Economic Model II

[Gary Thomas]

Rather than a special ability to find a special, a required tech (like in civ 3) seems better. You cannot find iron if you are still in the stone age, and noone looked for petrol/oil until modern times (unless it was used for naphta?).

That is a very good idea - less micromanagement!

I second Mark's point about aluminium being replaceable. You can always make do with something different from what you expected initially. Even for weapons, I think Aztecs and Incas didn't use metal weapons but stone, and they had quite evolved tactics.

This is not a coding or model issue - it is a data issue. My perception of the whole thing is that scenario designers can add the resources that they feel necessary for the development of the scenario.

One can certainly same the same about gold - it is replaceable. Is this an argument for not having gold among the resources? For some purposes it would be - a scenario set in the 20th century probably would not have gold as a factor. On the other hand, a search for El Dorado probably would.

The only thing I would consider mandatory as a resource is big animals (horses, cattle), which lacked in America. That prevented lots of progress in terms of agriculture and warfare.

Oddly, I don't think that horses (or cattle, etc) are resources as such - once discovered, they are bred everywhere, and the original resource is irrelevant. So we are not talking about an economic resource here, but a precondition for a technology advance. However, they would appear on the map as resources, until the appropriate technology is discovered ("Horse breeding" or whatever) at which point the "rersource" would vanish from the map.

Cheers

[Gary Thomas]

Mark:
I guess I am getting a better grasp of how things work. Can you clarify something for me: what is the relationship between the "the five sectors of the economy" referred to in the model, and "basic goods" referred to in the economics.xml draft? Are they the same thing?

Is it true that the the "resources" sector of the economy is an undifferentiated variable representing some kind of abstraction for all the "normal" (as opposed to "special") kinds of materials? I am not very clear on what is normal and what is special, though I suppose that we do not need to be very specific in the game. I do not, however, want people to be confused.

I am trying to get a coding system that has everything at the square level in a uniform framework. I order To achieve this I want to make a clean easily understood class structure.

Thus far, I have:

1. Basic goods, or sectors (this meaning of sector is not the one I understood when I did economics).

2. Special resources, both as production accelerators and as technology prerequisites.

3. Sites, mapping a basic good to a number in a square.

4. A list of special resources per square.

Cheers

[Mark_Everson]

Hi. I need to rush off to do something so I'll respond to the questions Gary needs resolved. I'll double back later to annoy everyone in more detail

Originally posted by Gary Thomas
I guess I am getting a better grasp of how things work. Can you clarify something for me: what is the relationship between the "the five sectors of the economy" referred to in the model, and "basic goods" referred to in the economics.xml draft? Are they the same thing?

The sectors of the economy each produce an Output that is the basic good in question. I think the five includes specials, so just consider it four for now. So the Farming sector produces the basic good Food. There is a one-to-one mapping between sectors and goods, but they are not the same thing.

Is it true that the the "resources" sector of the economy is an undifferentiated variable representing some kind of abstraction for all the "normal" (as opposed to "special") kinds of materials? I am not very clear on what is normal and what is special, though I suppose that we do not need to be very specific in the game. I do not, however, want people to be confused.

Yeah, you've got it. Normal is Everything that isn't deemed Special by the designer. My short hand way tto decide is that if the special would be available most everywhere then it is not worth treating as a special. In most cases its things like wood, stone, hides, etc. are the "normal" resources. But what is normal and what is special can definitely vary from scenario to scenario. FE if you're playing a scenario where access to ship mast timber is critical and isn't around everywhere, you could make that a special although we usually assume wood is either available or easily replaceable with other materials like straw-mud bricks.

I am trying to get a coding system that has everything at the square level in a uniform framework. I order To achieve this I want to make a clean easily understood class structure.

Thus far, I have:

1. Basic goods, or sectors (this meaning of sector is not the one I understood when I did economics).

Hopefully my expanation above cleared this up. The econ models' sectors are meant to be similar to those you learned in economics. But neither basic goods nor sectors will automatically exist in a map square. The Sites are what you need to keep track of. If the sites exist and the technology is available to use the special Then a sector will be created in the square economy to handle production of that good. This condition is satisfied for Dawn1 specials afaik. It would be good to have iron "appear" to the player as a result of tech advancement in Dawn2 IMO.

2. Special resources, both as production accelerators and as technology prerequisites.

3. Sites, mapping a basic good to a number in a square.

4. A list of special resources per square.

Sounds Good Gary!

Cya,

Mark

[Mark_Everson]

Hi Laurent:

I think your idea about just using tech level requirements to see if goods can be detected is a good one. Two further extensions occured to me that still have low micromanagement.

