NPC customer formula

[Topazan]

So let's say players can set up shops to sell items from their inventory.  These shops can serve both other players and NPCs.  How do you create a formula to determine what the NPCs buy and from who?

I vaguely remember some microeconomic concepts that could be useful here, but I can't remember the correct names to look them up and refresh my memory.  Maybe someone here knows the proper names.

1.  Certain products increase demand for other products.  Eg.  An abundance of cars will lead to a high demand for gasoline.
2.  Certain products reduce demand for other products.  Can't think of a good example, but it's usually because the products are similar enough that if you have one, you don't need the other.
3.  People buy more when the price is low and less when it's high.  I think this is usually represented as a curve.

If anyone knows the names of these concepts in economics, let me know.

So how do we make this into the formulas we need?  The questions are:

1.  How many of each item do the NPCs buy?
2.  Who do they buy them from?

For simplicity's sake, we'll assume that all items of the same type are identical, but there may be bonuses to certain shops related to things like marketing or advertising.

For bonus points, what about NPC sellers?

I'll post my attempt shortly.

[133794m3r]

if we're speaking of basic economics only, then have them decide it based upon how much advertising the person has done and as you said how well it's known so it could be.

commonality=(ads*ts)+(md*en);
probability of purchase = (commonality+disposition+sale+distance+prices);
ads=amount of ads that they've shown and to how many if the person know of them. ts=the individual effect it's had on the npc/player. md=marketing demographic aka how rich that they are etc. and en=how effective it was on the npc aka what demographic your'e going to put them in.

dispoisiton is how much they like the shop/'brand name' so thus it's a value of 0-100 0 being they ahte it and 100 meaning loyal to it.
sale=if it's on sale they'll of course go after it.
distance=how far they are from it the closer themore likely that they'll do it.
prices=how much they are compared to everyone else.

The more that someone buys an item from a certain store the more likely they'll come back again.


Also as you said about the demand and such will have to be something that you'll have to figure out yourself.

The reason for that other thing is because i cannot give you a formula for it, since i don't know what all you have within your game world.

[Topazan]

Thanks, that's a good start.  I didn't even think about demographics. However, I was going more for a system based on supply, demand, and competition between players than marketing.

I found this wikipedia article about various types of goods in economics.  It's very interesting, and a lot of it can be applied to a system like this.

If anyone's curious, these are the economic concepts I was trying to remember:
1.  Complimentary good.
2.  Substitute good.
3.  Price elasticity of demand.

Turns out there are even more types of good that affect demand.

Ok, how about for base demand there's an NPC population count.  Scaled with their population is a base demand for each type of item, and a budget for it.  When their demand for each item they need is fully met, their population will slowly increase.  If they fail to meet one or more of their demands, their population will slowly fall depending on how severe the shortage is.

budget & items consumed based on current population
items purchased = budget / average price
population growth = items purchased - items consumed

Where it gets tricky is once you factor in marketing.  If a higher priced shop has better marketing, will the NPCs blow their budget shopping there rather than buying enough to fulfill their needs?  That could get frustrating for players who want the npcs to expand their population.

Maybe you could make it so that NPCs get "smarter" when they're in danger of falling short of their demand, and just buy the cheapest in that situation.  But that could be frustrating for players who spend a lot of time on advertising.

That principle of price elasticity of demand could be tricky too.

[Chris]

2.  Certain products reduce demand for other products.  Can't think of a good example, but it's usually because the products are similar enough that if you have one, you don't need the other.

Answer in eRPeGish terms Bow and crossbow (substitute goods). Bow and arrows (complementary goods). Crossbow and bolts (complementary goods). Increased demand for bows decreases demand for bolts because these are complementary goods to substitute good (crossbows).

Take a note that in reality there never are full/pure substitutes. Like that might be a guild of crossbowers that do not accept members who use bows, for them bow and crossbow are not substitute goods. There might be also blacksmith that that made modified crossbows that can fire arrows instead of bolts which makes crossbow and arrows complementary goods.

3.  People buy more when the price is low and less when it's high.  I think this is usually represented as a curve.

More like stright lines, the meeting point of supply and demand line determine the good price.

If a higher priced shop has better marketing, will the NPCs blow their budget shopping there rather than buying enough to fulfill their needs?

No. Different target react different way to marketing effort. There is also the cost of the purchase (like if the shop is in a long distance it is expensive to you to buy it, that's why there still exist tiny shops in the era of supermarkets, you won't go to the other end of the city to buy matches, even if these were twice cheaper than those sold next to your house).


