Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[GavJ]

That is a deceptive argument because you have just offered me a higher affluence item than the regular chair.  If I already had a house full of such chairs then I would actually be stupid to accept the offer at any price.

This is the heart of the problem. You've latched onto this affluence thing and you're treating it as all-important. Whereas in reality, it is only one of dozens of relevant influences, and it has no importance at all in any situation where personal usage and appreciation is not the largest value influence (which is in fact most of the time)

Here's a couple more graphs to point this out:

You don't just use affluence, you use whatever one motivation has the highest value to you at any time. And since resale value completely ignores how affluent you are (buyers don't give a damn if you roll around in gold coins or live in a moldy wooden shack), it persists where your personal use doesn't, so it takes over, and you just sell the chair if you're too rich to need it for itself.



Notice how the value never hits zero. Also notice how it is beginning to take on the shape of a curve. This is not a coincidence. This is one of the basic concepts of where a demand curve comes form (although this itself is not a traditional demand curve)

[GavJ]

I am thinking of starting a new thread focused solely on economic ideas that can be done in the next set of major feature updates.  Small changes and small feature additions that will be playable and interesting, while laying the ground work for further improvements

Thread is long since cluttered. I was planning on just slogging along anyway and maybe testing things here and eventually maybe reposting a summary if one happens.

As for piecemeal additions, I dunno... doesn't hurt to suggest, but it doesn't seem to be the way Toady does stuff, and unlike world activation stuff, I really don't think an economy is too complicated to do in one update at all. The walls of text look complicated, but the coding itself just shouldn't be. I mean, not for the scale of an update by any means.

I am actually refining my ideas for a proposed system for DF by arguing, too. Not just pointless bickering. Having to teach/defend something is a great way to hone your own knowledge. I'm just trying to stick to explanations of economics and avoid the squishier stuff, is all

[Scruiser]

I am actually refining my ideas for a proposed system for DF by arguing, too. Not just pointless bickering. Having to teach/defend something is a great way to hone your own knowledge. I'm just trying to stick to explanations of economics and avoid the squishier stuff, is all

In that case I will help you with that instead of posturing and abandoning the thread  .  Just focus on criticisms relevant to the algorithms themselves, and ignore stuff that criticizes economics as a field of study.  If you get your idea together complete and developed, start a new thread focused just on the final algorithm idea.

Also, by small ideas, I meant stuff like the trade route pathing that you mentioned, or the reaction reagent-product tracking in world-gen like I mentioned, that could be done and would be good and complete in themselves, while setting things up for future improvements.

[GavJ]

The question of order of trade is relevant to the algorithm. Buncha industries, which ones get decided first? I.e. the stuff we've just been talking about (some of it).

I'm still not 100% sure how it should work. You can't just prioritize need for survival and then have everything else random, because everything MIGHT help your survival. A few extra food stocks? A bit more wealth to trade for food later in times of need? Another axe, if invasion is going to be the next biggest survival threat? Nothing is certain.

So really, somehow, something needs to rank or prioritize stuff, but it will change for every settlement for every situation. I'm sure there's some not-too complicated method, but not sure what it is yet. Pure ranking by item type doesn't work. Needs fluctuate by item type.

Maybe something sort of like the Sims? Where they have little needs bars for food and sleep and thirst and bathroom, etc.? Just keep track of a handful of survival relevant needs and dynamically rank based on that or something?

It could get folded into the demand curve ahead of time, or it could be a semi-complicated ranking like this afterward, but at some point its a question needs answering.

[GoblinCookie]

/start rant
Pretty much the conclusion I reached a few pages back.  GoblinCookie, economist don't have everything right, but it is not the "argument from authority" to say that the field of study that has worked on issues for over a hundred years might have some accurate things to say about those issues.  At the very least, you will need to give more justification for why what you have thought up in your spare time is so much more useful, accurate and true than what many other people have spent a long time working on.  In the future, could you just ignore my posts based around economics, and I'll ignore your posts that go against economics and hopefully we can avoid cluttering threads with long arguments that will go nowhere?  I really don't see that as productive for generating good suggestions.
/end rant

Well it is the Appeal to Authority fallacy to dismiss an idea solely because eminant economists X, Y and Z say otherwise. 

The problem here is really not that the economics is necceserily wrong in the real-world (though it is), it is wrong in dwarf fortress world.  That is because as mentioned economics is based upon politics (itself based upon a lot of other things!), if the politics is different then the economics are different. 

Initially ignoring economic theories can actually be to your advantage in a way because it allows you to work from the ground up so to speak based solely on the mechanics of the game and then you will arrive at the correct economic theories (that is not demand curve) as a result.

This is the heart of the problem. You've latched onto this affluence thing and you're treating it as all-important. Whereas in reality, it is only one of dozens of relevant influences, and it has no importance at all in any situation where personal usage and appreciation is not the largest value influence (which is in fact most of the time)

You do not seem to understand what I actually mean by affluence.  It is simply a representation of how demand of populations works, how the economy does not grind to a halt just because everyone got fed because and that was all they initially wanted.

You don't just use affluence, you use whatever one motivation has the highest value to you at any time. And since resale value completely ignores how affluent you are (buyers don't give a damn if you roll around in gold coins or live in a moldy wooden shack), it persists where your personal use doesn't, so it takes over, and you just sell the chair if you're too rich to need it for itself.

Here's a couple more graphs to point this out:

I do not know why you think writing it all in a graph will make me agree with you more, I understand the theory but I also understand exactly why it is wrong.  Yes affluence is irrelavant to buyers, but not for sellers.  It determines whether you will sell the chair to get bread or whether you will place it around your dining table.  If you stop getting bread you start to sell your chairs when previously you would have used them.

I have already represented the above graph mechanically through the concept of treating the demands of others you are presently trading with as secondary demands (to consumer demands).  If the chairs are in demand by a third party then the AI would buy them up to sell to the third party. 

