social combat in diaspora

In the first edition of Diaspora we kind of made a hash of a great idea, tricked by successful playtests into thinking we’d written excellent rules. And for a small number of people we probably had, but not for everyone and not even for most.

I think the idea was first brought to my attention (nothing is original) in an post in 2008 or so by Fred Hicks. I can’t find it now, but the gist was that maybe you could just use the existing zone combat rules and change the map to something notional rather that geographical and get a social combat system. Zones would be ideas or beliefs or other abstractions, but otherwise you’d leave the system intact.

This sounded like a brilliant idea. In practice it wasn’t — it led to the same problem I always had with social combat: you’re just figuratively beating each other, reducing a different kind of hit point. While the narrative is different, it’s still constrained to make sense of a combative scene, so it’s actually very limited. Like using the combat system to model acrobatics as well as gunfights. Yeah, yeah, Fate fractal, I hear you. Sorry, but endlessly and reductively using the same complex system for everything is boring. And strangely confining. It’s like saying Lego is best used for making a wide variety of giant Lego bricks with which you can make more and bigger bricks. Literally any kind of brick! Eventually you realize you haven’t got around to making anything but building materials.

Anyway, it sounded enough like a good idea that we wrote it up for Diaspora and what we published is pretty much where we stopped thinking about it. The rules give some unfortunately vague advice about what a map should look like and some contradictory rules for how to interact with the map. The thing is, our playtests with it were great. They were great, though, because we stumbled on some specific uses (that actually disobeyed many of the rules we wrote down) that were huge fun. I had actually failed to analyze what I did as a ref so I could mechanize that and instead mostly just wrote down the core idea that led us to a fun space. I had failed to give you the tools to reliably reproduce those good times. Since I hadn’t done the analysis, I didn’t even know why it worked when it worked or why it sometimes failed. Not all of the examples were every tested — they were just ideas that might or might not work.

Since then I’ve thought a lot harder about this. There are actually two different ways I’ve obeyed these rules (sort of) with good results and they are substantially different. These differences should be codified for a decent system to exist, and we might even want to just pick one.

The first is to have the protagonists and antagonists on the map. The objective is to move yourself to the Idea you want to dominate and to move your opponents to some place inert or favourable to you. Since this is the most important part of the resolution, the whole idea of beating each other’s composure hit points down should be dropped — that’s an attrition battle that distracts from the maneuver battle where we’ve invested all our energy. The map is a creative burden on the ref — what best handles the scene’s needs? But this is the biggest problem: not all maps work and there’s limited guidance as to what does work. And a lot of maps that look different are topographically identical.

The other way is to invert this and make ideas the pawns on the table and the map can be people, places, cultures. The geography becomes static and we move the ideas around it. The objective is to cluster ideas where we want them but now who rolls and against who become unclear. The few I’ve run with this inversion have been great but entirely ad libbed. I have no way to tell you how to reproduce this but, as it turns out, I have no way to tell you what maps will work in either case. So we’re not further ahead.

While tinkering with a very early version of Soft Horizon I started thinking about formalizing the map: let’s have one kind of map and all combat is social at this scope. How you resolve the map (where you move all the pawns to) determines just what kind of event this will be (warfare, violence, diplomacy, sorcery, and so on) and this has largely narrative impact: you make the final roll in the determined resolution space and if it’s WARFARE then you narrate your success or failure as a war. That is, the whole minigame is in the preparatory moves for the conflict and one roll resolves the conflict. This never got tested but I think it’s a step in the right direction: develop a single, generalized social combat map. I still think is possible. I still haven’t done it.

abstract plural units mapIt did get me thinking about a single abstract map for combat, though: surely if it was possible to generalize social combat to use a single map, then maybe you could do the same for combat! This is of course the same trap as thinking you can use a hit points combat system and relabel everything to make a decent social combat system. But I still think it might work, especially for mass combat which seems to demand more abstraction.

Could we find a similar common abstraction for social combat? Or is this physical combat really one sub-map of many different kinds of conflict? That seems to be more the case (and I’m excited by the idea of having a vastly richer social system than violence system: one that subsumes violence as a single special case of conflict and not the most interesting). A separate system for romance, persuasion, grifting, …

Can these be enumerated? Can all of them be reduced to a model that is fundamentally about maneuver and action in the context of position? I think if the categories are general enough the answer to the first is yes — but I haven’t found the categories for this to work yet. And I’m not working all that hard on it. And I’m pretty sure that all of them can be reduced to a maneuver model, to be put on a map. Lots of maps.