In short, there are so many factors involved you might just select an arbitrary number of the demand and use it, no one will say it is not realistic
I would say, make a shop radiate area of influence. The influence is extremely strong near the shop, even if the goods are total rubbish (shoping convenience and local passive marketing). Marketing and goods quality increase the area.

[saljutin]

just one thing...you want to create a GAME and NOT economical simulator
so yes you can "find" out your own formulas and test them first (like excel or smth) and see how they react, then test them in alpha version of game.

once I was making a "fast food restaurant manager" - and one of the formulas was that there was impact on how many people will buy goods - that was called advertising so if you advertised on radio 200 POTENTIONAL buyers come...so lets say 10 people wanted to buy "Double cheese burger" ... how many would actually buy it was dependent on price. so if you set price too low you sold em all but lost some profit, if you set price too high you lost costumers

so many things to add
- weather
- winter/summer
- even if you have 200 potentional costumers and you dont have LCD TV then eventually you loose them...instead of 5% people who want LCD there would be only 3%, and then you need some weeks to satisfy demands of those people (have excess of LCDs so they are allways X/X bought) then you can climb back to 5%
- no need to make real simulation, CHEAT to make it feel like its good and deep

[Nox]

I agree, wanted to write something like that before

I would consider for the script to take into account not just game-world situation, but also concrete player's situation, meaning let that when the player plays at least okey, he should get reasonable gains; possibly help weaker player

Plus you'll find out whether it's a good idea at all - Counter Strike: Source originaly had flexible prices of weapons&equipment, but it grew out of proportions and screwed it so they removed it altogether

[Topazan]

just one thing...you want to create a GAME and NOT economical simulator

How about an economic simulation game? 

Ok, y'all.  Suppose you just come up with an arbitrary number for demand.  The NPCs will buy that number at the lowest price.

What if there was a week or so when only one person happens to be selling arrows, at 1000000 gp apiece?

What if the player population spikes unexpectedly, and there's not enough NPC demand to go around, resulting in item inflation?

What if the player population falls unexpectedly, and the NPCs are snatching everything up before the players can?

What if constant routine purchasing results in currency inflation? 

What if a rich player griefs everyone else by selling everything the NPCs buy for 1 gp each?

A simple solution would be to let each shop have its own demand, and that might work fine, but it's not interactive and would lead to inflation in the long run.

__________________________________________
Alright, maybe I have gotten a bit carried away.  Let's set some limits.  The NPC market reacts only to what the player community as a whole stocks in their shops and at what price.  Disregard distance, marketing, and psychological factors.

Is it possible to make a system that adapts to the playerbase, provides challenging economic strategy gameplay, and creates a system of communal progression?

[Topazan]

I think this will work.

Core concept:

There is an NPC population count.

Each product has a base demand per capita and base budget per capita.

The NPCs will buy a given product up to their base demand for it at as cheap a price as possible.

NPC population grows or shrinks based on average (products purchased / base demand).

Possible extensions:
*Complimentary goods -  increase demand for another good when sold.
*Substitute goods - reduce demand for another good when sold.
*Illegal demerit goods- Things like poison, narcotics, illegal weapons.  Filling demand reduces NPC population rather than increasing it.
*Price elasticity of demand-  Could be useful to impede people who drastically undersell the market from driving the price artificially low.
*Demographics- Each demographic could be a different population.  Each one grows or shrinks according to how their needs are met.  Probably not really necessary from a gameplay point of view, but could add flavor.

Only thing I don't like about this is the idea that the NPCs will go straight to the lowest bidder.  Maybe there can be some randomness within a reasonable price range.

[Chris]

You do realize that what you wrote is more complex than any trading simulation game ever made on this planet?

[Topazan]

You do realize that what you wrote is more complex than any trading simulation game ever made on this planet?

You mean the thread as a whole or my last post?  If the latter then it's about time that changed. 

I don't know though, Simunomics seems to have a pretty complex formula for retail sales.  It at least takes a large number of factors into account, including item quality, brand reputation, brand awareness, store reputation, price, quantity, and time.

[Chris]

You do realize that what you wrote is more complex than any trading simulation game ever made on this planet?

You mean the thread as a whole or my last post?  If the latter then it's about time that changed. 

The moment you started talking/thinking/even considering for a second substitute and complementary goods. Games do not touch these.