Notice how the value never hits zero. Also notice how it is beginning to take on the shape of a curve. This is not a coincidence. This is one of the basic concepts of where a demand curve comes form (although this itself is not a traditional demand curve)

.......

That is where the theory goes wrong.  The value suddenly reaches 0 when you run out of customers for your chairs that you can affordably reached or can reach you. This is unlikely to happen for an individual in the real-world, (part of the sustainability of demand curve illusions) however in the dwarf fortress world it is a very grave concern because of the scale your are working on.

Or as I put it the demand cliff.  As soon as you have sold so many chairs that nobody else wants to buy any more chairs the remaining chairs suddenly become worthless and the wood they are made of becomes more valuable than the chairs themselves are. 

The question of order of trade is relevant to the algorithm. Buncha industries, which ones get decided first? I.e. the stuff we've just been talking about (some of it).

I'm still not 100% sure how it should work. You can't just prioritize need for survival and then have everything else random, because everything MIGHT help your survival. A few extra food stocks? A bit more wealth to trade for food later in times of need? Another axe, if invasion is going to be the next biggest survival threat? Nothing is certain.

The way it can work is that we have a stockpile value which in my opinion should just based on the fear value (what makes people arm their people for war). 

The more fearful a settlement is, the more it tries to stockpile items that it does not presently need for future need.

So really, somehow, something needs to rank or prioritize stuff, but it will change for every settlement for every situation. I'm sure there's some not-too complicated method, but not sure what it is yet. Pure ranking by item type doesn't work. Needs fluctuate by item type.

Maybe something sort of like the Sims? Where they have little needs bars for food and sleep and thirst and bathroom, etc.? Just keep track of a handful of survival relevant needs and dynamically rank based on that or something?

It could get folded into the demand curve ahead of time, or it could be a semi-complicated ranking like this afterward, but at some point its a question needs answering.

You want the game to play a game of sims for every single AI critter in the whole world?  Somehow that is going to take a mammoth amount of computer power.

The simple solution you are looking for is to use my system and then randomise the priorities somewhat.  Different settlements and civilization randomly have different items at different affluence levels.  And then if you like you can randomly have these recalculated ever so often. 

[GavJ]

If the chairs are in demand by a third party then the AI would buy them up to sell to the third party. 

The value suddenly reaches 0 when you run out of customers for your chairs that you can affordably reached or can reach you.

What, do the dwarves tell the traders "Oh hold on for 6 months while we go out and do market research to find out if there are buyers, and exactly where they are across the world"? No, they can't know this in time. The offer is unexpectedly thrust upon you by the caravan coming with a shit ton of chairs. You need to take risks. You have to buy things you EXPECT to be in demand but might be totally wrong about, and you might have to do so as a snap judgment.

In fact, in DF more than in modern life by far, this is what you have to do, because dwarves don't have telephones.

"Market research" is actually something I'm considering for algorithms, but only as a nice bonus now and then -- paying X amount of resources to finance a couple of members of your civ and some mounts to ride out and figure out trade situations around you (as a player or NPC) to better inform choices. However, this costs money (read: horses and labor time, etc. not literally money  ) and would only be done if info is expected to be out of date and profits likely being missed, and something that might only happen once a season or two tops.

Blindness is still the order of the day, for the most part, and the underlying base system needs to operate profitably in blindness, which means risks and guesses.

You want the game to play a game of sims for every single AI critter in the whole world? 

No no, per settlement. Not per critter. Although either would still be trivial computation, since you only need to calc them once per season (maximum frequency caravan schedule. (We're talking on the order of one or two arithmetic problems per tick even for critters. One or two per real life second for settlements)

[GavJ]

Sorry to double post, but only a few awkward minutes apart:

ORDER OF TRADE (for NPC-NPC, even though I use terms like "you")
EDITED entirely 8/5/14 2:18am central time, because my previous post was crap. This is an infinitely better algorithm:

1) First, go through both lists of needs from each trade party, cancel out any needs that match across both parties. I.e. if we both demand food, remove food from both lists.
2) Now just choose one party, doesn’t matter which, and go down their list of things to sell. Match up each one to the other party’s demand curves for those things to the seller's original supply curves, and decide on prices and quantities from equilibrium points. Keep track of each, as well as the economic surpluses and total prices per trade. You'll have a long list at the end.
(note: you can't actually change the amount you have with you in your caravan - too late now, but original supply curve is still important to make sure you aren't throwing away value. You would have used it earlier on to guess the right amount to send in the caravan, but now you're double checking your guess against reality. If the ideal quantity is above what you have on you, just sell it all. If the ideal quantity is below what you have with you, then you withhold the remainder and bring it back home with you. Hopefully you guessed spot on, and such inefficiencies will be minimal)
3) Sort that list FIRST by category of need of the buyer for those different goods (survival > defense > civ mandates > luxury > business trades), and then within each of those categories, sort by amount of consumer surplus (in the sense of the economics term "consumer surplus" NOT anything previously mentioned in this thread) of each trade for the buyer.
4) Go and make a similar list for the other direction of trade.
5) Add up the total prices (price per unit * total units) for every trade in both lists. One total will be lower. Take the lower total, and figure out how far down in the other list you get without overshooting. For example, the total price of one whole list might be 765 units of value. The other list might go (500, 100, 50, 80, 100, ...) So right after 80 is the cutoff (closest without overshooting), which adds up to 730.
6) Every trade up to that point and on the entire smaller list takes place - those items are all now swapped.
7) Whatever the disparity was (in this case 765-730 = 35) becomes an additional incentive to whichever guy would be getting screwed over. Take that opportunity cost into accoutn and recalculate that final trade equilibrium (the one that was 100 in the above list). Find a new equilibrium with the 35 as bias, and now trade that amount too, IF it adds up to equal to or lower than 35.
8] IF the person still coming out ahead has coins on hand, or some other small item of value that haven't already been traded, fill in any remaining disparity by just handing over whichever small items are most desired by the guy losing out still. All items are eligible EXCEPT survival-needs items owned by the host (the one who didn't launch the caravan), which are protected for this purpose.
9) Any trivial amount left just gets ignored. Cost of doing business in a mostly barter society.

