[suggestion] Gang modelling

[paktsardines]

While I'm at it I was having a wee think about gangs...

In the real world two people become friends through ongoing positive interactions.

In the Prison Architect world, the same could apply. Say, when two prisoners meet (ie: pass within a specified distance/radius) of each other, an interaction occurs. The outcome of that interaction (positive or negative) is determined by the 'chance of positive interaction' (cIP). The outcome also influences future interactions with that prisoner. Let's call this Friendship Level (FL).

Fundamentally these positive interactions are influenced by shared interests, traits etc. In the prison architect world, these could be criminal records, like of bacon, or whatever. We'll still call them traits.
Let's say we have a trait mapping function (Tf) which, for any two set of traits between prisoners gives an overall compatibility score between, say 0 and 100.

Introducing some some rules:
1. New prisoners have no friends.
2. Unhappy prisoners seek positive interactions (ie: will want to move towards the nearest 'friend' with the highest friendship level (FL)).

nIP(A,B) : number of interactions positive
nIN(A,B) : number of interactions negative

cIP(A,B) : chance of interaction positive between prisoner A and prisoner B
cIN(A,B) : chance of interaction negative

FL(A,B) : Friendship level between prisoner A and prisoner B
Then:

FL(A,B) = nIP(A,B) / nIN(A,B)
cIP(A,B) = (FL(A,B) + Tf(A,B)) / 100
cIN(A,B) = 1 - cIP(A,B)

I think this models friendship quite well. The more two people bump into each other 'nicely' the more they will want to hang out. Conversely the more bad experiences they have the less they'll hang out.


So, what has any of this got to do with gangs?

When you think about it, a gang is nothing more than collection of friendships between two people.
The more a prisoner hangs out with their 'friend', the more they will be bumped into (two people take up more area than one). The friends will gradually accumulate more friends this way. The more they all hang out the more they'll like each other, attract new friends and gradually become a gang, with likely similar traits.

The gang leader (if needed) can simply be the 'friend' with the most friends.


Hmm.. it sounded better in my head. Thoughts?


edit: Occasional inversion of of FL(A,B), to represent friends 'falling out', could also cause entire gangs to collapse and reform.

[Daimaju]

Brilliant!
Race should also be a huge, if not the greatest, factor.

[kongoman]

if(nIP(A,B) && cIP(A,B) > FRIENDSHIP_OVERLOAD)
{
homosexuality=true;
ShowerAction(A,B);

return SOAP;
}

[Daimaju]

if(nIP(A,B) && cIP(A,B) > FRIENDSHIP_OVERLOAD)
{
homosexuality=true;
ShowerAction(A,B);

return SOAP;
}

Omg, lol...

[paktsardines]

Hahaha... Conveniently SOAP is also an acronym for 'Simple Object Access Protocol'

But your post has also made me realise that, in reality, friends rarely 'fall out' due to occasional inversions of FL(A,B). Friends fall out primarily because Tf changes (that is, traits change and that affects Tf)

In otherwords, if, with this model, my mass-murdering neo-nazi prisoner's straight trait changes to a gay trait then he won't be as well accepted among his mass-murdering neo-nazi gang friends who've been gathering around him. Soon he'll have more friends among the pillow biting prisoners and he'll join that 'gang'. Friends of his in the neo nazi gang who didn't care whether he was gay or straight would continue to be friends with him, and may leave the nazi gang to follow him - particuarly if they too had gay traits.

Or, as another example, search and replace 'straight' with 'bacon' and 'gay' with 'vegetarian' - oh, and 'pillow' with zucchini.



edit: Forgot to also explain 'Unhappy prisoners seek positive interactions', as opposed to 'All prisoners seek positive interactions'. This ensures gangs are more likely to form when your prison is badly run (many unhappy prisoners). That is, happy prisoners are 'mostly' content reading a book (within certain margins).

[paktsardines]

Sorry, didn't include rule 3:

3. A Prisoner A, who's FL(A,B) is less than a given threshold, will try to avoid all interactions with prisoner B.




edit: Sorry, I don't think this rule is necessary.

edit2: But could probably be changed to: A Prisoner A, who's FL(A,B) is less than a given threshold, will try to fight/kill prisoner B.

[Penroyaltea]

In a gang, you probably have a respect level too.

Everyone knows the first thing you must do when entering a prison is knife the first guy you see so people know you mean business.

I don't think gay or straight exists in prison (from documentaries) even the nazi ones can have relationships as long as the other one looks like a girl enough. It's a power and respect thing.

[paktsardines]

For modelling purposes respect can be considered the same as FL(A,B).