But the point of this was that I am still, ten years later, flailing around trying to find good maps for this concept of social combat. So the idea that we could make a rule in Diaspora that basically said “first, invent a good map” was absurd. That’s why that section fails: we told you to take on the part that needs all the thought, all the testing, and was most likely to fail. And told you to do it on the fly.

Sorry about that, that was bullshit.



certain death and perception checks

Consider this scenario. I’ve seen it more than once so I think it might be interesting.

As ref, you describe a place that will definitely kill the characters if they enter. You don’t intend the adventurers to go in there — it’s just an illustration of the capriciousness of the owner. The adventure is somewhere else. So, say: the open archway is completely clear of sand, unlike where you’re standing outside of it. And right inside are the skeletons of three people who are dressed just like the people you know were exploring here last night. It looks like they crossed the threshold and died instantly then decomposed over at most a few hours.

From my perspective as ref I have made it totally clear that this path is barred. What I am selling here is the death of the friends not a puzzle. However, players will usually see a puzzle. So:

“I want to go inside and see what happened.”

Hmm, okay. It’s certain death. That doesn’t sound like fun. How about we tie a roll to it?

“Hrm okay. Let’s have a KNOW check then.”

Now this requires elaboration. The player may well think that failure means you don’t KNOW what’s going on. Success means you KNOW what’s going on. But this doesn’t lead anywhere interesting. I dunno means they go ahead and die. I know means I reveal it’s insoluable and the move on. Neither are very fun sounding. And what RISK would I apply here? Fail and die and also trigger a risk? Seems weird. Succeed and you can’t go there but also trigger a risk? Mmmmmaybe, sure. Success with no risk? It’s a hard success to celebrate so, hmmm, no.

And it fails to address the misunderstanding between ref and player. It maybe reveals it. So instead: explain the misunderstanding (you are humans, you can do this) and consider what the player really wants: they don’t want to know whether they can enter. They want to enter.

This reframes the player’s position in the context of the roll as “Do I KNOW how to enter safely?” At this point as ref I would be confident saying:

Risk is REVELATION (because I have an idea — the revelation will be that on entering you get the eternal enmity of the owning god but we won’t say that out loud right away).

Don’t piss off the ibis god. You won’t like them when they’re angry.

Fail is death. You KNOW you can just enter. But you can’t. This is extreme but we’ve already established in the narrative that it’s lethal. Say it out loud. The player can still decide against acting. They want to enter but given the evidence maybe it’s wiser not to try. Not only will they die, but thanks to the risk their they die entering and the god (which in this actual case the players need to continue the path they’ve chosen) hates them.

Success with risk means that the revelation happens and the character knows away around the certain death. This is the divergence from the main narrative that risk tends to produce.

Success without risk means that the character knows a way around the certain death and the ref should narrate at will with this new direction. The god is apparently fine with this.

What I take away from this is that revelatory checks should do more than just reveal a fact. They should acknowledge the further intention of the player — what do they want to do with that knowledge — and play with that instead.

This post is brought to you by my patrons.


resolution as narrative

Sometimes the resolution system in a game is the game. You spend most of the time resolving or trying to get to the point where you have to resolve. You’re looking for a fight, for skill checks, for conflicts. And these conflicts largely define the play — the narrative is a string of resolution rolls and their interpretation.

Soft Horizon, I just realized today, doesn’t operate that way. The meat of the system, the thing that drives moment-to-moment play is in fact the interpretation of oracles. The nature of community relations and their bearing on the characters. And critically, the debts that characters owe to other entities. This is not present as a roll, as a check, but rather is underlying the ad lib work the ref is doing.

That’s a Juan Ochoa original. See previous post. Thanks Juan.

What the system provides instead is pivot points: places where things go where the ref wasn’t expecting. When a risk is realized it creates a design space for the ref: you must now decide how this risk realization changes the direction of play. Some are automatic — if someone is HARMED they take a WOUND and they will want to resolve that before anything else. The direction of the narrative pivots around their self-interest to get rid of the hindrance on their character.

More nuanced is COST — the character now has a DEBT to an entity. They owe them something and until it’s resolved, the character is often at a disadvantage. This is really a quest mechanism except as ref you are going to have to think it up on the fly. Where your narrative path was heading one way, now a character needs it to go another and so off you go. A pivot. A hinge.

It’s very hard to railroad anyone in this. The pivot points aren’t entirely up to you.

All of the risks (except the runt of the litter, INEFFECTIVENESS) are roughly like this. While only HARM and COST have a mechanical bearing, let’s look at the rest.