This is much cleaner and more economically and computationally efficient than my earlier concept.

[GoblinCookie]

What, do the dwarves tell the traders "Oh hold on for 6 months while we go out and do market research to find out if there are buyers, and exactly where they are across the world"? No, they can't know this in time. The offer is unexpectedly thrust upon you by the caravan coming with a shit ton of chairs. You need to take risks. You have to buy things you EXPECT to be in demand but might be totally wrong about, and you might have to do so as a snap judgment.

In fact, in DF more than in modern life by far, this is what you have to do, because dwarves don't have telephones.

Nobody's demand is taken into account except those who you are presently due to trade with or are presently due to trade with  you.  The demand cliff is hit pretty fast, not pretty late as a result.  The idea of this is that the majority of your items are unsellable most of the time so that you have to produce specifically in demand items rather than whatever you

The name of this post afterall is more challenging trading, not produce anything you like and chuck onto the market (what we have at the moment).

"Market research" is actually something I'm considering for algorithms, but only as a nice bonus now and then -- paying X amount of resources to finance a couple of members of your civ and some mounts to ride out and figure out trade situations around you (as a player or NPC) to better inform choices. However, this costs money (read: horses and labor time, etc. not literally money  ) and would only be done if info is expected to be out of date and profits likely being missed, and something that might only happen once a season or two tops.

Blindness is still the order of the day, for the most part, and the underlying base system needs to operate profitably in blindness, which means risks and guesses.

It takes less resources to find out beforehand what the settlement wants to buy than it does to cart a load of unsellable resources across the world.  In my trade system humans and AI players sign trade contracts with eachother, promising to provide goods in advance. 

Throwing any old junk on the back of the wagon and praying you find a buyer is how things work at the moment and I am trying to get rid of that situation. Trade between companies and governments does not usually work on that basis, they usually buy and sell on contracts signed before any goods change hand and then the caravans/lorries/trucks are sent.  Basically large scale trade generally does not work in the same way that trade between shops and individual customers works yet at the moment the caravan behaves like a shop, which is complely economically stupid and unrealistic. 

In my system we have trade contracts and a system of competitive bidding.  People who want something and cannot get it will often try and outbid the existing trade partners this would effective drive down the prices.  As the demands of existing players increase they will try to increase the price of existing trade agreements, this drives up prices. 

If a caravan turns up in your settlement it will have some goods that it brought for you and other goods that are for other people somewhere else along the chain.  You can get those goods only if you sufficiantly outbid the AI player, at which point the AI settlement will then assume that you demand those goods and will likely try to acquire enough of those goods for both you and the other settlement both. 

No no, per settlement. Not per critter. Although either would still be trivial computation, since you only need to calc them once per season (maximum frequency caravan schedule. (We're talking on the order of one or two arithmetic problems per tick even for critters. One or two per real life second for settlements)

While that would work well it is still inferior to my own system.  Your system cannot represent the state of affairs by which a starving man will eat anything while a well-fed person will want to eat say specifically goat meat because that is what he likes.  If a settlement gets goat meat or plump helmets in your system, they are still getting their needs met. 

In my system the lower affluence need (any food) is met by being provided with plump helmets the higher affluence need for goat meat is not.  If you provide them with goat meat though, the lower any_food demand is met at the same time.  Both our systems would work, but mine better models consumer demand than yours does.

Sorry to double post, but only a few awkward minutes apart:

ORDER OF TRADE (for NPC-NPC, even though I use terms like "you")
EDITED entirely 8/5/14 2:18am central time, because my previous post was crap. This is an infinitely better algorithm:

1) First, go through both lists of needs from each trade party, cancel out any needs that match across both parties. I.e. if we both demand food, remove food from both lists.
2) Now just choose one party, doesn’t matter which, and go down their list of things to sell. Match up each one to the other party’s demand curves for those things to the seller's original supply curves, and decide on prices and quantities from equilibrium points. Keep track of each, as well as the economic surpluses and total prices per trade. You'll have a long list at the end.
(note: you can't actually change the amount you have with you in your caravan - too late now, but original supply curve is still important to make sure you aren't throwing away value. You would have used it earlier on to guess the right amount to send in the caravan, but now you're double checking your guess against reality. If the ideal quantity is above what you have on you, just sell it all. If the ideal quantity is below what you have with you, then you withhold the remainder and bring it back home with you. Hopefully you guessed spot on, and such inefficiencies will be minimal)
3) Sort that list FIRST by category of need of the buyer for those different goods (survival > defense > civ mandates > luxury > business trades), and then within each of those categories, sort by amount of consumer surplus (in the sense of the economics term "consumer surplus" NOT anything previously mentioned in this thread) of each trade for the buyer.
4) Go and make a similar list for the other direction of trade.
5) Add up the total prices (price per unit * total units) for every trade in both lists. One total will be lower. Take the lower total, and figure out how far down in the other list you get without overshooting. For example, the total price of one whole list might be 765 units of value. The other list might go (500, 100, 50, 80, 100, ...) So right after 80 is the cutoff (closest without overshooting), which adds up to 730.
6) Every trade up to that point and on the entire smaller list takes place - those items are all now swapped.
7) Whatever the disparity was (in this case 765-730 = 35) becomes an additional incentive to whichever guy would be getting screwed over. Take that opportunity cost into accoutn and recalculate that final trade equilibrium (the one that was 100 in the above list). Find a new equilibrium with the 35 as bias, and now trade that amount too, IF it adds up to equal to or lower than 35.
8] IF the person still coming out ahead has coins on hand, or some other small item of value that haven't already been traded, fill in any remaining disparity by just handing over whichever small items are most desired by the guy losing out still. All items are eligible EXCEPT survival-needs items owned by the host (the one who didn't launch the caravan), which are protected for this purpose.
9) Any trivial amount left just gets ignored. Cost of doing business in a mostly barter society.

