacle does not necessarily have to make the player who just defeated the obstacle want to hit it up again -- it just needs to make somebody at the table interested.  So Player A runs an obstacle for Player B, and after losing makes a cliffhanger that makes Player C's eyes light up, that's all for the good.

I'm conflicted about 'ignored' cliffhangers that are left hanging there.  There always is in serial fiction, so maybe it's not worth worrying about.  Alternately, it might be possible to remove an obstacle from the library if it's down to one or two complications and you resolve its remaining cliffhangers with a new obstacle somehow -- sort of tidying up the hanging plot threads.  That's an idea I need to think on.