Author Topic: Settlement "snakes"  (Read 1171 times)

De-Legro

  • M&F Dev Team
  • Sr. Member
  • *****
  • Posts: 3109
  • Karma: +105/-54
    • View Profile
Re: Settlement "snakes"
« Reply #30 on: April 23, 2017, 12:57:43 PM »
If it's per character then the characters will have the blobs, not the realms.

Also, instead of linear you could use something like the popular sigmoid style functions or tanh, where under a certain threshold, it's negligible, then it ramps up high after some point. You can customize those functions for where they start spiking up and how quickly it hits the peak.

That way, you can make it so that distance corruption doesn't kick in to some specified threshold, then ramps up nearly linearly to some cap point, then flattens out again, and all those values can be set as parameters of a single function.


I thought I posted about this before. The major change to the corruption system would be to get rid of the linear nature. My current set up allows for a few settlements per character with no corruption and ramps up from there.


I'm not convinced about distance corruption for the reason listed before, but also because it causes wars for mechanical reasons. Hey tiny realm to our east whom we have always been friendly with. Sad story we have expanded west as far as it practical, so guess we are going to have to take your land. No hard feelings though right.
He who was once known as Blackfyre

Demivar

  • Full Member
  • ***
  • Posts: 134
  • Karma: +1/-0
  • Purveyor of cunning plans.
    • View Profile
Re: Settlement "snakes"
« Reply #31 on: April 23, 2017, 01:00:50 PM »
It is worth considering that the costs of corruption are already exponential. If we assume there's a 1-player realm, every new estate that you take adds a flat amount of % corruption which affects all other estates.


With 1 1k pop settlement you pay 2 food, for 2 1k pop settlements you pay 4 each, hence 8.
22:34 - Roran Hawkins: Radovid's like you
22:34 - Roran Hawkins: but then insane
22:34 - Roran Hawkins: Dijkstra is like you

Andre

  • Full Member
  • ***
  • Posts: 365
  • Karma: +2/-2
  • Constant boredom
    • View Profile
Re: Settlement "snakes"
« Reply #32 on: April 23, 2017, 01:09:57 PM »

I thought I posted about this before. The major change to the corruption system would be to get rid of the linear nature. My current set up allows for a few settlements per character with no corruption and ramps up from there.


I'm not convinced about distance corruption for the reason listed before, but also because it causes wars for mechanical reasons. Hey tiny realm to our east whom we have always been friendly with. Sad story we have expanded west as far as it practical, so guess we are going to have to take your land. No hard feelings though right.

What do you think about my ideas though? Not specifically for corruption but for keeping realms in check in general, though you could probably incorporate corruption aswell. Afterall, people who lose stuff because of corruption wont be as happy.

Cipheron

  • Full Member
  • ***
  • Posts: 186
  • Karma: +7/-4
    • View Profile
Re: Settlement "snakes"
« Reply #33 on: April 23, 2017, 04:31:26 PM »

I'm not convinced about distance corruption for the reason listed before, but also because it causes wars for mechanical reasons. Hey tiny realm to our east whom we have always been friendly with. Sad story we have expanded west as far as it practical, so guess we are going to have to take your land. No hard feelings though right.

That wouldn't work with the setup I suggested, because remember I suggested a "closeness" metric for one knight's towns as a whole, rather than a distance from a set point. e.g. if you were required to defined a nominated "seat of governance" and then calculate corruption based on that, you would indeed get the issue of only wanted to spread out in a circle. but that's not what I proposed, my suggestion would work differently. The center shifts based on the geographic heart of your towns.

If you invaded to the east, because invading to the west was too-high corruption, then the "gravitational center" would shift east, meaning you now pay more for corruption as a result of your existing western towns now being further from the "center of gravity".

A fast way to calculate this could in fact be a population-weighted average location, then take the average distance from each city to that point, also weighted by population. So if you take a few low-pop towns it won't matter too much, but if you start capturing mega-cities then the "center" will shift towards them, increasing the corruption rating for your existing towns. The result would be that if you wanted to capture a high-population center you would think twice about keeping it in the same polity as another high-population center you already own, you'd want to make them separate zones to keep the corruption reasonable.