SPILLOVER creates a moral debt on the player not a mechanical debt on the character. Someone or something that shouldn’t have been harmed has been harmed and it’s the fault of the player’s character. The player will often feel compelled to fix this because being a fuck up stings extra when someone else gets hurt. Players that don’t feel this moral tug will be less likely to turn this into a pivot. Players that do will often use more energy to fix it than they would trying to fix a mechanical DEBT.

CONFUSION turns the situation from clear, directed action to unclear, undirected action. It pivots travel into study. It pivots combat into reconnaissance. What you knew is no longer certain. You’re lost. You don’t know who the enemy is. Something that made sense is wrong. It diverts action into search, movement into thought.

REVELATION forces the ref to invent a new hook, to bring in new information that they probably didn’t plan to bring in. Maybe that valuable thing isn’t really valuable. Maybe that uninteresting NPC is profoundly important. The narrative landscape changes. Again, the ref can’t really plan for this since it derives from the conflict which derives from play. This is huge creative space for the ref to pivot the whole narrative, revealing a deeper truth than what was previously thought to be complete.

WASTE creates a scarcity that wasn’t there before. You’re travelling from A to B and in the middle you run out of fuel. Or food. Or ammunition. This pivots the narrative to the story of either re-supply or living without.

DELAY makes something the players were going to encounter pass by. Got a meeting with the mayor? You missed it and that has repercussions. Trying to get to the airship dock before your ride leaves? You blew it, now what? Trying to get out of town before the invading force breeches the wall? No luck, you’re now in the thick of the invasion. It pivots away from an expected rendezvous and into whatever happens if you miss it. The creative space here is relatively small but easy. And the pivot is no less extreme than the others — everything can change by missing a vital appointment.

So the system provides pivot points and you don’t need a lot of these to make the narrative run like hell. One or two a session is fine — any more than that and you probably want to lean on the less disruptive ones occasionally, INEFFECTIVENESS say, just to rein in the chaos. Or, you know, go nuts. Embrace it.

Is it weird that one of the things that eases my anxiety as ref is to be forced to ad lib in specific ways? I find it makes detailed prep irrelevant, which means it’s very hard to show up unprepared. That’s where my anxiety is alleviated.

trusting the ref too much

Here’s a thing that happened in a game once (many many years ago — we may have just cracked open the box on a the freshly published Twilight:2000) that I never ever want to happen again.

IMG_0419 (1).pngThe characters were captured by some bad guys. Insert cool imagery here (I think it was a beached supertanker re-purposed as a fortress). Guards come and point to one of the characters. “You, come with me. It’s time for your execution.”

The player nods. “Okay I go with them.”

“The guards lead you down the makeshift steps — very rickety. What do you do?”

“Go with them.”

“You emerge suddenly into sunlight. It’s dazzling. Everyone covers their eyes for a moment to get used to it. What do you do?”

“I wait.”

“They tie you to a post and shoot you.”

What the fuck happened here?!

A bunch of things. As ref I thought I gave several opportunities for action to get out of this but the player never bit. Looking back on it now I see exactly why.

First, the player trusted the ref and did not believe the character would be killed so ridiculously. But the ref (me) had switched gears. This was serious business and there were ways out, but the player had to act and to take a risk. The player didn’t act because they didn’t get that this was their moment. They thought the moment to act would be later and trusted me not to kill them before their chance. To my mind, once we had poor Tim tied to the post, there was no way not to kill him. It was the stated intention, game mechanism would not longer save him (we weren’t going to roll hit and damage for a firing squad), and I felt bound to follow through.

Second, it was a system where the ref sets difficulty levels and the player assumed that at each possible action point the player decided action was riskier than inaction. They didn’t know what the difficulty would be and made an assumption. Based on prior gaming with me, for sure. But they didn’t know, they didn’t ask, and I didn’t offer. Because that’s how we played games then!

Third, the words “what do you do?” had no culture associated. It’s not a phrase that the rules command you to say to indicate you are expecting action. It’s just words in a conversation. It has no weight. I intended it to have weight but for that to happen we’d have had to have a discussion about it and agree on that. And frankly none of us were that self-aware about our gaming to realize that that was even what was happening.

Finally, everyone knew I fudged the dice. They were reasonably sure I would fake a bad roll at the critical moment and let them squeak out. But I had no more rolls to roll, in my mind.

Now there are things you can do about this without changing the rules, but they all involve changing the people and changing people is bullshit. You could say “well, players should be more proactive”. What if they aren’t? You could say “well, players should ask about difficulty levels and possible actions”. What if they don’t? Should the game just fail embarrassingly (and it did — the player was pissed, I was flustered, and we stopped early)?