[LennyLeak]

I like this idea, but I also think it kinda downplays the structured nature of organized crime. Some street gangs might be nothing more than friendships, but other gangs (especially prisongangs from what I've heard) are pretty top down oriented - and pretty much run like the (illegal) business that they really are. If someone "changes" he would be totally excommunicated and perhaps targeted for assassination. I like your system to handle person to person interactions and relations, but for gang structure I think more is needed. Gangs in prison have shotcallers and members function more like the army than a commune. My point is that for gangs to be believable you need to have rules and orders handed down that dictate the interaction of individual members, more than their personal preferences in regards to others. Prison gangs will often have friends kill friends.

[Great Magical Hat]

It's an interesting model, but I don't think it really belongs in this game.

First off, it completely ignores a lot of things that can make for interesting gameplay, such as things like prisoners being affiliated to one another because of what they did before coming to (this) prison, a gang being present in max sec and min sec and still coordinating things from time to time even though they only have dinner time to communicate and prisoners being affiliated with a group because of things they have no influence over (such as race). Of course, those could be modeled in, but then you come to my second point:

I don't believe this tackles the problem they are having a problem with. I am confident that when the time comes Chris will come up with a good way to define which prisoner belongs to which group. The problem, as I see it, is that you would need something that makes you care about the gangs and in which gang a prisoner is, while it can't break the gameplay as it stands. If it's just a minor thing that can (almost) always just be ignored, the game may well be better off without it. If I get Chris's vision for the game correctly, it also needs to be possible to build a prison that runs without your input, which then is yet another constraint you have to put on the gangs feature, as the game would have to be able to handle gangs without user input.

[paktsardines]

such as things like prisoners being affiliated to one another because of what they did before coming to (this) prison

Criminal history could simply be another trait to be included as part of the trait mapping function, along with things like duration in prison, what sort of crime, race, maxsec/minsec etc, personalities... or whatever. You could even include a prisoner's current mood, so all your pissed off prisoners end up hanging out together and getting up to mischief.

Just to clarify, it was the intention that the model would run without any interaction/input from the player. That said, as you say, it will also likely run without any emotional involvement from the player, unless there's some element(s) introduced to make the player care.

[Great Magical Hat]

Just to clarify, it was the intention that the model would run without any interaction/input from the player. That said, as you say, it will also likely run without any emotional involvement from the player, unless there's some element(s) introduced to make the player care.

I understood that. However, the gang system would consist of (at least) two parts: the part which defines who belongs to what gang and the part which has gangs do something (stage riots together, fight with other gangs and things like that. You only described your solution to the first part, but I don't think that part is the problem. The second part is hard because it really needs to balance the different things I described above. I didn't mean to say this balance was missing in your solution.

[FourWinds]

I could see that 'gangs' would perhaps add something to the game: I'm thinking that factors of discontent would propagate faster through gang units and this would make members less discrete perhaps. That said, beyond that I'm not sure they would bring a lot. Personally I'm more of a fan of older prisons (Alcatraz, Porridge, Shawshank), and I note that they are all pre gang (as most would think of prison gangs) era. Perhaps I'd like to see prison cliques rather than gangs.

The same thoughts come to me when thinking af all the execution types I see discussed, as well as the discussion of how many types of firearms could be squeezed into the game; I'm expecting someone to ask for a BFG, attack helicopter, or a rocket launcher soon.

Also tunnels: who tunnels out of modern prisons?

The whole thought breaks down with CCTV of course, as that's clearly a good idea...or is it?

One of the reasons I'm glad I brought into the alpha is that I get to see how all this evolves.

[Great Magical Hat]

Also tunnels: who tunnels out of modern prisons?

At either Minecon or Eurogamer (or both the two talks were largely the same) Chris talked about tunneling and said that there's not all that much tunneling going on in reality (and he also added that present day prison cells actually tend to be metal framed), however he also said this was one of the instances where they were willing to break from reality for the sake of gameplay. Personally, I think that alongside the gameplay, just the general idea of prisoners tunneling being a really nice one (and one ingrained in popular culture) was part of the whole decision, but he didn't say so.

[asmo0]

At either Minecon or Eurogamer (or both the two talks were largely the same) Chris talked about tunneling and said that there's not all that much tunneling going on in reality (and he also added that present day prison cells actually tend to be metal framed), however he also said this was one of the instances where they were willing to break from reality for the sake of gameplay. Personally, I think that alongside the gameplay, just the general idea of prisoners tunneling being a really nice one (and one ingrained in popular culture) was part of the whole decision, but he didn't say so.

That can't have been the right topic...?