This is much cleaner and more economically and computationally efficient than my earlier concept.

1.  How are the trade parties determined?
2. 
3.  Essentially we are getting rid of the stuff belonging to categories that are superabundant and acquiring things belonging to categories that are relatively scarce. Is that the case? 
4.  Both parties are doing this. Okay.
5.  How are the prices determined? By the hardcoded value or items or some artificial system?
6. We could have made this simpler and just had them trade their actual surplus for their actual demands directly.
7. Is that really neccesery?  Why are we not simply creating a fair trade to start off with.
8. There is no attempt to rank the relative importance of said items like there is in mine?
9.  Okay. 

[GavJ]

It takes less resources to find out beforehand what the settlement wants to buy than it does to cart a load of unsellable resources across the world.  In my trade system humans and AI players sign trade contracts with eachother, promising to provide goods in advance. 

Not sure what you're on about. I said in the thing that you just quoted, that settlements would invest money to go gather information now and then if it's out of date.

But that's a far cry from auction "competitive bidding" system on contracts. Can you please explain how you propose 50ish settlements in a trading region to not only keep up to date on all their information, but send out contracts to each other every year, and all bid on everybody else's contracts "competitively" (are they outbidding in response? etc.). This implies hundreds of trips every year. Which A) how is that possible in medieval times? They don't have interstate highways or light rail... and B) this is no longer necessarily implies a cost savings, because dozens or hundreds of trade diplomacy trips are not necessarily cheaper than a caravan or three.

Happy to consider logical efficient systems if possible though.

Your system cannot represent the state of affairs by which a starving man will eat anything while a well-fed person will want to eat say specifically goat meat because that is what he likes. 

If you look back at the earlier post with the first demand curves posted, there were both survival and also "luxury" ones.
To clarify: a good is not restricted to one category. It can be listed (and generally should/would be) under all categories that apply.

This is a slight economics shortcut to save massive calculations over the alternative of recalcing curves every single object traded. Regardless, it will definitely do as you say, because they'll continue to also buy food as luxury items, after having bought other food for the survival purposes.

Essentially we are getting rid of the stuff belonging to categories that are superabundant and acquiring things belonging to categories that are relatively scarce. Is that the case?

No, that's only half of the equation. What you seem to mean by "surplus" is something along the lines of an inverse of "how much you value this stuff"
What I mean by surplus is this, the actual economic term:
http://en.wikipedia.org/wiki/Economic_surplus
which by contrast (for a seller) is "the difference of how much I would have been willing to accept as exact value of this thing to me, versus how much you actually gave me"
In other words, profit.

The highest profitability COULD be from those dusty old warehouses full of scrap metal you have in the back of your base.
OR the highest profitability COULD be from your most highly prized possession! If the amount that the other trader prizes it is sufficiently higher than how much you prize it, a bigger gap than any other good, then it's the best thing for you to trade to gain value for your fort, and you trade it first.

All depends on the other guy as much as you. That's why you need curves, for finding equilibrium between two parties and economic surplus values.



Doing it with only half the equation is like trying to do linear geometry with slopes, but never intercepts. Might be slightly better than nothing, but still terrible at efficiency.

[GoblinCookie]

Not sure what you're on about. I said in the thing that you just quoted, that settlements would invest money to go gather information now and then if it's out of date.

But that's a far cry from auction "competitive bidding" system on contracts. Can you please explain how you propose 50ish settlements in a trading region to not only keep up to date on all their information, but send out contracts to each other every year, and all bid on everybody else's contracts "competitively" (are they outbidding in response? etc.). This implies hundreds of trips every year. Which A) how is that possible in medieval times? They don't have interstate highways or light rail... and B) this is no longer necessarily implies a cost savings, because dozens or hundreds of trade diplomacy trips are not necessarily cheaper than a caravan or three.

Happy to consider logical efficient systems if possible though.

No all settlements guess what the other settlements supply and demand should be based upon what they should be producing and demanding based upon their population and environment.  That is how trade routes are initially formed.  They bring the goods they guess should be in demand.  Once they have visited however, they start to keep track of the actual state of affairs in the settlement not what they guess should be the case. 

It is only once the trade routes are formed however that bidding wars and organised systems of trade contracts start to happen.  If a caravan goes through a settlement, the cargo it is carrying is either earmarked for sale to their settlement (in which case it can be bought immediately) or it is reserved for sale to another settlement. 

One advantage of this over the present system is no long will we be bombarded with a list of every good under the sun.  If we (or an AI settlement) want a certain good badly enough however and it is earmarked for sale to another settlement on the trade route, we can outbid the other settlement and get those goods instead of them.

If the good is not actually scarce in the area, all this will mean in the long run is they will bring double the amount, one lot for you and another lot for the other guy.  If the good is scarce then the other guy , desperate to try and meet his demand will outbid you offering an even better deal to the deal you are offered causing 'prices' to rise.

This is essentially the mechanics of how supply and demand affect prices.

If you look back at the earlier post with the first demand curves posted, there were both survival and also "luxury" ones.
To clarify: a good is not restricted to one category. It can be listed (and generally should/would be) under all categories that apply.

This is a slight economics shortcut to save massive calculations over the alternative of recalcing curves every single object traded. Regardless, it will definitely do as you say, because they'll continue to also buy food as luxury items, after having bought other food for the survival purposes.

Our ideas on consumer demand are basically similar then.    

Trouble with your idea is still that it cannot really represent the relative nature of luxury and the way that luxuries become treated as necessities as the wealth of an individual or group increases (what affluence represents).  A luxury good is a just a luxury good in your system, but a bed to a settlement starving to death is a luxury in a way that a bed certainly is not to a prosperous settlement. 

