Managing the narrative complexity of Cultus: A keyword based approach?

[Le_Wang]

Alright people, i'm not going to lie, but the degree of narrative complexity i'm seeing thus far, both from design document and what proposals that myself and others have made...
This is going to be a programming nightmare... or is it?
You see, while I haven't seen this mechanized explicitly, one of the features of Apocalypse World (and all other titles that built on it, including Blades in the Dark) used to define the fiction NARRATIVELY was through keywords. E.g. when creating a cohort in Blades in the Dark, players give this cohort qualities and flaws drawn from a pool of descriptors. E.g. a gang can be LOYAL, but SAVAGE. Now, this doesn't have any mechanical implications per se, but if you've watched people GM/play games like this, you've seen instances where, when decisions to be made in the fiction, people will often call up these keywords for reference.
This, imo is a powerful concept, because it allows us to
1. Break down otherwise complex actions into simpler components e.g. Using a ritual murder to set an example would have VIOLENT and RITUAL and PUBLIC as potential descriptors
2. Have a progression system that boosts certain type of actions based on their descriptors. Taking the example above, the player would get bonuses to his/her success based on any bonuses they have to these specific keywords.
3. It's important to note that keywords don't have to be limited to descriptors, but stats and resources as well e.g. FOLLOWER or WEALTH as keywords
There's probably a few i'm missing, but off the top of my head
4. Programming wise, this should make selecta's life INFINITELY easier (tag keywords to decisions rather than individually having to build every decision point as a unique choice instance with unique mechanics and checks)

Qualifiers
PROMISCUOUS (Action that is sexually explicit in nature, or relies of loose attitudes to sexuality)
VIOLENT (Actions that result in visible and tangible physical harm, including death)
RITUAL (Actions that involved a tangible religious element that goes beyond the practical nature of the task itself)
PUBLIC (Actions that take place in the public, or are devised to appeal to public attention eventually)
SECRET (Actions that take place outside of the public eye. Bribes, for example, would be CIVIC and SECRET)
CIVIC (Actions that involve the civil service in one form or the other. Resisting a police enquiry would be CIVIC, but resisting a police arrest would be CIVIC + VIOLENT)
PATRIARCHAL (Actions that invoke/enforce male ascendancy)
BENEVOLENT (Actions that are explicitly charitable in nature)

Resources
WEALTH (All actions involving Wealth, either acquiring it, or spending it)
INFLUENCE (All actions involving Influence, whether to check if you qualify for access, or acquiring more of it)
FAITH (All actions involving Faith, whether to check if you qualify for access, or acquiring more of it)
CHARISMA (All actions involving Charisma, either acquiring it, or spending it)
FOLLOWERS (All actions undertaken by followers, or with follower involvement)
PERSONAL (All actions where the PC is actively involved)

Others
RELICS (all interactions involving relics)
RIVAL (all interactions involving rival faiths)
FRINGE (all interactions involving criminal/extremist NPC groups - Generally signals unique decision chains, like the one mentioned in the choice/consequence thread)
ICON (all interactions involving major named female NPCs (the ones mentioned in the design document)

[Selectacorp]

Wow-- a whole new dimension here you've unveiled-- like it lots!

[Le_Wang]

Warning: Long exposition - Skip to keywords if you want to know about how those work, and don't care for all the build up
Before we talk about how keywords fit in, we need to look at how actions will be implemented in game
i think there's 2 routes to go down, depending on how much you hate RNG
1) %based choice - The one we are familiar with; namely, the %based approached in SotU and ML, where you have an innate %chance of succeeding any given action.
SotU took this one step further, by allowing you to have a "prep" action that boosted this %chance of success. It's functional, but well, how much you like this depends on how much you like RNG. Imo, unless you're a fan of tabletop gaming, most people span from being indifferent to it, to HATING IT WITH A PASSION, mostly because of instances of "missing an 80% action 2-3 times in a row", which, statistically, isn't impossible, but in practice, is difficult to digest ESPECIALLY if the % system is supposed to be simulating complex actions (as opposed to just hitting/missing in a swordfight).
2) >100 - The alternative is to have a "build up to 100%" approach. I.e. the action has a base number which starts higher or lower depending on it's difficulty. Let's say it's 50 for a medium difficulty challenge. The only way for the action to succeed is to get that number past 100. The way you do this is by performing other actions, or having met sub requirements. E.g. If you perform the "preach" action, this will increase the base number by 10 for all NPCs in the AoE of the shrine.
Note: functionally option 2) is, for most intents and purposes, the same as 1) in how it works, the only difference being that it doesn't let you perform the task before you've guaranteed success.
In terms of tradeoff, it's more of the feel factor