And there's actually a paper on county size, and why counties created at different times are different sizes. It turns out that the deciding factor was transport technology at the time the county was created. Counties tend to be the size that minimizes the travel time for authorities to reach person X, where person X is any typical person in the county. The optimal county size varies based on population density. For more dense counties, the optimal size is smaller, for less dense counties, the optimal size is larger. And historical land divisions do actually adhere to the formula, and in a fractal sense, e.g. up to nation sizes.

So yeah, efficient region structures do actually play a big part in history, because the regions that were closest to the optimal shape outperformed those which were not. It might be mechanical, but that's not necessarily unrealistic.
« Last Edit: April 23, 2017, 04:55:26 PM by Cipheron »

Demivar

  • Full Member
  • ***
  • Posts: 134
  • Karma: +1/-0
  • Purveyor of cunning plans.
    • View Profile
Re: Settlement "snakes"
« Reply #34 on: April 23, 2017, 04:53:08 PM »
That wouldn't work with the setup I suggested, because remember I suggested a "closeness" metric for one knight's towns as a whole, rather than a distance from a set point. e.g. if you were required to defined a nominated "seat of governance" and then calculate corruption based on that, you would indeed get the issue of only wanted to spread out in a circle. but that's not what I proposed, my suggestion would work differently. The center shifts based on the geographic heart of your towns.

If you invaded to the east, because invading to the west was too-high corruption, then the "gravitational center" would shift east, meaning you now pay more for corruption as a result of your existing western towns now being further from the "center of gravity".

A fast way to calculate this could in fact be a population-weighted average location, then take the average distance from each city to that point, also weighted by population. So if you take a few low-pop towns it won't matter too much, but if you start capturing mega-cities then the "center" will shift towards them, increasing the corruption rating for your existing towns. The result would be that if you wanted to capture a high-population center you would think twice about keeping it in the same polity as another high-population center you already own, you'd want to make them separate zones to keep the corruption reasonable.
I believe that there was a lot of talk and good ideas for the implementation of seats of power, however. Generally speaking, administration and infrastructure sprawls out from courts as centres of power.


I'm actually not a fan of hard coding corruption to be more penalising as it already bites hard enough. It's even worse in regions where land is less valuable, not only are 'bad' settlements lacking in resources, they're already hard to manage as the provinces are a lot larger anyway. Distance based corruption would add to those issues even more.


« Last Edit: April 24, 2017, 12:14:05 AM by Demivar »
22:34 - Roran Hawkins: Radovid's like you
22:34 - Roran Hawkins: but then insane
22:34 - Roran Hawkins: Dijkstra is like you

Cipheron

  • Full Member
  • ***
  • Posts: 186
  • Karma: +7/-4
    • View Profile
Re: Settlement "snakes"
« Reply #35 on: April 23, 2017, 04:58:01 PM »
You could scale per-settlement corruption based on a function (e.g. log function) of (population), to ensure it's not a big hit on a small town. The most useful one is the sigmoid type functions. Let me explain that:



Basicall you'd take the linear amount of corruption and multiply it by the sigmoid function over population, and you'd get a situation where very small cities effective get no corruption, then the multiplier increase to x1 at some arbitrary high level of population and you get the regular linear corruption from that point onwards. You can scale the -6 and +6 points to be equal to any start and end population levels you like, so the sigmoid function is handy for this sort of thing.

Also, corruption wouldn't have to be greater in a distance-based system. The average corruption in the game could be exactly the same, but it's scaled by distance from the court and a function of population size.
« Last Edit: April 23, 2017, 08:23:30 PM by Cipheron »

Demivar

  • Full Member
  • ***
  • Posts: 134
  • Karma: +1/-0
  • Purveyor of cunning plans.
    • View Profile
Re: Settlement "snakes"
« Reply #36 on: April 23, 2017, 05:09:42 PM »
You could scale per-settlement corruption as log(population) or something however, something to ensure it's not a big hit on a small town.

