Economy - business simulation game.

Player has factories that make goods and shops that sell them. Sales volume is determined by the price, quality and competition. This is not "ideal market model" but "realistic market model" so even if competition sells goods that are better and cheaper you still can sell some because people will visit your shops that are near their house or becauce they never heard about competitor's product or because they want to have something different and unique than everyone else even if it is inferior.

So far no problem, we make a formula with border volumes easier/harder to reach and it works. But how to handle products of the same producer? These seem to be competing among themselves...

1) Competition sell family car at price X and they sell it in Y shops. You sell family car at price A in Z shops.
2) Then you put on sale family car at price B but in the same Z shops.

What percentage of market you get by introducing step 2)? You should lose some sales of the first car, but the competition should lose some sales too.

Even more interesting it become if we assume the same price and the same quality. Like you sell black cars, then you introduce red cars. You should get more sales for sure, you should also lose some sales of black car, but less than the gain from red car (overall you got a bigger share of the market by acquiring clients that want only red cars no matter what).


This could lead to degenerated cases when player wants to pump as many different cars as possible (Detroit) so it should be countred somehow, for example by the cost of developing a new model (even if it uses exactly the same technology than the old car).


So, how to simulate the sales of similar (but at least slightly different) goods of the same producer (and seller)? Taking into account competition as well of course.

Have the brand play a role. If you create different black cars, they will all share the same brand and will fight for the same market. If 2 players have 50% market share with 1 car each and one decides to create a clone of the same car, then they will both retain 50% share, except one will have 25% share for each car.

The player can create a new brand to get a higher market share, but will need to pay for marketing, which should balance it out. The game Capitalism had a similar scheme where you could choose to have a corporate-wide brand or a specific brand per product.

In Capitalism you always had 1 type of each goods class. Like only 1 TV type sold per shop. So, there was no problem of the same good being sold "twice" in the same shop.

I meant a game where you make one class of goods (cars only) and you sell them. Advertising is global, per company, not per product, so brand is perfectly equal in such case.

If 2 players have 50% market share with 1 car each and one decides to create a clone of the same car, then they will both retain 50% share, except one will have 25% share for each car.That's what I want to avoid. If you have second model of very similar good then you grab a bigger share of the market, because some people who do not like your first car will choose the second one instead of going to the competition. Of course, it should not be that with second car you get 2/3 of the market, since you still mainly compete with your first car, but you still should be able to claim a bit of competition's market share this way...

You could rebrand goods at the store in Capitalism if I remember correctly.
Add brand as a requirement in the consumer desirability matrix.

For example, you have 3 criteria : Color, GPS and Brand.

Color :
- Black 25%
- Red 50%
- Don't care 25%

GPS :
- GPS included 30%
- Don't care 70%

Brand :
- Brand A 25%
- Brand B 25%
- Don't care 50%

Let's say you produce Black GPS enabled cars. Out of 1000 consumers, 18 absolutely want a Black Brand A with GPS and as long as this type of car is provided, they will buy it. They may switch their allegiance if the price is too high, but as equal price, they will go for that specific car no matter the competition. 18 will also want the same car, but for Brand B. That leaves 36 consumers who wants a Black GPS enabled car that can sway toward one brand or another. When creating a new brand, you have to fight for the "Don't care" consumers until you build the brand. Assuming equal market share for the "Don't care" category, you end up with :
Brand A : 18 + (0.33 * 36) = 30
Brand B : 18 + (0.33 * 36) = 30
Brand C : 0 + (0.33 * 36) = 12

The result of creating the same car under a different brand results in a gain of 6 consumers. That bumps you from 50% market share to 58% market share. I left out the "Don't care" consumers for the other 2 categories, but the end result is the same. Creating the same car under a different brand will increase your market share, but will not increase it proportionally to the number of brands until you build consumer loyalty.

OK, maybe first sort out the easier case.

Computer
Marketing capabilities: 8
Model A (quality 5, price $500)

Player
Marketing capabilities: 4
Model B (quality 5, price $450)
Model C (quality 6, price $700)


How many of which model they sell? And how the sales would change if model C would be unavailable?

Good question! But I have no definite answer.

The way I would model it is with a multi dimensional array representing consumers and their desires. Each "turn", I would run the consumer desires simulation. This would check available prices and qualities and move them around the matrix. Since the player has poor marketing capabilities, some consumers would shift toward the computer. The higher quality of Model C would shift some consumers from the other 2 models to Model C.

Next would come the buying phase. Here, the price becomes a factor. Some consumers may want Model C, but since it's so expensive, they cannot afford it(yet). For each bucket in the matrix, a percentage of consumers buy the models based on the price. When in the Don't care category, they pick a car randomly from any available and attempt to buy it(price vs income). If they buy a car, they are removed from the bucket.

Next, the consumers reenter the available consumers by being randomly placed in a bucket.

