COMMUNITY - FORUMS - SOULBORN ENGINE
Socioeconomics and NPCs

I'm really looking forward to some more specifics on NPC behavior. If I'm reading between the lines of Caspian's comments about using "Five Factor Model" psychological theory to populate individual variables in personality, ambitions are very high! I hope they're realized.

If the devs are still taking suggestions, here are some other features I'd like to see:

NPCs able to interact with the economy based on the locality's economic variables. If just being in a city means you're assumptively getting food and water, it stands to reason its citizens have some knowledge of the cost of basic goods and services too.

When the system of agricultural units and their production is determined, then cities should have their own dynamic profile of running commodity prices. This way NPCs know roughly how much or how little to pay for something. The smarter or more specialized they are, the less likely they are to 'get taken.' (NPCs can then participate in initiating contract negotiations as well?) Since the economic profile is unique to the city--it keeps the local flavor and doesn't unduly homogenize the Kingdom. (How boring would that be?)

I do think if an urban center is assumptively furnishing food and water, then that should require ample food sources supplying it. (I know there's a requirement for wells, but I'm not recalling anything about food sources.) It can be as simple as the good old Warcraft (1 farm: 8 units) formula, or maybe contracts in place with food producers, who then have to meet the production quota or else the town's food supply dwindles enough to raise food prices.

A significant enough rise in the city's food prices should cause some unrest. If fewer people can be supported, then citizens on the bottom of the food chain should emigrate, protest, or revolt--and/or switch from their profession to subsistence farming on the outskirts of town (land leases be damned!) The motivation to turn to crime should also be ticked up. When food productivity rises and there's a glut, food prices should drop and farmers think about developing other skills...

...and with this mechanism of agricultural/food supply/social stability cycle in place, some very fun possibilities about more specialized economics become tantalizingly easy to add...

I know there's only so much computation possible and there's a lot of priorities already on the table. But I really do think there's a dividend in adding a system of "economic consciousness" to NPCs. With a little forethought, I bet it can be accomplished with lean RAM/server demands. No need to try to recreate "Tropico," but that's not a bad example of a superb...and lean...abstraction of the give and take between Maslow's Hierarchy of Needs, economics, & politics that would go a long way to manifesting what Soulborne was conceived to do.

The payoff is a realistic challenge for aristocracy and nobility government. (It is the job they wanted, after all.) It's a more realistic environment for townsfolk, too. The wilds have their intransigent hardships...so should city life. With NPC social cohesion affected by PC decisions, that's game-worthy drama. And it helps keeps the promise of "if it would change, then it does."


DPBoD2.jpg

4/21/2017 11:24:51 PM #1

Some more thoughts: NPC migration

As mentioned previously, loss of food supply (or housing for that matter) should give NPCs some motive to travel somewhere they believe is having fewer troubles, as well as increase their likelihood of 'deviant' skill development. When basic sustenance and safety are threatened, it just won't bring out the best in people.

However, there're other legitimate triggers for NPCs to make drastic life changes. Higher up on Maslow's Heirarchy of Needs, there's 'Self-Actualization.' If they're doing well enough on all the basics underneath, then they might just invest any nest-egg in themselves.

If we can assume the "OCEAN" phsych build for NPCs can produce a "Career Change Risk Aversion" value for an NPC, perhaps they should periodically check it against the opportunity to switch careers based on their skill sets, or skill aptitudes & training availability. If [perceived probability of success] x [potential gain in income] + cash on hand > [career change risk aversion], then they just might make a drastic life choice if they find their current station in life unfulfilling.

In a disaster, unfortunates on the low end of the socioeconomic scale aren't blessed with the most reliable information about what opportunities really lie where. Migration vector may as well just be random.

But if cities, institutions, and realms have Reputations as do their inhabitants, then perhaps that's partially manifested by a distance value--and NPCs would take that into account when they make these big life decisions. A famously successful and bountiful realm ought to be an intentional target for the economically displaced. A skilled craftsman may hear of better wages in a city down the road, and take the leap of faith even though his survival isn't threatened by staying put. The same equation may even tie into a rogue's decision to stay in town or move on---depending on how much 'heat' is on him. And NPCs stay automatically responsive to the technological advances of their evolving economies.

