When you hear the phrase ‘choices in games’, what do you think of?

It might be this:

Image result for fighting fantasy

or this:

or indeed this:

But equally it might be this:

If you can’t read the text, the player’s just moved a Great Prophet to a city and chosen to use them up to build a Wonder

or this:

Image result for ftl game
The helpful captions aren’t mine – they’re from a Steam strategy guide. I love the hectoring tone of the advice. ‘PAY ATTENTION, ENSIGN!’

Lots of choices! Where are you going to send your crew, what are you going to repair, should you teleport your boarding party home and bug out?

Or this:

I have literally no idea what’s going on in this screenshot, but I take it on faith that it’s something complicated and interesting that 2% of my readers could explain to me.


The first three sets of choices are all what I’m going to call tight choices: at a specific point, you have a specific set of options. You pick one and move on.

The second three sets of choice are all loose choices. You don’t have to make a choice at all: you can just watch things happen. When you do make a choice, there are often bundles of similar choices (use the Prophet here or in the next city over; repair the shields now, or in five seconds from now).

I didn’t want to say closed/open or finite/infinite because I don’t want to suggest a value judgement. There is nothing fundamentally better about the second kind of choice.

Generally, loose choices are more system-driven, and tight choices are more scripted. This is almost a tautology – the thing that makes tight choices tight is that they are hand-designed and hand-placed.

All the tight choices above are all choices from explicitly narrative games. This isn’t a coincidence: narrative games tend to be more scripted experiences. But lots of games contain both kinds of choice:

Image result for civilization iv diplomacy
Tight choice in Civ IV


Loose choices in Sunless Sea


In fact, this is what we expect from a CRPG. You do your walking, killing, spelling and looting in a 2D or 3D world as loose decisions – do I cross the lake or look for flowers on its edge? do I do some or all of the quests in this zone first? and when you need to make important or characterful decisions, you usually make tight decisions in a dialog tree – do I nuke the city or not? Do I side with the templars or the mages? Which school of poetry do I pick?

So I said above that narrative experiences tend to be more about tight choices. And even in a hybrid game, the most narrative-y bits are tight choices. But there have been, and are, attempts to create narrative games that rely more on loose choices, and some of these attempts have been really, really fun. This is where Cultist Simulator comes in.

I’ve seen (generally speaking) two kinds of approach when people try to make stories with systems and with loose choices: Grand Ambitious Projects and Focused Indie Experiments in Poetic Design.

Grand Ambitious Projects often try to solve a big problem in interactive storytelling – sometimes even the general problem. I’m generally pessimistic about their chances (although I think they often throw up fascinating insights along the way) and I don’t have the ability, inclination or patience to do that. So I want to be clear that I’m not trying to solve any big problems here, I’m just trying to do my own artisan indie thing over in the corner. Right? Right.

Focused Indie Experiments in Poetic Design are things like Papers Please

or Her Story

or Rimworld:

All of these have scripted material, of course. But they also have a game-mechanic layer that you spend most of your time interacting with, where you make loose decisions constantly to narrative effect. That’s what interests me right now – though I don’t have any grand reason for this beyond creative curiosity. So here’s my thinking for Cultist Simulator.

I want to make a narrative game about the tension between the prosaic and the mundane, and the temptation of the sublime, even when it’s self-destructive. So I could do something like this:


That looks familiar, doesn’t it? Tight choices, branching or stateful narrative, something like Twine or Ink or a homebrew system. I’d probably do a decent job and I’d have fun making it. But that was very close to the kind of thing I was doing at Failbetter (when I had time to write at all), and I want to bang some other ideas together.

So I’m going to try to do something like this (WIP; sorry for the horrible greybox and placeholder art)


We’ve got a basic dichotomy there; Work or Dream? (There are another five usual actions, but that’s where we start.)

We’ve got any number of things you could use your Work action on; here, we’re giving up some of your health.

When you use this combination, you’ll get a Labour activity which will count down, use up your Health and gain you some cash. You might instead have Dreamt using Reason or Health or a number of ritual implements; you might keep spamming Labour to get the money to live, or you might find a number of other ways to support yourself, or you might descend into obsession with all that dreaming. It’s the same thematic choice as you would have in a more traditional branching game – it’s just presented in a more systemic way.

I want to try to make the whole narrative of ambition and disaster out of these combinations and options – no menu-driven dialog-tree tight choices at all. That will last until and unless I find that this constraint is getting in my way. I’m not a purist about this.

I said above that loose choices aren’t better or worse than tight choices. They’re not. But it is what excites me at the moment. And when you have your Dream/Work dichotomy alongside a passel of troublesome Followers who may need disciplining, a Hunger that needs to be fed either Health or Followers, an ongoing Investigation by a stern gentleman in dark blue – and, if you want it, a work promotion – all as loose choices for the player to juggle as strategy alongside the narrative – then I think we’ve got something fun.