How much of each model will be sold will depend on the current simulation state. Initially, an equal amount of consumers will fall in all buckets. If the simulation is built in a way that makes it hard to retain a full bucket when comparing an equal quality, then consumers will naturally flow toward the emptier buckets, which will be the ones with the lowest prices. For example, the quality category. If the quality 5 has 10 consumers and quality 6 has 20, then some consumers will flow toward quality 5(assuming they have an equal weight) so it becomes 13-17. This means that on the buying phase, quality 6 cars will have less potential consumers while quality 5 will pick them up. In the end, this should simulate offer and demand.

Take for example a Toyota and a Ferrari. Everyone wants a Ferrari, yet most are driving a Toyota. This is simulated by making consumers flock toward the higher quality cars. There will be lots of consumers, but none will be able to buy them. After a while, they will move down the quality ladder and end up in the Toyota range because the low quality bucket is emptier than the high quality one. At this point, they can afford them and are removed from the available consumers. A few lucky end up with a Ferrari, but most fall down and buy a Toyota.


One more thing, to make it work, the percentage of sales needs to be a small percentage of available consumers. The higher the purchase-to-consumer ratio is, the more random the simulation will become because consumers do not have enough turns to fall in an appropriate bucket.

The nice thing about this is it then becomes easy to simulate world trends. If owning a red car becomes hot stuff, then you can make consumers flow toward the red bucket and red cars will have more potential consumers each turn, resulting in more sales. If a brand gets bashed on the news, you can make consumers move out of that bucket until only the die-hard fans remain.

The way I would model it is with a multi dimensional array representing consumers and their desires.

I hate that one for several reasons. I really, really strive to avoid it at all cost
- It's complex, too complex
- Old games (actually, I think any games) never used such complex system. This was done simplier back then, still these games worked (Detroit, Old Timer, etc). How they did it?
- With enough detail such system will fail sooner or later because computer will be unable to process it (colour, shape, speed, fuel usage, safety, handling, glass transparency, radio set quality, James Bond drove the car in a movie, etc, etc)

I think the solution can not be discrete but fuzzy. Like there are no boxes with people that are very poor, poor, below average, average, above average, almost rich, rich, but a fuzzy container with people being in some state between poor and rich.


Maybe forget about the simulation part for now and focus on the game part. It does not have to be realistic or accurate, it only has to appear realistic. And it must be playable and fun.

Maybe think what values we want the model to return? What the player would find realistic and what would not cause imbalance in the game?
If there is model A and B then I would say computer gets 66% of market and player 33%. But since the player made the price lower he should get the 33% bumped up to 40%. Right?
Now player introduces model C. It's better quality but more expensive... So, some clients should dump model A and B, because they are richer and want the better quality. Still not the less rich because the price is still high. So like more richer only Hmmmm, I have no clue what kind of sales change I would find realistic here as a player... Well, I would find it strange if suddenly 50% of market switched to model C. That would be too much. But if it steal 10% from A, I would accept, not sure if I would accept if it steal 10% from *my* B through Sid Meyer was talking about this kind of illogical thinking of players in some podcast too I find it realistic that my ultra quality and expensive C model steals clients of this pathetic model A while I would find it ridiculous that my fine B with its awesome pricing got reduced just because it is a bit lower quality that this outrageous expensive C model!

What are your thoughts? What changes in sales you would expect as a player?

It would depend on the context. If consumers care about quality in that region, then it should be higher. If consumers are poor, it should be lower. No idea what would feel realistic. As long as I have a sense of control, I guess it would feel realistic.

OK, if everything or almost everything feels realistic in player's eyes, what do we want as designers? What would make the game more fun? What we want the player to do? What are degenerated cases we want to avoid?


As a designer & coder I would prefer to:
- Make separate classes of models like family car, luxury car, sport car. These would be handled completely independent, no substitutes at all. Like completely different goods just named all as "cars" for mood only.
- The easiest would be to handle whole marketing per producer not per product. Like you get 100 base customers per marketing level (computer's marketing level ignored for now), then the actual number of clients is modified by price and quality. Models B & C compete among themselves then. Of course done separately for each product class as above (so E & F which were produced as luxury cars would not affect it at all).
- Still, this means that one model within the same producer would be perfect and would use up all available customers... And there would be no point making different priced model (which maybe is better from gameplay point of view?) and you could not get rid of obsolete cars unless you reduce the price a lot (in which case obsolete cars would eat up all customers and new one would get less/none). Hmmmm, this does not sound very good. I think the car preference should be taken into account somehow, like some customers want the outdated one because it is cheaper. And I think if you have 2 similar cars you should get at least a bit more customers BUT I don't want it to turn into a feast of making dozens of similar (or even identical) cars with different price. Maybe something like only 2 models are considered no matter how many you have, one as a cheap and one as expensive?
- I want to avoid the Detroit mechanic where pumping new cars each turn was benefitial because of the way it calculated sales (each model seemed to generate some (few) customers that do not compete with your other models, at least that's how it felt). But on the other hand I liked the way Detroit handled old cars, it's as if some clients got used to them and were buying these even if new one become available and the price difference wasn't too big.