If you resolve that issue, you basically have a more complex version of my idea with redundant categories rather than 0,1,2,3,4,5,6,7 etc. 

No, that's only half of the equation. What you seem to mean by "surplus" is something along the lines of an inverse of "how much you value this stuff"
What I mean by surplus is this, the actual economic term:
http://en.wikipedia.org/wiki/Economic_surplus
which by contrast (for a seller) is "the difference of how much I would have been willing to accept as exact value of this thing to me, versus how much you actually gave me"
In other words, profit.

What I mean by surplus is the dictionary definition of surplus.
http://www.thefreedictionary.com/surplus

What you are using is a concept that only exists within a particular economic theory which is opposed by other economic theories.  Obvious you are a supporter of this economic theory, as is your right; but others do not hold to this economic theory and I appear to be one of them.
http://en.wikipedia.org/wiki/Mainstream_economics

It is better to use dictionary definitions only, if your economic theories are true within the context of dwarf fortress society then they will emerge spontaneously; the game does not exist to be your economist guinea pig essentially. If you want to apply economic theories then this is fine but use dictionary definitions for things and do not attempt to overtly impose them as game logic. 

The highest profitability COULD be from those dusty old warehouses full of scrap metal you have in the back of your base.
OR the highest profitability COULD be from your most highly prized possession! If the amount that the other trader prizes it is sufficiently higher than how much you prize it, a bigger gap than any other good, then it's the best thing for you to trade to gain value for your fort, and you trade it first.

What is profitability?  Profitability in present game terms is merely the accumulation of an arbitrary set of values, at present essentially fixed by the raws.  My system is not based upon any arbitrary or calculated set of economic values, ideally the economics/trade system will eventually be able to dispense of them entirely. You will be able to trade this amount for that amount, that is all that will exist. 

The AIs in my system are programmed to attempt to acquire as many in demand goods (including for trade with a third party) as possible using their finite surplus goods.  A good deal is simply where they get as many in-demand goods as possible in total without sacrificing any in-demand goods.

Yet as they get in-demand goods however, their affluence increases and they start to demand more 'luxury' goods.  This means that the economy never grinds to a standstill because the more you get, the more you want and eventually what you want is an infinitely increasing amount of money. 

All depends on the other guy as much as you. That's why you need curves, for finding equilibrium between two parties and economic surplus values.

Doing it with only half the equation is like trying to do linear geometry with slopes, but never intercepts. Might be slightly better than nothing, but still terrible at efficiency.

The things you describe will spontaneously emerge in my system without anyone having to actually add them in; if they actually exist and work the way they are described that is.

They do not have to be arbitrarily imposed on the system by the game mechanics!

[GavJ]

It is only once the trade routes are formed however that bidding wars and organised systems of trade contracts start to happen.  If a caravan goes through a settlement, the cargo it is carrying is either earmarked for sale to their settlement (in which case it can be bought immediately) or it is reserved for sale to another settlement. 

Yes of course if you are passing through, you'll take note of their trade situation to bring back as info later. What I'm still confused about is the logistics, though:

1) How does one settlement telepathically negotiate contracts with a passing caravan on behalf of another settlement 5 days' ride away...?
2) Why would you be "just passing through" settlements that you didn't set out to bring goods for? This may happen rarely, but most settlements are not going to just happen to be on the route to the one you're sending a caravan for. Most of them are going to be way off to the side and not your current destination, so how do you learn about them? If a place is of your own civilization then they MIGHT share their most recent knowledge of places they have traded with, but that's now at least two links out of date, and possibly several links out of date (could be years and almost totally  useless). And if they aren't your same civ, they likely would not share such information for free with a potential competitor. Perhaps for a fee, but you have to now decide if you pay it, which is essentially the same model as sending out diplomatic missions.

Trouble with your idea is still that it cannot really represent the relative nature of luxury and the way that luxuries become treated as necessities as the wealth of an individual or group increases (what affluence represents).  A luxury good is a just a luxury good in your system, but a bed to a settlement starving to death is a luxury in a way that a bed certainly is not to a prosperous settlement.

Luxuries never become "necessities." Not for humans in real life, not even in DF's thoughts system (dwarves do not start getting more negative thoughts if you don't upgrade them). So no, my system doesn't do that... on purpose, because it's not a real thing.

The other concept here is the question of whether you would trade one really high end chair for, say, 12 cheaper chairs whose overall value adds up to more. For humans, this is a tricky question. Generally you SHOULD make that trade, even if you insist on only sitting on nicer chairs. Why? Because if the 12 add up to more total value, then it means you can just turn around and sell them again for an even nicer single chair than you began with! So it's still nearly always a good trade, as long as you make profit.

But also, we have to consider that dwarves have no conception of single nice things being better than lots of cheaper things that add up to more. The value of belongings and rooms is linear: if you literally cram a room with 12 chairs covering every square inch, worth a combined total more than one golden chair, the dwarves will consider it to be a nicer room. Even if that's not your personal preference, you're not a dwarf / the game works differently currently. And the model captures current rules perfectly as described so far.

If one did, however, prefer to model single, nicest items as having a bonus preference over many cheaper ones, then this could be done with a simple morphing of demand curves to have a slightly steeper negative slope.

What is profitability?

I quantitatively and precisely defined what I meant. Here it is again:
http://en.wikipedia.org/wiki/Economic_surplus
This is how it works. Again, I'm not arguing the basic validity of economics until/unless you provide more human subject and market data to the contrary than experts have gathered for the last 300 years.

[GoblinCookie]

Yes of course if you are passing through, you'll take note of their trade situation to bring back as info later. What I'm still confused about is the logistics, though:

1) How does one settlement telepathically negotiate contracts with a passing caravan on behalf of another settlement 5 days' ride away...?
2) Why would you be "just passing through" settlements that you didn't set out to bring goods for? This may happen rarely, but most settlements are not going to just happen to be on the route to the one you're sending a caravan for. Most of them are going to be way off to the side and not your current destination, so how do you learn about them? If a place is of your own civilization then they MIGHT share their most recent knowledge of places they have traded with, but that's now at least two links out of date, and possibly several links out of date (could be years and almost totally  useless). And if they aren't your same civ, they likely would not share such information for free with a potential competitor. Perhaps for a fee, but you have to now decide if you pay it, which is essentially the same model as sending out diplomatic missions.