How about we change the rules instead?

If difficulties are fixed then the player knows what’s what.

If the risks have to be declared then the player can make substantive choices.

If the ref never rolls dice then the ref never fudges dice.

If the culture is that the question “what do you do?” invites concrete (go to the dice) action then the cues are real cues and not just part of a conversation that might only be conversational.

So in some ways the Soft Horizon system is designed to heal this 25 year old wound.

There are no difficulties. What changes when you roll is the risk, not the difficulty. When you roll you already know the odds of success and the ref has already communicated (in a vague way) the risk of failure (or imperfect success). As player you already know you can succeed.

The ref never rolls.

“What do you do?” is codified in the text. This is what you say when you expect the players to act. Maybe not go to the dice, but certainly narrate something pro-active, something that progresses their interests. It’s a declaration that what the player says next is important.

Yeah I stole all this from smarter people. I’m not proud. It works.

when the system engages invisibly

When playtesting it can be frustrating to get a session where there are no rolls. After all, you’re trying to test the system and where was the system? But the system is more than the dice you roll — there are other mechanisms, usually, and there is also the negative space: when the roll doesn’t happen because of what it would mean if it did. That is, when the system engages the story by being declined. Exactly because of what it will do.

Yes, this is a true story.

Our last playtest session had no rolls. So how did the system engage?

Negative space: there was one situation where an action led to an offer of a roll but was declined.

JB: So no passages leading off? Just a fuck-you huge pit? I’ll try to commune with a vegetable machine. Like grab a less mobile appendage and see if I can somehow mind-meld or something.

Brad: You see nothing that seems sentient. Just vegetable labour. So nothing talks. But that sounds like SOCIALIZE to me with risk HARM. Only because you want to get close to what’s essentially industrial machinery.

JB: Hmm. [examining character sheet which has DENY for SOCIALIZE] Ah, fuck it.

Brad: But your scar counts.

JB: So 2d6 or just 1?

Brad: They aren’t insects, so 1d6. Very probable injury.

JB: Seeing the OSHA-violating implications, Marc beats a hasty retreat.

A good conflict system means that sometimes you don’t go in guns blazing: it gives you measurable factors to weigh and decide. Sometimes you do, though.

So nothing happened but we have an interesting choice: Marc isn’t that confident in his new ability to communicate with alien vegetable monsters, and also reinforces his denial of SOCIALIZE as a method. It’s not what he’s good at, it makes him uncomfortable, and adding the significant risk of an industrial accident, he lets that opportunity go by. The system is doing what it’s intended to do: the player plays into the choices they made about their character (denying SOCIALIZE in this case) and the player is making risk calculations based on the risk and sometimes not taking an action. The possibility of information from the vegetable machines is abandoned and a different path must be taken through the narrative because of the system. Working as intended.

This is the same as a party in a D&D campaign deciding to rest for the night because they are wounded, branching the narrative from pressing on wounded through the wilderness to resting, telling campfire stories, and getting jumped by bugbears while asleep in their armour. Negative space is important: if every choice to engage the dice is answered with “yes please!” then there is a missed opportunity for the system to forge another path.

And then there’s instruction to the ref. In Sand Dogs there is a ref prep sheet, just a few lines, and one of the prompts is “introduce someone interesting”:

Brad: You hear a high pitched buzz of insect wings. A speck on the other side of the pit slowly resolves as a large flying insect person. What do you do?

Dune: Hail!

JB: I’ll wave at them.

Dune: Duarte waves his arms as well.

Brad: The bug thing flies over the pit to you. It is multi-segmented like a flying centipede and has dozens of arms and/or legs. As it comes close it strokes a beard of tiny eating legs thoughtfully. “I thought you were Tik but you are not!”

JB:  “Tik has asked us to help.”

Brad: “Help how? This vegetable machine thing makes no sense at all. What was she thinking?” It throws a dozen limbs upwards in exasperation. “Would it be rude for me to land?”

Dune: “Please land.” I make room.

Brad: It lands. “Kanikalakiwinazzztakila. You may call me Kan.” It waits expectantly.

Yes, sure, you can add an NPC any time you want without a rule. But I have a rule: when you’re in a lull, wondering how to push things forward, check your prep sheet. And one of the things there is to invent someone interesting to talk to. System engaged. Kan now exists and branches the narrative. Casual statements like “flying insect person” turn into story inflections — if the characters make friends can they fly with Kan? That’d be handy.

You might think of it on your own. But with this you have a recipe and you won’t flounder because you didn’t.