Also, corruption wouldn't have to be greater in a distance-based system. The average corruption could be exactly the same, but it's scaled by distance from the court and population size (or some log value of pop size). The average amount of corruption in the game could be the same or even lower than now.
Still can't quite agree with that. I'd prefer not to tie population with corruption, else a realm with loads of players and a lots of tax would burden down the liege (I know you suggested a log, but the principle still stands). It'd also be even easier to take swathes of the map, as the population part of the corruption would be negligible for impoverished land.


There's a confusion in a lot of people's minds between corruption and how centralisation should work. Corruption is tied to the player's account, and I don't think it should be tampered with too much. (De-Legro's planned change should fix an exploit with this, but particularly given the state of the game, I'd only bother with fixing exploits rather than any other changes).


As I said, Seats of Power was a thing talked about a lot, and many good things can come from it that would be very healthy for the game. It would promote the things that are good for M&F and is a no-brainer in my mind, really.

Edit: I posted before you included the log graph. Not that it matters, but that's a good thing to include for a lot of people. I've got a personal vendetta against pseudo intellectuals that like to reference parts of Maths that a lot of the world's population doesn't know about, and it kills discussions off. I get it, but it doesn't mean that others do too.
« Last Edit: April 23, 2017, 05:13:31 PM by Demivar »
22:34 - Roran Hawkins: Radovid's like you
22:34 - Roran Hawkins: but then insane
22:34 - Roran Hawkins: Dijkstra is like you

Gustav Kuriga

  • Full Member
  • ***
  • Posts: 459
  • Karma: +22/-34
    • View Profile
Re: Settlement "snakes"
« Reply #37 on: April 23, 2017, 07:48:59 PM »
Still can't quite agree with that. I'd prefer not to tie population with corruption, else a realm with loads of players and a lots of tax would burden down the liege (I know you suggested a log, but the principle still stands). It'd also be even easier to take swathes of the map, as the population part of the corruption would be negligible for impoverished land.


There's a confusion in a lot of people's minds between corruption and how centralisation should work. Corruption is tied to the player's account, and I don't think it should be tampered with too much. (De-Legro's planned change should fix an exploit with this, but particularly given the state of the game, I'd only bother with fixing exploits rather than any other changes).


As I said, Seats of Power was a thing talked about a lot, and many good things can come from it that would be very healthy for the game. It would promote the things that are good for M&F and is a no-brainer in my mind, really.

Edit: I posted before you included the log graph. Not that it matters, but that's a good thing to include for a lot of people. I've got a personal vendetta against pseudo intellectuals that like to reference parts of Maths that a lot of the world's population doesn't know about, and it kills discussions off. I get it, but it doesn't mean that others do too.

Why do you have a personal vendetta, and why do you call them pseudo-intellectuals?

Cipheron

  • Full Member
  • ***
  • Posts: 186
  • Karma: +7/-4
    • View Profile
Re: Settlement "snakes"
« Reply #38 on: April 23, 2017, 07:59:18 PM »
I think Demivar meant people who try to cite things just to sound clever, then don't try and explain it, it can come off as a type of elitism.

He was just saying that the visual helps because showing is better than telling. You just have to look at it and you get the feeling for how it would work, no need for the equations etc.

Quote
It'd also be even easier to take swathes of the map, as the population part of the corruption would be negligible for impoverished land.

That's sort of already the case. If you get 10 500-people towns then the total corruption costs (in "hard dollar terms") are 50% of having 10 1000-people towns.

One way that could benefit the current system would be to have "base corruption rate" which is just the normal 0.2% per settlement, and then each settlement gets an "effective corruption rate". The idea would be that it looks at the total account-wide resource burden from corruption, then skews it so that less of the cost is coming from your small towns. A town 50% smaller than average is currently only costing you 50% of the "average" towns corruption, that could be skewed to 25%, and bigger than average towns pay a bit extra to make up for it (but it wouldn't be that much more, proportionately). This could be calculated pretty fast just by knowing your total pop in your account, each town would get a +/- to corruption based on what share of population it represents and the total number of towns.
« Last Edit: April 23, 2017, 08:21:49 PM by Cipheron »

Demivar

  • Full Member
  • ***
  • Posts: 134
  • Karma: +1/-0
  • Purveyor of cunning plans.
    • View Profile
Re: Settlement "snakes"
« Reply #39 on: April 23, 2017, 08:09:43 PM »
I was going to respond to that, but Cipheron pretty much covered it.