1.  There is no telepathy as such (long range anyway).  I have decided that the noble that 'owns' the caravan should travel seperately to the caravan itself visiting settlements according to the caravan logic that is used in order to negotiate trade contracts with settlements, including the human one.  Once trade contracts are negotiated, then the caravan is rerouted accordingly but the caravans normal journey is not initially changed (basically it does not wander about, it's noble does that instead).

2. I was talking about existing trade routes.  A player can 'outbid' existing trade contracts by offering more in demand items in return than is being offered already.  That allows the 'stealing' of trade off eachother.  The visiting noble described above can also outbid as well in making his initial caravan offer. 

Luxuries never become "necessities." Not for humans in real life, not even in DF's thoughts system (dwarves do not start getting more negative thoughts if you don't upgrade them). So no, my system doesn't do that... on purpose, because it's not a real thing.

You could not be more wrong.  Neither a bed nor a chair are neccesities, in fact only some kind of food and water are neccesery in the strict sense of the word and it is not 'neccesery' strictly speaking that you survive. 

You can sleep on the ground, you can sit on the ground.  In your view there are now very few items that are not luxuries given that one can live an utterly wretched existance without anything but what you need merely to survive.

A person that has never had a chair or a bed is not as desperate to acquire either item as a person that has always had them and taken them for granted.  In this sense the chair/bed is treated as a neccesity when for the other person it is a luxury.

The other concept here is the question of whether you would trade one really high end chair for, say, 12 cheaper chairs whose overall value adds up to more. For humans, this is a tricky question. Generally you SHOULD make that trade, even if you insist on only sitting on nicer chairs. Why? Because if the 12 add up to more total value, then it means you can just turn around and sell them again for an even nicer single chair than you began with! So it's still nearly always a good trade, as long as you make profit.

The question is actually really a question of utility(?).  Why are seeing to acquire those chairs in the first place and how many chairs do you need?

If you are seeking a single chair for your party's use then it is irrational to carry out that trade.
If you are seeking a single chair for your paty's future use then this trade is rational.
If you need 12 chairs for your party's use, then it is rational to trade your single superior chair for 12 inferior one's.
If your party has already met it's demand for chairs then what you are describing is rational.

You see what you are describing is exactly how the AI could spontaneously works in my system when it wants to acquire the ultimate chair but has a surplus of ordinery chairs.  It can take a surplus (dictionary definition) chair of good quality to one settlement, then trade that chair for 12 normal chairs and then trade it to a third party for the ultimate chair that is in demand. 

But also, we have to consider that dwarves have no conception of single nice things being better than lots of cheaper things that add up to more. The value of belongings and rooms is linear: if you literally cram a room with 12 chairs covering every square inch, worth a combined total more than one golden chair, the dwarves will consider it to be a nicer room. Even if that's not your personal preference, you're not a dwarf / the game works differently currently. And the model captures current rules perfectly as described so far.

Perhaps you are right.  It does not matter because the demand for chairs is quite seperate from the demand for a nice room.  Dwarves get positive moodlets for having a nice room and negative moodlets for not having enough chairs.

If one did, however, prefer to model single, nicest items as having a bonus preference over many cheaper ones, then this could be done with a simple morphing of demand curves to have a slightly steeper negative slope.

That is a problem with my system.  I have not thus far really modelled quality rather than quantity of type of item in my system.  Since I am going away in 2 days, I really need to explain how it could work. 

The basic idea is that quality of an item is essentially just a competitive advantage over equivilant items.  Trades would have a quality factor, which is 0 in the case that the items of a type are all of base quality.  All else being equal, AIs will make trades that carry the highest quality factor.  The catch here is that they would only tolerate a small negative divergance from the prior quality as valid, if you give them a total of items worse than the existing quality they will not initially do anything but they will secretly look to replace you as trade partner even if means them taking a worse deal.

Quality is mainly of commericial use then in the competitive bidding, if you offer better quality goods then you can outbid your competitor without offering a better price than they do.  However this comes at a catch, it means that you are expected to continue to provide them items of equivilant quality.

The other question is how to model quality of AI products.  I suspect the way we should do it is to have fortresses as a whole gain experiance at producing certain items over time and lose it from employing new people or losing existing workers to old age. 

I quantitatively and precisely defined what I meant. Here it is again:
http://en.wikipedia.org/wiki/Economic_surplus
This is how it works. Again, I'm not arguing the basic validity of economics until/unless you provide more human subject and market data to the contrary than experts have gathered for the last 300 years.

You seem not to have heard of.
http://en.wikipedia.org/wiki/Heterodox_economics
There is no 'validity of basic economics here', there are different and opposing theories of economics around to the one's you expound. You are perfectly free to believe 100% in mainstream economics but that does not mean you can dismiss everyone else who does not as economically illiterate/idiot.  There are other points of view, they may be wrong in your opinion but that does not mean you can pretend your people are the only economists in town.

The reason I ask the question about profit is quite simply because of a certain line in the wikipedia description of your favourite economic concept. 

Consumer surplus or consumers' surplus is the monetary gain obtained by consumers because they are able to purchase a product for a price that is less than the highest price that they would be willing to pay.

What does this monetary gain consist of in dwarf fortrees?  What are people saving by not selling a surplus item in dwarf fortress.  The value of the item is measured in what?  At the moment it is measured based upon a fictitious currency called 'dwarfbucks' which is as artificial and contrived as the present economic system of caravan trade is. 