1. Instead of having a single tech threshold for when the sites are detected there would be a range. If we have different sites have different difficulties for detection (FE one oil site is easy, and another very difficult) then they would become visible to the player as the tech matching the difficulty became available. One example would be when oil is first starting to be widely used the shallowest sites would show up. As drilling/prospecting tech becomes better, others from deeper wells would become available.

2. We could have an Overall prospecting budget, and tie detection into this. This would be low-micromanagement but still give the idea of letting the player accelerate detection if they deem it useful.

Hey Gary:

Originally posted by Gary Thomas
This is why I made my little list of [specials] characteristics (to which I would also add a defense strength).

Defense strength? As in how resistant it is to bombardment or destruction by troops? What exactly did you mean here?

Laurent said:
The only thing I would consider mandatory as a resource is big animals (horses, cattle), which lacked in America. That prevented lots of progress in terms of agriculture and warfare.
Gary said:
Oddly, I don't think that horses (or cattle, etc) are resources as such - once discovered, they are bred everywhere, and the original resource is irrelevant. So we are not talking about an economic resource here, but a precondition for a technology advance. However, they would appear on the map as resources, until the appropriate technology is discovered ("Horse breeding" or whatever) at which point the "rersource" would vanish from the map.

I like Gary's approach here, although I'm not sure about the details. For the time being doing it as stated, and then modifying the approach as issues appear seems useable.

[Gary Thomas]

Defense strength? As in how resistant it is to bombardment or destruction by troops? What exactly did you mean here?

I mean resistance to sabotage or destruction for other purposes such as scorched earth.

I have got to the point where I hope I am closing in on the code requirements. Really the only remaining problem I have is that I do not know how the economic model tells the difference between basic goods and specials. Since they behave differently I am inclined to have them in different lists.

My personal preference for site building is to do it through the existing buildable interface, or at least design the code so that when things are refactored it becomes easy to incorporate the buildable system. Everything that is done through that kind of system (Buildable, Order, SpecialAbility, Command, Administration) becomes very much easier to control, maintain and extend.

My present, updated, list of resources is:

1. Basic goods (four of them).

2. Special resources, both as production accelerators and as technology prerequisites.

3. Sites, mapping a basic good to a number in a square.

4. A list of special resources per square.

5. A Feature class, a generic class which indicates that a special feature is present. This has no direct impact on economics, but is used for the horses, cattle and so forth. At a later stage it will have the framework present for disasters, disease and so forth. Generally it will hold ephemeral phenomena.

Cheers

[Mark_Everson]

Hi Gary:

Everything you say sounds good to me, though I haven't had a chance to reflect upon it.

Here is how you tell a basic good from a special:

public class GoodsInfo{
    /** Holds the actual Goods of the economy, keys are the Good names  */
    private static HashMap theGoods = new HashMap();

    private String goodName;
    //  for a special good like gold, it is eventually converted into services
    private String basicGoodItBecomes;      
    //  when converting a special into good it becomes, how many?
    private float goodConversionRate;      
    private boolean basicGood = false;
    // Eventually need technology required to make good...

(SNIP)

 GoodsInfo( String _goodName, String _basicGoodItBecomes, float _goodConversionRate){

          goodName = _goodName;
          basicGoodItBecomes = _basicGoodItBecomes;
          goodConversionRate = _goodConversionRate;
          if( goodName.equals(basicGoodItBecomes) ){
              basicGood = true;
          }
          else basicGood = false;

The boolean basicGood tells whether its a basicGood or not. It for now is set in the constructor shown by comparing names as you can see. For now you can have a lock on the econ code to change this around if you want to add a more robust way to handle basic vs special goods. But I would prefer to leave basic and special goods together in the same GoodsInfo list for now since I am not sure about what changes would be needed to separate them. I will send you by email the draft econ xml ini stuff I worked up a while ago in case it is of use to you.

[Mark_Everson]

Sites per terrain type

Well, now that we have choices of where to settle (coming in D7) we need to have a reasonable approach for what type of sites (farm, resources, and specials) are associated with each type of terrain. This ties in with what I want to do this weekend in rationalizing the economic framework to deal with different numbers of sites on different types of terrain.

The Specials can possibly be put in by hand per scenario, but I think its time to specify that the current "flat" terrain has so many of each type of site, and forest has a different set of values. For now the values could be constant, although eventually they'd probably be picked from some sort of distribution.

So Gary, how would you prefer I do this? Or do you want me to just give my best guesses on the correct number of sites of each type for each existing terrain, and you implement it? Your call

Once this is done we will have a sensible driver to make population diffusion work.

[Gary Thomas]

Tell me what you want and I will implement it. All the infrastructure is there.

Cheers

[Mark_Everson]

Sounds good Gary!

Here is my first shot at it, comments welcome. Note these are values for terrain that hasn't been modified. FE no drainage in swamp. We will get to terrain improvements some other time...