[Topazan]

Granted.  They're not essential to the system I laid out anyways.

My thinking was that the existence of complementary goods would add another element of strategy to playing the market and predicting what will sell well.  Substitute goods would go hand-in-hand with this, with different substitutions of the same demand having different complements.

It wouldn't be *too* difficult to implement in an oversimplified manner, but you're probably right that the simplest solution that achieves the goals is the best.

That said, since when is "no one's ever done this before" a reason not to do something? 

[Chris]

That said, since when is "no one's ever done this before" a reason not to do something? 

You mean like commiting suicide by getting oneself eaten by a whale? Nah, I would personally go for more traditional stuff like throwing oneself into a pit full of hungry lions

What I mean is, I get a feeling you didn't studied enough how economic games work. Especially older ones when computers were unable to compute it perfectly: Detroit, Motor City, Supersonic Air Business, Pizza Tycoon.

Another thing, you overestimate reality. No one really knows how economy works, not even people whose job is to know it and not even those who teach it at universities. Which is better Mercantylism or Laissez faire? Nowadays and if you live in a rich country you would say Laissez faire, so New Deal was wrong then? Or go to any african country and say you are from the IMF, you will be killed on spot So you say mercantylism then? Still wrong, just recently we got this issue in europe and most countries went for mercantylism while polish prime minister decided on laissez faire and for a while Poland was one of the two countries only with positive GNP, also you can't obviously miss USA with traditional strong LF (but again with occasional betrayal of this ideal via things like New Deal ). So, you are trying to perfectly simulate a thing that no one really know how it works

[saljutin]

it is GAME
it is YOUR GAME
YOU set rules of economy

if you write in rules/faq/help that selling more ducks would increase need for pig food people will KNOW what impacts on something even if it is not logical
so make it logical and playable and FUN

so lets say you will use cron job to buy goods - first in cron count active members (logged in last 24h or smth), and then each active member "produces" 50 costumers with % demands
so when there are low amount of active people lets say 60% will buy bread and 30% will buy milk and 15% will buy alcohol
so many possibilities without any need to make REAL economy simulator...make it FUN and LOGICAL -> more condoms, less baby diapers

[Topazan]

Where are any of you getting the idea that I'm trying to make a perfect simulation?  You're getting overly caught up on complementary/substitute goods, which I only mentioned as a possibility in passing.  It is not the focus of what I'm trying to achieve here.

Never mind what other games have done, look at the area where I marked "Core concept" and tell me what's complicated and difficult about that.  I'll even quote it for your convenience.

Core concept:

There is an NPC population count.

Each product has a base demand per capita and base budget per capita.

The NPCs will buy a given product up to their base demand for it at as cheap a price as possible.

NPC population grows or shrinks based on average (products purchased / base demand).

[Chris]

Where are any of you getting the idea that I'm trying to make a perfect simulation?  You're getting overly caught up on complementary/substitute goods, which I only mentioned as a possibility in passing.  It is not the focus of what I'm trying to achieve here.

It does not matter if it is focus or not. The very fact that you considered it even for a tiny litle second means you went way, way, way overboard.

Never mind what other games have done

You will regret it in the end You are trying to build a space rocket without the basic knowledge of algebra. The outcome is easy to predict
 

The NPCs will buy a given product up to their base demand for it at as cheap a price as possible.

Completely unrealistic and unplayable. This is called "perfect market model". You ommited transaction costs completely which are actually the most important factor in all economy games. You made it so customer from the depths of amazon jungle will fly to Paris to buy a box of matches. In such world can exist only one big supermarket per planet with all goods available, everyone else will go bankrupt.


Want to make a mini coding challenge? The theme would be "a shop/company that sell goods, you try to get the most money in 50 turns/years", singleplayer minigame. It would let you understand the basic techniques how to skip corners in economy games (and of course I would be deligheted to compete along you )

[Topazan]

Where are any of you getting the idea that I'm trying to make a perfect simulation?  You're getting overly caught up on complementary/substitute goods, which I only mentioned as a possibility in passing.  It is not the focus of what I'm trying to achieve here.

It does not matter if it is focus or not. The very fact that you considered it even for a tiny litle second means you went way, way, way overboard.

Why?  