Once again, I don't think it has to require too much computation to make this happen.

Aim high, esteemed developers! Assure me of enough happy specifics on NPC sophistication, and I'll finally shell out some $Dough$ for you.

(I'll spare you my ontology on NPC reactivity to 5 senses & object sensory profiles. Maybe for the sequel?)


DPBoD2.jpg

4/26/2017 12:08:46 AM #2

Consent of the Governed:

So this post is pretty much 'whipped cream,' great if you can swing it. Might be a bridge too far. No worries.

Going back to "Tropico," with a nice, manageable equation for 'happiness' levels per Maslow category, a citizen's complacency with the government is generated. In the system I've suggested, an NPC has access to some comparisons of economic prosperity at home and a certain distance away. When an NPC gets unhappy enough to think about moving yet doesn't have the means or courage to move, s/he can use that comparative data to direct their ire!

Is it the Mayor who's underperforming relative to neighboring towns? Or is the Count? Is the whole Duchy in a mess? Or a Kingdom on the verge of rebellion?

New checks against that fancy OCEAN profile are made: do they have the gall to openly protest? To take arms? Or do they quietly starve to death when the Queen suggested they eat cake? Does the percentage of government soldiers stationed in the town have a bearing on that decision?

Surely PCs will be the ones leading the charges, but maybe this system of NPC economic cognition can help determine how much support the PCs have....

Times change and people get more spoiled. I'd say the minimum standard of living to keep NPCs politically docile ought to slowly increase as Elyria's 10 year plan rolls by.

Like I said, there's no need to recreate "Tropico." But a good system of NPC responsiveness to economic conditions will really make Elyria come to life for players in a way no one can completely guess. (Fun!) Food production simply becomes important--and defending farmland has some appropriate urgency for the realms' political security.


Looking forward to seeing what you come up with re: NPCs! I'm sure you'll blow me away.


DPBoD2.jpg

5/7/2017 7:36:58 PM #3

"There go my people. I must follow them, for I am their leader."

If CoE really does facsimilate the NPC sensitivities of Post 3, then extending this out to NPC rulers is actually pretty simple. When the people are up-in-arms, their rulers have to dance the side-step or risk finding their head in the guillotine. (Talk about motivation!)

With some economic statistics defining political stability of the realm, its ruler has some choices:

Neutralize a destructive agent. [Is there a Grendel responsible for the destruction of farmlands, or PK mob, or invading force? Throw the army at it? Ask for alliance support?]

Acquire missing resource. [Quickly settle a new region and rev up production, make an import agreement, or invade a neighbor?]

Of course, not all rulers are going to be inclined to address the real problem. Some will think 'Let them eat cake!' and otherwise take advantage of the situation to...

Propagandize! [Blame a political rival. Blame the people themselves! Beg for patience, or maybe just simply quash the dissenters for some ruthlessly efficient 'economic restructuring.']

A realm can be made to do its best to keep its economic profile within certain ranges in order to avoid political crises. There should be a gravitational force towards 'steady state' placidity, and it might be assumed that the government successfully decrees these actions intended to stabilize most of the time. Yet...this is also where the NPC's "Big 5/OCEAN" capabilities can find a real canvas...how NPC rulers decide to govern--and plenty of them will have no qualms about tipping the apple cart.

So this chain of thought sketches out a way to make NPCs believable C's....Characters. Instead of spritely meandering in circles waiting for PCs to trigger their scripts, NPCs trudge along like we all do, figuring out how to get by in the world. PCs up and down the ladder of power will have their hands full...as it should be. With the chaos that brews and spreads from the human condition, PCs will have plenty of opportunities, too.


DPBoD2.jpg

5/7/2017 8:25:36 PM #4

This is extremely complex AI behavior, and while I would love to witness such depth that one cannot distinguish between a Person and an NPC, its fairly improbable.

Turing test aside, I do wonder how much effort SBS will put into the development of realistic AI.


5/7/2017 8:41:38 PM #5

I'll admit it's very easy to let it get overly complex, and therefore completely unfeasible.

