What all is involved in simulating an economy for a game that involves trading? | Hacker News Reader

What all is involved in simulating an economy for a game that involves trading?

old.reddit.com

30 pointsyamrzou10 comments

The interesting bit of simulating an economy for a game is not the tradeoff of realism vs computer resources, but realism vs fun.

A game where AI actors are making prices come very close to costs, as they make the most profitable trades, and demand makes actors switch to alternative goods whenever their preference is priced too high, will turn into a spreadsheet. Not unlike in real life, the work will be just identifying where you have left actual inefficiencies, and end up with an algorithm that exploits it. Not so much fun.

Most games that involve trading leave giant holes in the economic model on purpose: The parts of the economy that they copy are the parts that make the game more engaging. It's not that in, say, Railroad Tycoon 2 you couldn't build a more realistic economic model: It's just that the more realistic model would be less fun to play with. Same as in boardgames. Many people love playing Power Grid, and like its market mechanic, but it's 100% attached to a very unrealistic turn order for purchases, which attempts to balance out who is receiving more income. It builds a tradeoff between being the leader or falling behind, which would in no way occur in a more realistic model. Then the winners would just keep winning no matter what, and some people might as well accept defeat in the first turn of the game. This is not just speculation: People in the game industry have playtested games with a stronger economic model all the time, and the more realistic the model, the harder it is to make the game fun.

So the real question is, what parts of an economic model are necessary to make the game realistic enough to be engaging, and which parts harm your game?

One thing I distinctly remember about one of the Railroad Tycoon games is that it was almost always a better bet to invest in non-railroad businesses, and use the profits from that to build the bare-minimum railroad required to get the top score.

Don't sleep on the fact that, 8 years ago someone pointed to a piece of software someone wrote - and that the author of that software replied 20 minutes ago (as of this comment)!

https://old.reddit.com/r/gamedev/comments/24fjp9/what_all_is...

The author is active on here, that's probably why :)

Does modern ai simplify the simulation of an economy? As, ai is much more capable of learning complex games now than it was 8 years ago.

Economists could publish their models and anyone could adjust the assumptions to test the stability of the conclusion. Anyone could watch the agents go about their days and see if the decisions relate to reality. Almost allowing economists to run experiments.

The simulations would probably need to be narrow rather than the entire economy.

Is this worth pursuing or is it intractable?

I keep looking for an excuse to apply this research paper.

It details how to have actors in the economy react to supply and demand by each having their own belief of what an asset's price is "supposed" to be. The gap between their individual beliefs and actual market value drives price movements.

Emergent Economies for Role Playing Games, Jonathon Doran and Ian Parberry,

https://web.archive.org/web/20181222160146/https://larc.unt....

You can simulate a wealth-backed currency (defined in terms of in-game resources):

https://github.com/pjkundert/ownercredit.git

This has been used in games; however (just like real wealth-backed currencies), it doesn’t respond well to games that simulate the destruction of wealth (such as war games).

https://www.true-world.com

(2014)

would be interesting what would happen if something like eth existing within a game so that users would be able to form complex financial institutions beyond the scope of the game design team.

havn't thought much about this, but even games like Eve Online that claim to have an interesting economy are still very primitive.

Reply on news.ycombinator.com