BTW, can we add two flavors of flat similar to the distinction between grassland and plains in civ? For now I'll give an intermediate value, but I expect we will soon want to make a distinction. There is no need for them to be integers, so if people think 1 for dunes is too high we can make it 0.3 or something. The quote below gives a description of the general scale of these numbers.

Terrain type, Farm sites, resource sites

flat____10, 2
rolling_5, 4
broken__2, 6
mountain_1, 8
ocean___0, 0
coastal_12, 4
swamp___2, 2
forest__3, 5
jungle__4, 2
dunes___1, 1
tundra__1, 2

River would also give a bonus, but we're not there yet. Services sites are needed by the code, but irrelevant in the model. I have been setting those sites to 1.

One issue: These values, and the lack of merchants moving food, will make square populations above 15k or so almost impossible with low technology. (Current code has farm sites = 90, res = 3 for all squares) This could cause big problems with the carthage scenario in that the populations of the cities would not be sustainable. We can temporarily fudge the numbers up for farm sites, or come up with a temporary square type that has especially good farmland as a fix. Ideas?

Mostly for my own reference I'm going to copy here something I wrote in the Vegetation, Climate, Ecology, and Pollution thread.

Some numbers for sites...
I think I've already given numbers for this in the discussions above. For the record (and this may change with playtesting) a lousy square (desert, tundra...) might have one agricultural site, steppe might have a few, rainfall agriculture-type squares would probably range from 5 to 20, and vastly productive agricultural lands like ancient Egypt might have something like 50 sites per square along the Nile. Irrigated desert agriculture poses a bit of a problem, since unless the irrigation is there, the lands almost useless. So I think we might need a special case to handle such lands. But for demo 5 I think we can't ignore the distinction and see how the system works in general.

The rule of thumb is there should be as many sites as ancient agrictulture could support population over the area of the square. So if on average a square in China supported 30K people, then there should be 30 sites. The exception is lands where heavy investments in irrigation are required. There I would cut the pop number by 2/3 or so since the investments in irrigation will increase the yield above the usual amount just given the sites.

Hope this is what you need... For resources I would just wing it for now. Sites will usually be in the range from 1-10.

[Gary Thomas]

This is now implemented, except that I added a plains terrain and changed the parameters to:

flat 12, 2
plains 8, 3

Tell me what you want for these and I will fix them.

Cheers

[Mark_Everson]

Thanks Gary, those values look fine to me for now. When we see how they play out, we can think about modifying them.

[Mark_Everson]

I posted this by accident in the Demo 7&8 thread, so I'm going to re-post here where it really belongs:

I'm now working on getting colonists to distribute themselves effectively into the different economic sectors (farming, etc.). For now there is no influence upon this distribution of whether the economy is market or traditional. My question (for the longer term) is should colonists that are primarily traditional keep their fixed patterns of distribution into the sectors when colonizing? I don't know if its realistic, but I'm sure it will piss off players to have settlements start out hobbled because there are an inappropriate number of say farmers compared to other sectors given the sites available. My notion of how to handle this for now is that settlers with a traditional economy will distribute themselves near-optimally but after that the traditions are set for that locality, and the proportions of workers in the sectors stay fixed (in a pure traditional economy, which will be rare). My hand-waving argument is that the colonists knew that traditional patterns could be shaken up when they moved.

In the other thread Laurent replied:

Mark, I think your proposal is OK. I think the traditions hold only when there is some amount of infrastructure. Those who migrate are usually ready to start from scratch and seize whatever opportunity is open to them.

[Simon Loverix]

Hello again

I think those traditional colonists will try to keep their traditional occupations, but after that they will react normally (= according to their culture) to changes. The effect will be that a part of them strengthens their traditionally strong sectors and another part adapts to the new situation. Think, for example, of the Indian immigrants in the UK where the caste system continues to play a certain role, though there is no longer an economic or juridical base for it.

Concerning the terrain values, is there anything I can do? I did something on the food and production sites related to the ecology, should I elaborate or revise it?

ciao

[Mark_Everson]

Hi Simon!

The problem with that is that if you take the proportions in labor in different sectors from a given traditional economy and plop them down in terrain different from that in which their economy evolved, they may starve or be largely worthless. Anyway, I went ahead and did it the "versatile settlers" way for now. We can always change/refine it later...

On the sites, we just ended up hard-coding site number by terrain type, based partly on the numbers you and I worked on a bit ago. I think we are set for the moment, but thanks for offering! If you've got the time, there has been a lot of discussion on various things since the last time you were actively here, those topics could all use some fresh eyes .

Cya,

Mark

[Simon Loverix]

I just thought they might try their traditional sectors first, and change quickly afterwards if they can't make a living. Migrating people are likely not very rejective to change, except when conservatism is their reason to migrate.