If you look at the average person's education, a lot of people won't know what a log is/what a log graph looks like or even remember. If you want to reference those kinds of things, it's always best to make sure everyone knows what you're talking about so that people aren't excluded from the discussion.
22:34 - Roran Hawkins: Radovid's like you
22:34 - Roran Hawkins: but then insane
22:34 - Roran Hawkins: Dijkstra is like you

Gustav Kuriga

  • Full Member
  • ***
  • Posts: 459
  • Karma: +22/-34
    • View Profile
Re: Settlement "snakes"
« Reply #40 on: April 23, 2017, 08:15:21 PM »
I was going to respond to that, but Cipheron pretty much covered it.


If you look at the average person's education, a lot of people won't know what a log is/what a log graph looks like or even remember. If you want to reference those kinds of things, it's always best to make sure everyone knows what you're talking about so that people aren't excluded from the discussion.

Ah ok, no problem then.

De-Legro

  • M&F Dev Team
  • Sr. Member
  • *****
  • Posts: 3109
  • Karma: +105/-54
    • View Profile
Re: Settlement "snakes"
« Reply #41 on: April 24, 2017, 02:27:39 AM »
Account and Character corruption is something can be be modified/added quite easily to the current code base and the effects are also reasonably easy to predict. The same is not true of most forms of distanced based corruption. That is not to say it might not be added in the future and may or may not entirely replace the current system, but it is certainly not a priority for me.

If such a system was to be added it would need to address the problems BM always had with the system for example
  • Realms that never go to war because they simply can't control any more estates further from their capital
  • Realms that have "colony" or "Ally" realms that are for all intents and purposes the same realm and exist only to bypass control issues.

In my head corruption is generally a function of administration. Administration difficulty is a function of distance, amount and some other factors not yet discussed (cultural differences, independence/rebel groups, corrupt officials etc). Thus we should be able to mitigate some effects by things like de-centralisation,  which in turn I guess bring there own penalties or risks.
I'm not a fan of a "geographical" center. Firstly it is not apparent nor easy to predict for the average player. Secondly is doesn't to me make logical sense. If corruption is the effect of distance from a bureaucratic central location, why would it change unless that central location is changed?
He who was once known as Blackfyre

Andre

  • Full Member
  • ***
  • Posts: 365
  • Karma: +2/-2
  • Constant boredom
    • View Profile
Re: Settlement "snakes"
« Reply #42 on: April 24, 2017, 09:10:50 AM »
Feels bad when your idea for how corruption or control issues could be done are completly ignored. :(

Andrew

  • Game Master / Lead Developer
  • Administrator
  • Sr. Member
  • *****
  • Posts: 1797
  • Karma: +75/-7
  • Mildly Amused
    • View Profile
    • Lemuria Community Fan Site
Re: Settlement "snakes"
« Reply #43 on: April 24, 2017, 10:52:27 AM »
What about some form of "issues" game? As in settlements will have issues such as monsters from ruins and dungeons and such coming out (good reason to explore them and try to get rid of them), (NPC) bandits, too high taxes (outgoing trade, gold tax sent to lord/realm ruler if that is ever implemented), armies marching trough (whether friendly or not doesn't matter), lack of supplies (any resource), lack of militia, recently conquered (smaller penalty if from a slumbered lord), possibly even the distance penalty aswell.

Ofcourse one of these being slightly in the negative wouldnt do anything, you would probably have to have a lot of problems for it to remove the lord and realm allegiance.

I am tempted to have certain things appear and maybe have them interact with settlement stats, but only if I can also make those things, themselves, a fun occurence. Otherwise, any sort of little issue is basically something the character can outright ignore. Things like too high taxes are difficult to define (is it a tax or a trade or a tribute?), and lack of supply is something that already affects people--if anything I'd make that effect harder than it is now, but not as part of this concept.