Another ref cue is “recall a missed hook”. Ages ago the characters met and befriended a tough sand dog named Rachael but they left her in the previous plane. She wasn’t that interesting. So I find this cue and think, “maybe she is”:

Brad: Tik says “Well don’t I look foolish. What a waste of land the vegetable machines are.”

JB: “Harrison told you it would be good?”

Brad: “Not-Harrison did. Harrison just demanded things. I thought perhaps you were all mentally ill.” [Tik was under the impression until very recently that humans were a hive mind, all the same, because they look so similar and she had named them all Harrison.]

JB: “Well, he was.”

Brad: Tik says, “Well the meet is moot. Will you do a last thing for me? I cannot bear the waste.”

JB: “Sure.”

Dune: “Of course.”

Brad: Tik says, “Find not-Harrison and bring her to here. It seemed so important. She went by Rachael.”

Hurray! I get Rachael back into play. I liked her then and now she’s way more important than previously. Now it’s implied that she’s a planewalker and maybe an enemy or rival of this Harrison villain. The story opens up wide.

The system is more than just the conflict mechanism and its side effects. It is also in how the conflict system inspires people to sometimes deny operating it. And it is in all of the bits and pieces that guide the referee. Don’t focus too hard on your conflict system — there’s a lot of other game to play.


designing to counter ref stress

I design games for me. I sell them because I think some of you are enough like me to make them worth your time.

dinner diceI am coming to realize that one of the things I like is really an old-school feel to a game — that is, it’s about well-defined characters interacting with an environment to discover, explore, and grow. This does not, however, demand old school mechanisms: it’s not clear to me that they were ever the best way to do this. For me.

So aside from that, the new designs (the Soft Horizon work) are built to accomplish that but relieve my particular pain points. My stresses. The things that cause me to feel relief when a session is canceled (which is a sucktastic way to approach gaming). Here are my pain points:

Prepping story. I am not a plotter. I don’t want to plot. I want to discover and enable the players’ interests. Current designs therefore avoid this by providing motivation and twists mechanically. Motivation is baked into characters through WOUNDS and DEBTS: when you have one of these you need to solve it or your functionality is degraded. And they derive from prior action (though you get one for free when you start play) so they are regenerated throughout play. Twists arrive through the risk mechanism: when a risk is realised the onus sometimes is on the ref to create new information and inject it into the story. Weirdly, I am fine ad libbing this in a well defined way but not fine writing it as prep before play. I’m pretty sure I know why this is: I don’t like homework and so much homework becomes irrelevant in play. By contrast I find that inventing an amusing twist based on immediate information is fairly easy — and when it’s not, I offer a risk that doesn’t do that. Mmm, I’m relaxing already.

Figuring out what to do next. This is a place I need to prep — what will happen if the players aren’t moving forward in pursuit of their own goals? But I don’t want to do a ton of preparation — I want to do the minimum necessary to get play progressing. And so I have a set of very simple cues that you can fill out before play with one phrase or less. You won’t use all of them in play. You might not use any. And any that you don’t use stay on the prep sheet for later. So what happens is, before the first session when you’re super hyped and full of ideas, you jot down a dozen or so half-formed ideas based on cues provided. And then before every session you have a look and maybe update one or two.

Controlling player satisfaction. Making sure the players are happy is a major stress point and this is where system so often lets me down. It lets me down when it takes more time than it’s worth. It lets me down when the mechanism makes someone feel bad (they get mind controlled or they die or they lose a level from undead draining or whatever). So the system is fast. Nothing takes a terribly long time to resolve. There are specific mechanisms to stretch out the narration when one roll and a speech feel inadequate, but you choose to extend the process only when you want to extend it and not based on some external pressure like “it’s combat time”. And characters don’t die unless they want to. And it turns out that they want to more often than I thought. In fact they die (or otherwise choose to exit play) more often than in games I play where death is dictated by the rules. Because I usually cheat in those games.

Cheating. I don’t want to fudge the dice. I don’t want the onus to be on me to arbitrarily resolve something that should be mechanically resolved. But some games have random outcomes that are not fun. Not for me and not for the players. So instead two things happen in this game: results are always fun (or at least never not fun) and I as ref don’t roll at all. I can’t cheat if I don’t get to roll.

This is the purpose of the Soft Horizon games: quit making me feel sad about playing and happy about cancelling. Most other games, honestly, I’d rather just chat with my friends than play. But I didn’t realize that until I made a game that fixes it.

The Soft Horizon SRD is available for free from

The King Machine, the first instance of the Soft Horizon is available in PDF and hardcopy at DTRPG.