The item's value is measured in what? Plump helmets, coins, gemstones, metal bars?  You see the whole problem with the application of your economic idea, there is no inherantly existing thing against which to measure the value of anything except the fictitious dwarfbuck measurement. 

A chair in this context has a plump helmet price from the food producing settlement over there and an iron price from the mining settlement over here.  You are aiming to get rid of as many chairs as possible to as people as possible in order to collect as many different sorts of 'prices' as possible.  Hoarding a stockpile of surplus chairs is fundermentally undesirable, since it means not collecting on all possible prices that the chair could fetch; inherantly they are worthless to you. 

If you get rid of the mining settlement we can fetch only kind of price, the plump helmet price then you will happily trade 100 chairs for 10 plump helmets that you need.  If there are multiple possible prices however it pays to be stingy and only give them 10 chairs because you can use then use other 90 to get iron.  Needless to say if there is no other prices, it's charcoal time for the chairs. 

[GavJ]

I have decided that the noble that 'owns' the caravan should travel seperately to the caravan itself visiting settlements

This doesn't address the question. Noble, diplomat, whatever, the problem isn't one of rank. The problem is that as soon as you leave your settlement, you don't know what's in the settlement anymore or what their demands are. And until you get to your destination, you don't know whether things will have changed there. So you can't have competitive bidding, because no matter what pattern you run, the most people available at any given time with full knowledge of what's available to trade = 2: a caravan knowing what it carries and a settlement it is sitting at, and vice versa.

Nobody else outside (such as your next destination) knows either of those things up to date. So how can they put in a higher bid to protect their goods? They can't, because they neither know what's going on, nor do they have a means to actually place the bid. So then what happens is you arrive there, and find out that they are furious at you for dishonoring your trade agreement by selling their earmarked stuff along the way, AND they would have been willing to outbid the other guy anyway maybe! So you're ripping people off AND you can't even know that it's profitable to do so. That's not very strategic.

The best you can do is shuttle info around to everybody you meet with a caravan or diplomat, but all that does is keep them more up to date than if they had to collect info on their own. It doesn't allow 3- or more-way dynamic bargaining and bidding. For that you'd have to have, like, a trade summit or something, which is out of time period and also STILL involves working on out of date info (since all but one of the summit members won't be up to date on their home settlements while they are at the summit...)

You could not be more wrong.  Neither a bed nor a chair are neccesities, in fact only some kind of food and water are neccesery in the strict sense of the word and it is not 'neccesery' strictly speaking that you survive. 

Arguing that OTHER things also aren't necessities does not contradict my claim that luxuries aren't necessities...

Quality is mainly of commericial use then in the competitive bidding, if you offer better quality goods then you can outbid your competitor without offering a better price than they do. 

What you are describing is economic surplus. Higher value for the same price = more profit above what you would have been willing to pay.

However this comes at a catch, it means that you are expected to continue to provide them items of equivilant quality.

I don't have any experience with this "humans are quality ratchets that can't spin down ever again once they spin up" idea in my own life. I quite commonly do things like buy a really nice tweed 3-piece suit and then go out the next day and buy a t shirt and jeans. Or buying "statement pieces" of furniture and then buying and using a bunch of cheaper stuff around them. Nor would I have any grudge against a salesperson who formerly sold me such things, listening to my budget later and simply offering whatever was best for me at the later time, nicer or less nice.  Where are you getting this concept from?

http://en.wikipedia.org/wiki/Heterodox_economics

Note that "stuff you made up for purposes of this thread" does not qualify as "heterodox economics."

If you have something specific that you've already argued or wish to argue that you think would be better for dwarf fortress that is proposed by a non-mainstream legit economist, then by all means, explain what it is, which professional economist(s) supports it, and why it's a better choice for the game, and you might be right. But you still have to cite professionals and make a solid case for why a couple of dudes with a weird theory are more right for our game design purposes than 5,000 dudes supporting the mainstream.

Consumer surplus or consumers' surplus is the monetary gain obtained by consumers because they are able to purchase a product for a price that is less than the highest price that they would be willing to pay.

Yes, they wrote it like that because every person in the modern world reading the article lives in a monetized economy, and that's what's relevant to 99% of people's modern context.
The same concepts hold equally true for bartering, though, if you happen to be the 1% of people programming a medieval dwarven economy simulator.

Value is just value. Chemicals in the brain, more pleasure less pain, whatever. You don't need to know the science of exactly what value is physically to model it economically. The fact is that people do trade, and that implies that there does indeed exist some metric by which people decide that two things being swapped have some sort of equivalency to them. I.e. value. If different things had no common ground in holding value, then trade wouldn't exist.

"Dorfbucks" is an unfortunately misleading term implying money even though money is not involved, but it really is a measure of something along the lines of amount of seratonin/dopamine/lack of cortisol/etc.

[GoblinCookie]

This is the last day I can post before I go away for 2 weeks.

This doesn't address the question. Noble, diplomat, whatever, the problem isn't one of rank. The problem is that as soon as you leave your settlement, you don't know what's in the settlement anymore or what their demands are. And until you get to your destination, you don't know whether things will have changed there. So you can't have competitive bidding, because no matter what pattern you run, the most people available at any given time with full knowledge of what's available to trade = 2: a caravan knowing what it carries and a settlement it is sitting at, and vice versa.

Easy: he knows what he has been sent to buy and he also know what surplus your settlement had to sell when you left. Things that are due to be traded are not considered surplus. 

And should a settlement somehow manage to lose the goods it was due to trade it can normally make more and it has all the time until the caravan returns to base to do this. 

Nobody else outside (such as your next destination) knows either of those things up to date. So how can they put in a higher bid to protect their goods? They can't, because they neither know what's going on, nor do they have a means to actually place the bid. So then what happens is you arrive there, and find out that they are furious at you for dishonoring your trade agreement by selling their earmarked stuff along the way, AND they would have been willing to outbid the other guy anyway maybe! So you're ripping people off AND you can't even know that it's profitable to do so. That's not very strategic.