Complementary goods:

Code: [Select]

$arrowDemand = $arrowBaseDemand + $bowSales;Substitute goods:

Code: [Select]

$armourDemand = $armourBaseDemand - $chainmailSales - $platemailSales - $scaleMailSales;

Never mind what other games have done

You will regret it in the end You are trying to build a space rocket without the basic knowledge of algebra. The outcome is easy to predict
 

The NPCs will buy a given product up to their base demand for it at as cheap a price as possible.

Completely unrealistic and unplayable. This is called "perfect market model". You ommited transaction costs completely which are actually the most important factor in all economy games. You made it so customer from the depths of amazon jungle will fly to Paris to buy a box of matches. In such world can exist only one big supermarket per planet with all goods available, everyone else will go bankrupt.

Well, look at this. A tangible, solvable problem.  That's what I like to see.  Problems like these are why a complicated system is sometimes necessary.

I think the main thing is that you're assuming that players will have unlimited productive capability.  The problem only exists if the players don't need to spend time or resources to expand their businesses.  The beauty of this system is that demand stays just ahead of productivity.  It ensures that there's always a new niche to fill.  There are also numerous ways the system could be slightly tweaked to solve this problem.

Want to make a mini coding challenge? The theme would be "a shop/company that sell goods, you try to get the most money in 50 turns/years", singleplayer minigame. It would let you understand the basic techniques how to skip corners in economy games (and of course I would be deligheted to compete along you )

Maybe, I'll think about it this weekend.

But tell me, if you know so much about economic games, where's your solution?    The only think you've given me is "arbitrary demand", and you didn't respond when I pointed out the numerous problems with that kind of system.  Add your problem of transaction costs to that list.

[Chris]

Where are any of you getting the idea that I'm trying to make a perfect simulation?  You're getting overly caught up on complementary/substitute goods, which I only mentioned as a possibility in passing.  It is not the focus of what I'm trying to achieve here.

It does not matter if it is focus or not. The very fact that you considered it even for a tiny litle second means you went way, way, way overboard.

Why?  

Such system was never implemented in even the most hardcore simulations. In games made for 100% pure economy maniacs, made by people who make only economy simulations. They didn't include it. You, while making an economy subsystem that will be a small part of your game, are planning to outdo the biggest complex monster that ever existed. That's what I call getting overboard

But tell me, if you know so much about economic games, where's your solution?

That's the point (and that's the point saljutin was making too), such wide knowledge on the subject does not give me any real advantage over you It is irrelevant, more theorethical knowledge here (I'm not talking about theorethic knowledge being useless in other cases, only about making economy models in games case) is of very low importance. In the end you are making your own game with your own rules and players will accept these. That's why I said you should play other games first (old and primitive ones, that's where the solution is). As for being more concrete I gave you one solution, a shop radiating area of influence and gathering customers from surroundings, that's how probably all or at least huge majority of economy games do it. Build a "shop" in Pizza Tycoon, you will have access to X "people" from surrounding buildings of a specific demographic, you can expand by building more shops or by increasing area of influence/stregth of influence via marketing. In Detroit you select a region and build office levels (progressive cost), each office level enlarge the penetration of the region it is located in. Capitalism I & II, pure shop influence radiation (probably with losing power at larger distances from the shop).

The only think you've given me is "arbitrary demand"

It's not "ONLY" it's "THE" solution

[saljutin]

I think you should try this
http://www.kapilands.com/
pretty good but it does not have "on the edge" effect

one more to help u think - if u want to create REAL economy simulator: people who buy your goods need to GAIN money from somewhere -> from working for you or some other player, so how are you going to make that

[Topazan]

Where are any of you getting the idea that I'm trying to make a perfect simulation?  You're getting overly caught up on complementary/substitute goods, which I only mentioned as a possibility in passing.  It is not the focus of what I'm trying to achieve here.

It does not matter if it is focus or not. The very fact that you considered it even for a tiny litle second means you went way, way, way overboard.

Why?  

Such system was never implemented in even the most hardcore simulations. In games made for 100% pure economy maniacs, made by people who make only economy simulations. They didn't include it. You, while making an economy subsystem that will be a small part of your game, are planning to outdo the biggest complex monster that ever existed. That's what I call getting overboard

Great, maybe I'll get a Nobel Prize for those two lines of code. 
 

As for being more concrete I gave you one solution, a shop radiating area of influence and gathering customers from surroundings, that's how probably all or at least huge majority of economy games do it. Build a "shop" in Pizza Tycoon, you will have access to X "people" from surrounding buildings of a specific demographic, you can expand by building more shops or by increasing area of influence/stregth of influence via marketing. In Detroit you select a region and build office levels (progressive cost), each office level enlarge the penetration of the region it is located in. Capitalism I & II, pure shop influence radiation (probably with losing power at larger distances from the shop).