But the idea here is for a relatively small routine sequence of NPC 'need' checks to reference another contained dataset of a location's economic indicators. The computation doesn't have to be overwhelming at all to give NPCs a vast array of behaviors that simply 'make sense' and can 'Wow' the player witnessing it regardless of economic/social strata. (Think of the good ol' Lorenz Attractor...a simple enough equation that still manages to produce an unfathomable array of possible expressions.)

Several strategy games/population simulators have been doing exactly this for many years. Without getting anywhere nearly as nuanced, CoE could anchor its NPCs into the vital game mechanic that a government rpg deserves.

The developers have dropped hints of stuff far more ambitious with regard to NPC psychological profiling--so if those expectations aren't too high, then my pitch here shouldn't be either. (In fact, I don't rule out the possibility they're already ahead of me on all of this.)

It's all a long way from approaching, or even aiming for, passing the Turing Test. We're not talking about conversational language generators.


DPBoD2.jpg

5/13/2017 5:35:51 AM #6

This was a really interesting series of posts. Never played Tropico, but maybe I'll have to check it out. Thanks for a thought-provoking read!

5/13/2017 5:45:45 AM #7

If any of you is interested in the more statistical approach to phycological test, i would recommend the International Personality Item Pool.

IPIP

In order to introduce certain AI behaviors i think that the state machines have their uses.

However, the last time i studied anything remotely related to computer science was two years ago so...


5/13/2017 6:45:14 PM #8

Posted By Huntsmaster at 12:35 AM - Sat May 13 2017

This was a really interesting series of posts. Never played Tropico, but maybe I'll have to check it out. Thanks for a thought-provoking read!

Thank you kindly for reading it, Huntsmaster! If you take a look at Tropico & develop a lifelong love for Caribbean music, don't say I didn't warn you!

Posted By Hludowig at 12:45 AM - Sat May 13 2017

If any of you is interested in the more statistical approach to phycological test, i would recommend the International Personality Item Pool.

IPIP

In order to introduce certain AI behaviors i think that the state machines have their uses.

However, the last time i studied anything remotely related to computer science was two years ago so...

Haven't seen that one--Looks like an interesting read! Thanks Hludowig!


DPBoD2.jpg

5/14/2017 7:09:34 AM #9

Posted By Ecir_Edyah at 7:45 PM - Sat May 13 2017

Haven't seen that one--Looks like an interesting read! Thanks Hludowig!

I am happy to offer you a new approach to psychometrics of certain personality models.

Posted By Ecir_Edyah

No need to try to recreate "Tropico," but that's not a bad example of a superb...and lean...abstraction of the give and take between Maslow's Hierarchy of Needs, economics, & politics that would go a long way to manifesting what Soulborne was conceived to do.

Besides the Maslow's Hierarchy of Needs, there are more theories that could be interesting to see implemented.

For example, we have the ERG theory of Alderfer, which is a "reinterpretation" of the Maslow's Hierarchy. Alderfer divide the hierarchy of needs in three categories:

-Existence Needs: The first two levels of Maslow. Security and physiological needs.

-Relatedness Needs: The next two levels of Maslow. From the "esteem needs" level only the external esteem.

-Growth Needs: The last level of Maslow plus internal esteem.

The difference with Maslow's Hierarchy is that Alderfer don't think that such a rigid relationship exist between the different categories of needs. Is possible for someone to fulfill his Growth Needs even when the other two are not completely satisfied.

Also, each person may have his own personal preference when trying to satisfied his needs. In that case, if a need of superior order, in the personal hierarchy of said person, is not satisfied his desire to satisfied a need of inferior order will increase.

Another example would be the McClelland theory.

This theory introduce a new set of needs for each person.

Need of achievement: People who desire to "make things better". People with this need will avoid to perform tasks too easy (no improvement) or too difficult (impossible to accomplish or with too much risk). This kind of people like to search solutions to problems and need to receive continuous feedback to know if they are doing their work right.

Need of affiliation: like the social needs of other theories. the individual prefer a cooperative situation to a competitive one. You can see it as my notes describe it "the desire to enjoy the acceptance and esteem of the people."