3 thoughts on “Cultist Simulator Devlog: Loose, Tight

  1. Personally, I believe that to achieve a loose-choice based narrative the game has to first off be “real-time”, and the NPCs need to have their own objectives and motivations. Let’s say you kill a child in the presence of a parent NPC, then revenge becomes the motivation for that parent NPC, and their actions are based on a set of skills and resources available to them. So let’s say they decide to hunt the player down for revenge, and the NPC is a Ranger, then it’s very likely that the player is gonna be hunted, ambushed, etc. until either they are dead or the NPC is dead. The interesting aspect to this is that when you die, you can visit your timeline and reset back to any point in your timeline.

    Here’s how that might work. Let’s say you, as the player, can kill any NPCs, and you have abilities that make that death either clean, or prolonged. Let’s say you shoot an NPC, but because that NPC has its own motivation to survive, it runs away. You either give chase, in which case any other NPCs that were present would be motivated to stop you, or arrest you, etc. Let’s say you kill them all, that leaves a bloody trail of deaths in your timeline, and if the NPCs you killed are pre-assigned parents, siblings, spouse, etc., then that related NPC hunts/haries you as the player. Now let’s say you made the NPCs that you killed “suffer”, then if/when the NPC’s/related-NPC hunts you down, they choose to kill you slowly giving you some exposition of why they are torturing you, then if you die, you are presented with your timeline, and you know to reset to a time before you killed the NPC, and then choose to either kill them a different way, or not kill them at all. Let’s say you killed the NPC cleanly, without even a second thought, then the relative NPC may just ambush kill you, and you’re presented with your timeline and you have to figure out at what point you might have pissed off a “relative” NPC to hunt you down.

    I think the key is to give the NPCs their own personal motivations and objectives. Do they, when harmed (either directly or indirectly) seek revenge or do they seek justice, etc. Let’s say you kill an NPC that may have had a quest for you further down that NPC’s own timeline, then that chain of quests aren’t available to you, so you may have to figure out why. Part of that might be asking other NPCs what they know about the missing person, and you might put together that someone killed their child (that someone being you), so you might have to restart your timeline WAY back and then “redo” all those choices, or if you’re like “screw that I don’t wanna redo all that ‘work'”, then you suffer the consequences. Let the Quests guide the the timeline, and let those objectives be subject to previous impacts on the NPC/target that may have been perpetrated by the player. Let’s say that the only way to kill a “boss” might be with a weapon that you obtain through a quest line, then the player may have to reset multiple time to figure out “where they went wrong”. To make it a little less frustrating, each time a point in the overall timeline is revisited, there’s more exposition/narrative either given (or made available by other NPCs if you choose to interact with them) to help the player figure it out.

  2. ZacWolf’s comment just reminded me of the nightmare of following a series of Windows Restore Points to undo something malicious that happened. I’d rather not make a game of that. It also reminds me of having several fingers tucked into the pages of a Fighting Fantasy gamebook, each one representing a ‘savegame’ of deeper decisions. That could become a confusing affair too (oh, but we all did it! So there’s something to be said for a systemic peace of mind feature. You could even use literal fingers as an homage to the Choose Your own Adventure method of saves. Represented, perhaps, by an occult Hand of Glory?)

    Personally, I always opted for timer bars to freeze whenever I opened the Actions menu in a game like Final Fantasy. Anything else was an exercise in frustration and haste-based mistakes, not tactics. But that’s just me.

    Some mixture of the two might help here. There are timers we would choose to always run in real-time, like the Work spam, the daily grinds that sustain our hearth and health. And then there’s timers we’d prefer to handle manually, and with a clearer head on our shoulders. (To summon the metaphor from before, it’s a bit like automatic updates. Some programs we’re quite happy to let manage themselves. They’ve earned trust by never going horrendously wrong on us, or by being vital to security. Then there’s optional updates, some of which have been known to crash systems, are effectively bloated nagware, or have turned out to be kind of snooping on us.)

    With so much smallprint potentially scrolling by on all sides of the UI, I’d like to have my hands on the wheel when pivotal decisions are being made, but I’m quite happy for the menial labours to ticker-tape on by with barely a glance and a thoughtless click (we might even want to establish a set of Good-Til-Cancelled automated clicks. Like a Direct Debit.) That’s where design will play an important role. In Fallen London, devs could use bold text to warn players of potentially costly or irrevocable decisions. In Cultist Simulator, something more colourful might be required since the luxury of time isn’t on the player’s side, from what I’ve seen so far.

    If Cultist Simulator is to be a completely loose game of button-bashing real-time timers, I’d dabble with it just to milk that lovely Alexis lore. But as a game, I’d retreat from it. I want to like the intended experience of it! But I may be just as happy consuming it via Wikia contributions over a cup of tea.

    Wishing you the best of success, with this and the rest.

    ~Shalinoth / Jay.