This is supposed to be simpler than my idea, is it?  Because I don't think it is.

Maybe I should have been more clear at the beginning, but this is a multiplayer game I'm talking about.  The systems that work in single player games may not work as well here.  What you describe becomes even more complicated when you factor in competition between players.  I'm not planning on making the market instanced, as that defeats the whole point.  The other thing you need to factor in is the player-player economy.

With that in mind, all five of the problems I pointed out with arbitrary demand would exist in a multiplayer version of this system.

The only think you've given me is "arbitrary demand"

It's not "ONLY" it's "THE" solution

"THE solution" doesn't work. 

And do you realize you've been going on about how complicated and overboard my idea is, then you suggest a system that's even moreso?

I think you should try this
http://www.kapilands.com/
pretty good but it does not have "on the edge" effect 

  I have played kapilands actually, it was pretty good, rather similar to simunomics.  I don't remember what was unique about the retail system though, maybe you could refresh my memory.  Also, what's the "on the edge effect"?

one more to help u think - if u want to create REAL economy simulator: people who buy your goods need to GAIN money from somewhere -> from working for you or some other player, so how are you going to make that

That's a different topic, but the short answer is "raw material sales".   

[saljutin]

my friend told me there is some "deep" economical game:
http://virtonomics.com/

On the edge - you can go bankrupt if u use bad decisions...in kapilands you must be stupid to go bankrupt

[Chris]

And do you realize you've been going on about how complicated and overboard my idea is, then you suggest a system that's even moreso?

I meant something like this:
- you build a shop you are getting 10 customers per day
- you click upgrade shop button which cost 100 gold (each next upgrade cost more) you get +1 customer per day
- if you reduce price by 50% of average price of other players you get 3 times more customers
- if you run marketing campaign you get +10% customers (cummulative with other bonuses)

[Topazan]

And do you realize you've been going on about how complicated and overboard my idea is, then you suggest a system that's even moreso?

I meant something like this:
- you build a shop you are getting 10 customers per day
- you click upgrade shop button which cost 100 gold (each next upgrade cost more) you get +1 customer per day
- if you reduce price by 50% of average price of other players you get 3 times more customers
- if you run marketing campaign you get +10% customers (cummulative with other bonuses)

Ok, so similar to my idea, key differences being:
1.  Instanced- each player has their own NPC market.
2.  Things other than sales determine population.

I'm sure it would work great in a single player game, but for a multiplayer game it's just not interactive enough.

[Chris]

I don't see it any less interactive than the NPC idea, at least the way I understand what you wrote. In my example the interaction is by the global price affected by all players, in yours NPCs move to the shop with lowest price - that's the same level of interaction to me...

Also, I still have no clue how you intend to prevent "one shop with the lowest price to claim 100% NPC customers and make everyone else out of business".

[Topazan]

Slight revision:  NPCs have a need level for each item, and a certain amount of money, which for the sake of this discussion may be arbitrarily scaled with population*.  They will buy the cheapest goods they need until they run out of money or all their needs are filled, at which point the population increases and the needs are increased.

Help me understand.  You say:

if you reduce price by 50% of average price of other players you get 3 times more customers

Is that an example or is that the one and only pricing rule?  If the latter, why not just charge 1billion gp for everything?  In either case, what would you say guides the players in their pricing decisions?

Perhaps I was wrong to say that it doesn't have enough interaction.  But I'm not clear how the market will determine the price.  Couldn't you control the price by making a lot of dummy accounts to bring the average up or down?

The interaction with what I said:
-There will be an incentive to provide the less popular products.
-The amount of money entering the system will be measurable, and players will compete with those in the same industry for a share.
-Players will have a symbiotic relation with those in other industries .

As to the problem you mention:

1.  As I said, players will not have unlimited production capability.  A single player can't monopolize the entire market because they just won't have enough products.
2.  By its very nature, the market will continue to expand.  A single player selling their good cheaply will increase the price of other goods the NPCs need.
3.  The chance to make a profit in the player-player market still exists.  That megastore will still need materials to produce its products.

Sometimes I wish this forum had subthreads, so we could talk about different models in depth.

*The risk of inflation is one of the flaws with this system.  One quick fix is to cap production per player, as I said.