So, I've been thinking about mercantilism for awhile, and how to make it really gameable. I've read the Axioms about it, but I think my solution is going to be to abstract it to an even greater level. I don't think that having umpteen goods types is very valuable; players don't care whether they're salt merchants or glass merchants by default, and adding rules to differentiate between salt, glass, meat, grain, fish, fur, rare wood, common wood, ale, animals.... therein lies madness.
My suggestion looks like this:
Each City, based on market class (6 minus class? Some other number?) can support a certain number of TRADE ROUTES. In addition, it has a TRADE CAP, based on the number of families. This represents the maximum amount of profit that all trade routes combined can generate. XX (half??) of the TRADE CAP forms the RESERVE. The RESERVE is divided into shares and assigned to its trading partners based on their market class.
Example: Capital City, a Class 3 Market, can trade with three other urban settlements. It's currently trading with Beach City and Kakariko village. It has a population of XX families, so it has a TRADE CAP of 20,000 gp. It's RESERVE is divided between Beach City and Kakariko Village. Since Beach City is a class 1 and Kakariko Village is a class 5, Beach City gets 4 shares and Kakariko Village gets 1, for 8000 and 2000 respectively. A dozen major caravans from Beach City bring in almost 20,000 gp worth of goods, but only the first 18000 get sold (Beach City's share of the RESERVE plus the unreserved portion of the TRADE CAP.) The 12 caravans actively compete to try to avoid being the one who brought too many goods. (MECHANISM FOR COMPETITION YET UNWRITTEN. FIRST TO ARRIVE? LARGEST SIZE?) Meanwhile, the Kakariko wagon brings exactly 2000 gp worth of goods and sells all of them, since it has its share of the RESERVE. No matter how many goods the Beach City caravans bring, there will always be some people who need/want whatever goods are unique to Kakariko Village. (In this case, probably kitschy hand-carved trinkets, scary masks, weird ethnic food, and "fictional" stories about the False Hydra that lives in the well.)
To participate in a Trade Route, you need two things:
1. A caravan. (Guards, camels, barges, a single flying ship protected by magic skulls; whatever actually moves goods from one place to another safely.)
2. An initial investment (Determined... somehow...I'm still working on this bit.)
To create a WHOLLY NEW Trade Route, you need a third thing, which is
3. A map of the route. Since maps don't exist by default, PCs usually create a map by travelling along said route.
New trade routes are very profitatable, and then decline in value over time as the miraculous becomes merely novel and then common. (I'm thinking something like x4 gold for the first trip, x2 for the second, and then normal value afterwards.)
So, my goals here are to
1. Make the DM's life easier bookwise. Saying "This city should have five trade routes" and then drawing five lines to the five nearest major settlements seems very straightforward, in a way that generating and maintaining a large table of prices and demand modifiers for every city does not.
1B. Make the DM's life easier from a creative writing standpoint. It's easy to imagine a table of Vaugaries of Trade Routes. Bridges collapse, religious zealots burn your caravan from Sodom, nobles demand transport, a troublesome citystate tries to embargo you.
2. Allow for intuitive choices by PCs. If the main merchant-adventurer action is the discovery of routes to new cities, then the choice between "Find a route through Manybees Forest" and "Find a route over Murderpit the Death Swamp" has level of difficulty stamped all over it.
3. Allow for scaling and world-integration. Just as a party of fighters start out punching individual kobolds, progress to burning whole kobold villages, and finally end up as kings oppressing kobolds as a whole, a party of venturers starts out as freelance cartographers, progresses to Columbus-style explorer-conquerers, and finally end up as wagon emperors who command a vast spiderweb of trade.
Things I'm still working on:
I forgot trade routes connect TWO cities, so profit should be generated at both ends sometimes (always?) and I haven't even CONSIDERED the math for that.
I think maybe a small diversity of good types would make sense over just assuming that every city is selling wholly unique goods. Maybe a system of tags? The sliding scale of complexity to ease of simulation is cruel.