The lose their goods, that is the point.  It is supposed to punish people who are too stingy and press too hard a bargain just because they can at the moment.

Arguing that OTHER things also aren't necessities does not contradict my claim that luxuries aren't necessities...

We are talking in economic terms here are we not?  An item starts to be treated economically as a neccesity as affluence increases rather than a luxury.

This is the main reason why there is so much consumer debt in the real world. 

What you are describing is economic surplus. Higher value for the same price = more profit above what you would have been willing to pay.

No I am not. The higher quality items are not adding any value to themselves, they are simply making other lower-quality items worthless. 

They are not worth any more, other items are worth less.  That is because they are now 'mere surplus' and thus utterly worthless to anyone. 

However this comes at a catch, it means that you are expected to continue to provide them items of equivilant quality.
I don't have any experience with this "humans are quality ratchets that can't spin down ever again once they spin up" idea in my own life. I quite commonly do things like buy a really nice tweed 3-piece suit and then go out the next day and buy a t shirt and jeans. Or buying "statement pieces" of furniture and then buying and using a bunch of cheaper stuff around them. Nor would I have any grudge against a salesperson who formerly sold me such things, listening to my budget later and simply offering whatever was best for me at the later time, nicer or less nice.  Where are you getting this concept from?

You are not talking about a tweed 3-piece suits vs a teashirt.  Those are not items of comparable quality, they are completely different items.  It is a better tweed 3-piece suit vs a worse one, not suits vs teashirts. 

If you consistantly shop for teashirts at a store and suddenly the teashirts you are sold are moth-eaten and full of holes, you are probably going to go to another store.  That is where I got the concept from. 

Note that "stuff you made up for purposes of this thread" does not qualify as "heterodox economics."

If you have something specific that you've already argued or wish to argue that you think would be better for dwarf fortress that is proposed by a non-mainstream legit economist, then by all means, explain what it is, which professional economist(s) supports it, and why it's a better choice for the game, and you might be right. But you still have to cite professionals and make a solid case for why a couple of dudes with a weird theory are more right for our game design purposes than 5,000 dudes supporting the mainstream.

We are not here to argue economic theories.  We are here to come up with suggestions to improve the game that make economic sense solely in the context of the game.  It may well be that in the real-world one economic theory is correct while a completely new economic theory is needed for the game.  I am not making stuff up for the purposes of the thread, I am created a dwarf fortress economic model (rather than taking one off the shelf from the real-world as you have done). 

Yes, they wrote it like that because every person in the modern world reading the article lives in a monetized economy, and that's what's relevant to 99% of people's modern context.
The same concepts hold equally true for bartering, though, if you happen to be the 1% of people programming a medieval dwarven economy simulator.

They do not hold equally true for bartering.  That is the problem with the applicability of your theory to the game.  They can only apply in a world where things can be easily moneterised (our modern world).  My theory on the other hand is based on the actual economics of bartering in dwarf fortress terms. 

In bartering trade operates according to a model by which everyone divides the surplus (dictionary definition) up among the number of people who can offer you something in demand.  The complexity comes in when you are demanding things because they are demanded by others.  You try to get as many in-demand items as possible and divide up your surplus among them. 

Surplus is totally worthless, that is what a barter economy is built on.  The barterer has utterly no use for the item he is selling, he loses nothing by selling it; assuming that is that noone else is around to offer him something else he demands in return for that item. 

In a barter economy the objects being sold are ideally worthless to the seller and their entire value comes from the demanded goods that other parties can offer in return for the items.  In a non-ideal state relative 'luxury' goods are being traded to acquire 'neccesities' by one or both party.

You aim to avoid trading any objects that actually have value to you if at all possible.  That actually means all trade is profitable as long as you do not sell in-demand items or use up in-demand items in transaction costs or use up the labour time that can be used to produce such items.   

Value is just value. Chemicals in the brain, more pleasure less pain, whatever. You don't need to know the science of exactly what value is physically to model it economically. The fact is that people do trade, and that implies that there does indeed exist some metric by which people decide that two things being swapped have some sort of equivalency to them. I.e. value. If different things had no common ground in holding value, then trade wouldn't exist.

"Dorfbucks" is an unfortunately misleading term implying money even though money is not involved, but it really is a measure of something along the lines of amount of seratonin/dopamine/lack of cortisol/etc.

Value then is demand, exactly what it is in my system.  Surplus objects (dictionary definition) have no value as they do in my system.  As you describe it I am modelling value rather well.

Yet Dorfbucks are the only thing that can make your mainsteam conomic models actually apply to the game.  They are a contrived concept imposed on a barter economy in order to make the equally contrived present caravan model work. You could build a whole contrived mainsteam economic model for the world based up dorfbucks and it will probably work as long as you fix the values correctly.  That however does not mean I think that you should go down on this route, building a contrived world economy instead of a contrived caravan visiting one settlement.

You have yet to explain what profit means without using dorfbucks.  I want to get rid of dorfbucks ultimately because they should not be needed with a proper economic model.

[Scruiser]

GoblinCookie,  I don't have the motivation to both explain and justify economics in general via forum post, so I will just ignore your posts for now.

GavJ,  Keep in mind when you are using demand curves, that they tend to form those nice smooth curves when there are multiple buyers and sellers.  You can get some pretty weird slopes on the curves when it is just a single buyer and/or a single seller with only certain levels of information on the other buyer or seller.  I think a micro-economics textbook might be able to give you some examples.  The best example I can recall is with a monopoly where the seller will sell a lower quantity at a higher price to maximize profit.  The quantity and price are still determined by drawing graphs and all, but instead of looking at where the supply and demand curves intersect, the seller will draw an additional curve... can't quite remember how to do it (been a few year since microeconomics), but you get my overall point.  Take into account that with a medieval level of understanding, the single seller won't actually know the optimal price and quantity to get the most  out a monopoly and you can get some pretty complex situations.