Need of power: the person desire to have influence in other people. Enjoys status and competitive situations.

There are many options, though i don't know their viability. I also don't know if this post of mine is necessary, but it helps me remember my notes so is not that bad.


5/14/2017 8:11:07 AM #10

Awesome post, Hludowig! Since all we can do is kill time until SBS chimes in with details, then you're post was absolutely necessary!

I like the 'simple' take for Alderpher. For programming purposes, it's not so much the way we group the related 'needs' as it is how they're sequenced:

Combat threat? (Y/N) if Y Combat mode, if N then check...

Breathing threat? (Y/N) if Y then Find Air, if N then check...

Environmental threat? (Y/N) if Y then Escape Environ, if N then check...

Physical Damage threat? (Y/N) if Y then Seek Medicine, if N then check...

Dehydration threat? (Y/N) if Y then Find Water, if N then check...

Starvation threat? (Y/N) if Y then Find Food, if N then check...

Sleepy? (Y/N) if Y then Best Affordable Sleep Boarding Within Range NPC's own room if possible, if N then check...

Thirsty...

Hungry...

So far this sequence seems fairly easy to peg, although splitting hairs is always possible; all of this is applicable to animals as well as NPCs. In fact these are pretty much basic minimums for CoE to execute a responsive ecosystem for any particular population biology. So far I haven't picked up any clues that they intend to cut corners here. If that's true, humans aren't adding too much complication:

Moving forward...

Employed? If Y, Execute Contract Terms for Day, if N Acquire Currency [which gets into the Psych matrix for determining how to go about it, but for the most part Inventory Personal Skills; Sort by Probable Profitability; Filter by Personal Morality; Filter by Personal Risk Tolerance, Seek Contract If Unsuccessful, Repeat process for next most-profitable skill. If Successful...

The 'Seeking' range here is determined by the character's transportation assets as well as those Economic profiles of the localities within that range.

Needs not requiring daily attention but getting a check routine every now and then, with individualized magnitudes and frequencies determined by the Psych profile:

play

socializing

home ownership (?) [other economic goals]

mating

spiritual health

career advancement

This order is much more flexible in my mind, and might start manifesting differences in Need Theory, Hludowig. Whichever way it's sorted, the beauty of it is that any need check can refer to a common syntax: identify solutions in range, sort preference, and attempt to satisfy.

And the need checks don't need to be running dozens of times every second. Once the Sequence is determined, it's the Needs that are either scheduled to occur or triggered by encounters. When a need is satisfied (or not satisfied by the check's duration deadline,) it updates the character's contentment level (with any appropriate political implications.) Then it checks to see the most urgent need lacking attention and runs the new routine.

Hopefully this isn't drivel. Apologies if it is...I'll edit it in the morning. :P


DPBoD2.jpg

5/14/2017 10:38:56 AM #11

Posted By Ecir_Edyah at 12:24 AM - Sat Apr 22 2017

Aim high, esteemed developers! Assure me of enough happy specifics on NPC sophistication, and I'll finally shell out some $Dough$ for you.

(I'll spare you my ontology on NPC reactivity to 5 senses & object sensory profiles. Maybe for the sequel?)

I can't be sure, but perhaps if the "pre-elyria" offline demo includes static NPCs, our first cracks at generating OPC scripts could pave the way for greater complexity. If that's the case, you can be pretty sure they will be flexible.

All we need then are measures and methods.

In regards to Having a "config" section that's clearly labelled, I had planned on doing exactly that. Makes life so much easier to tweak and change even more so for non programmers.


5/14/2017 12:44:32 PM #12

That sounds very interesting, Oracle! I'd be interested in seeing what you piece together! (Sounds like you're well on your way to a marketable mod, too!)


DPBoD2.jpg

5/14/2017 3:42:32 PM #13

Great read, indeed. So thanks for that!

Considering Improbable just went to be an half a billion company, I think you're allowed to think big and can safely turn your expectations up a notch or two.

Personally, I've got to calm myself down, as to not being eaten up by hype, but I certainly can't wait to see SpatialOS in action. They've got quite the disruptive tech, right there.