I think we could also add something like trade variety and goods bonuses, as in, if the settlement has a lot of trade from distant places coming in, then they are happier. And maybe a thing similar to building focus, where you set how many goods you use, whether that be 100%, 150%, 200% or 300%. And the higher it is the higher the bonus is, with 100% being 0 bonus.

Until we start tracking the origins of goods, which is impossible without rewriting the trade system, this is impossible. Though I would like to reward diverse settlements down the line.

We could also do some kind of fear thing, where you can for example set troops to activly disuade people from trying to secede, but this would also reduce opinion, and so they are that much more likely to secede some time after the troops leave. And troops present that aren't disuading would also slightly reduce chance of secession when they are present, but wouldn't leave an opinion effect other than the one for troops marching trough. (Which I think should depend on how often troops march trough, one warband marching trough every few weeks likely wont do anything and might even make the mortals excited for seeing an army.)

Similarly, troops from a foreign realm being present in a dissatisfied town would make it likely that they secede to them, or just speed up TO's, as long as the troops aren't causing trouble. 

Like Tom, I'm not going to limit you from doing something you can actually decide to just do.

Additionaly, we might be able to both increase fear and excitment with executions of First Ones. Fear as in they fear someone who could just execute another First One just like that. And excitment because it isn't often you see a First One die, especially fun for the citizens if they are allowed to humiliate the First One maybe, a new execution or punishment? Humiliation? And maybe more effect if that First One is one who was recently causing trouble there, or from a realm from which someone was recently causing trouble.

We might also be able to construct a colloseum or similar thing, where we let thralls fight, and possibly also trained soldiers. This would increase happiness and also generate gold income I guess. Slave armies anyone?

Hm.

That wouldn't work with the setup I suggested, because remember I suggested a "closeness" metric for one knight's towns as a whole, rather than a distance from a set point. e.g. if you were required to defined a nominated "seat of governance" and then calculate corruption based on that, you would indeed get the issue of only wanted to spread out in a circle. but that's not what I proposed, my suggestion would work differently. The center shifts based on the geographic heart of your towns.

If you invaded to the east, because invading to the west was too-high corruption, then the "gravitational center" would shift east, meaning you now pay more for corruption as a result of your existing western towns now being further from the "center of gravity".

A fast way to calculate this could in fact be a population-weighted average location, then take the average distance from each city to that point, also weighted by population. So if you take a few low-pop towns it won't matter too much, but if you start capturing mega-cities then the "center" will shift towards them, increasing the corruption rating for your existing towns. The result would be that if you wanted to capture a high-population center you would think twice about keeping it in the same polity as another high-population center you already own, you'd want to make them separate zones to keep the corruption reasonable.

And there's actually a paper on county size, and why counties created at different times are different sizes. It turns out that the deciding factor was transport technology at the time the county was created. Counties tend to be the size that minimizes the travel time for authorities to reach person X, where person X is any typical person in the county. The optimal county size varies based on population density. For more dense counties, the optimal size is smaller, for less dense counties, the optimal size is larger. And historical land divisions do actually adhere to the formula, and in a fractal sense, e.g. up to nation sizes.

So yeah, efficient region structures do actually play a big part in history, because the regions that were closest to the optimal shape outperformed those which were not. It might be mechanical, but that's not necessarily unrealistic.

I like the logic here, but it kills islands, and doesn't take into account that we have instant communication as part of the game lore.

For what it's worth, I don't mind the current system. It's easy to figure out, and it's highly predictable. What I don't like is when I do database queries to find that one guy controls a relatively HUGE portion of the map because all 50 character slots all have like 3 estates on them each. Mind you, I've not done a query like that in a while, but I do sit down to note who does or does not control ridiculous amounts of land.

I'm not against someone having a bunch of land, or a character that controls 15 estates, but that's way more than one character can be reasonably expected to actually manage.
Standing for the creation of interesting things since Year 1, Week 5, Day 4.
Favorite cold beverage: Strawberry Shake
My hobbies: Fixing computers, video games, anime, manga, some other stuff, sleep (in no particular order)