1) Gives players more freedom, but also causes more "#rngfeelsbadman" moments.
It also gives more design possibilities in terms of "punishing players when they fail"; note, if there are no big consequences for failure, then the RNG approach often takes away from the strategic aspect; most players will generally tend to just spam low-rng actions until they get lucky, rather than actively try to boost their chances of success; The above is far more meaningful that people think
Warning: Math incoming
Say an action has an 40% chance of success natively. A supporting action can boost it to 80% success chance. In practice going for 2x 40% instead of prep into 80% has a 64% of success, making the prep action a better statistical use of your time. However, because the prep action CAN result in players failing the roll, most people will just go for the 2x40% route instead. In fact, if you consider the very real possibility of failing the 80% roll an extend that to three rolls, there is a ~80% chance of succeeding by going 3x40%, and a 96% chance of succeeding by going prep+2x80%. In absolute terms, therefore, using the prep action doesn't give you THAT much of a better chance over 3 rolls.
Translation: Even if you might think you're being clever by giving people the statistically optimal choice, because most human beings struggle to rationalize statistics, a lot of people will typically end up disregarding any element of strategy, and just roll the dice on >40% success chance actions, especially if there are no consequences.
What if there are consequences: It's touched upon in the design document, but balancing success chance vs. the consequence is a tricky thing, because you then have to weigh the mechanical penalty against the expected outcome of the roll. If there are ways to boost your chances, that adds another layer of complexity; a punishment might be fair for a given chance of failure, but feel excessive/trivial if you can adjust that number up or down. The consequence also has to make fictional sense; trying to murder someone should have a negative consequence regardless of whether you succeed or not. you would... anyway, it's all very complicated, and why many games default to the same consequences regardless of success or failure.
In practice: Systems where you can affect the %chance with prep actions generally see a lot play only in systems where a) there's no possibility of success w/o them (i.e. base chance of success is 0%) or (b)if it guarantees success (100%). This is the reason why option (2) makes for a great compromise, because it's essentially an RNG system that compels the player to go for (100%), and is easier to balance. More on that below.

2) Gives players less freedom on the surface, but feels better in practice
The main argument for this approach is that it's easier to
a) Balance
You just have to worry about the base difficulty, and how challenging it is to clear the 100point requirement cap.
b) Build up interesting decision points
Does the player build up shrines so they can lower the difficulty requirement across the board for all NPCs in the AoE? Or is there an NPC at the edge that they can get w/o having to go that far...
c) Most importantly, it's easier to play
Even though they are very similar in how they are designed, route 2), because there's no probability involved, is far easier for the average player to connect with, understand and, accordingly, play.
Consider this; if you're a D&D vet, you know that Str 20 is great is because that gives a +5 to a lot of stuff, and translates into a +25% chance of success and +5 min/max damage which, given how base HP and difficulty checks work, means you are now guaranteed to one shot far more low tier enemies, and have a decent chance of hitting harder to hit enemies hard. If that didn't make a lot of sense to you, then you are probably not a D&D vet, so more likely than not, you are going to undervalue a +2 Str bonus at 18 Str (because it sounds lame, and feels like a marginal improvement) compared to getting a perk, which has ALL THIS COOL TEXT attached to it.
TLDR: Statistically complex game are only fun for people who think statistics are fun, which isn't most people.


What about keywords?
Essentially, the game will keep a stat sheet (maybe visible, maybe not) of a list of Keywords.
At the beginning of the game, all of these Keywords will have a +0 to them. As you progress in the game, these numbers will start going up (acquire new tenets, grab relics, get specific upgrades, make decisions).
In parallel to this, every action/npc you undertake/interact with in the game world will have a number of Keywords(w/e make narrative sense) attached to them. When you select these npcs/actions, the game checks your stat sheet for all relevant keywords, then gives you a bonus accordingly.
What's the point of doing this?
If you look at SotU, there only way to affect your chances of success with each individual NPC is on a case by case basis. Shrines and map based AoEs (which I proposed) lets us have elements that affect NPCs and interactible "regionally". However, if we want things that both have GLOBAL yet SPECIFIC reach, like tenets, then we need a way for the game to both KNOW that this effect is global, and in WHAT WAYS the effect is specific.
The Keyword system lets you 1) Record and recall these specific global bonuses and 2) gives you a way to assign case by specificity to each NPC/Interactible
I'm confused: what's the difference between Keywords and Descriptors and...
I'm using the term "Keyword" as to reference a "specific word or group of words" that will have a game Mechanic attached to it. Not all narrative descriptors (say: Disillusioned Preacher) will be Keywords, but some narrative descriptors (those from the list in the OP), will be Keywords.
Special descriptors
For those of you who've keeping tabs on the tennis game across multiple threads SC and me have been playing, he asked how we could use Keywords to implement "special NPCs", e.g. for the purposes of determining who's a descendant; imo, you can just have DESCENDANT as a Keyword that is part of the stat sheet, to be recorded as it increases but you also want to be efficient with the number of Keywords in the game to limit design clutter and complexity. Imo, ICONS (or w/e Keyword we use to describe the super NPCs) are good enough as a requirement for the "descendant" feature he was describing, and will exist in the game as their own feature anyway. Letting these ICONS also serve as DESCENDANTS for the purposes of qualifying for the old faith ending is just expedient.
So i read through that part about %based choice and >100 choices; how does this change based on keywords?
You did? Uh, i mean that's great. Well, it doesn't really matter which of these systems you use; The Keyword system works fine regardless. If you're going for %choice, it adds 10% to your chance of success, where it would otherwise add a flat 10 if you were going for >100.

Now this all makes intuitive sense to me, but i doubt it's that straightforward for people who aren't into applied mathematics or game design. Feel free to ask questions.