Internal trade routes - abstracted

[tlaurila]

I've played a bit of DoNE and more of RaR now, and read some of the discussions about development here. It seems that carting goods around is a critical issue, both for human playability and for AI function. So I had this idea that just maybe could find resonance:

Why automating moving the wagons (and ship) is the thing, to apparently quite a frustration, instead of altering the concept of the wagons themselves to something that would work more easily?

So the idea is to automate intra-continental trade in a manner that draws from the custom house: Allow wares to move automatically (and instantly) between cities, as long as investment is made to infrastructure. So the player gives for a city limits for the stored amount of a ware that trigger import and export. A wagon unit is attached to the city that wants to import, giving the city certain "pulling power". Then the goods move automatically from export to import city, spending an amount of "pulling power" that equals distance between cities times the amount of goods moved.

Connections and distances between cities need to be determined, ideally both by land and by sea, but probably good enough to start just by land.

A mechanism for "attaching" transports to cities need to be come up with, but this could take any form that allows to "invest to transport infra". One possibility could be that empty wagons in the city that have full movement left at end of turn give the city "pulling power" according to cargo slots times movement points (likely divided by something like 4 for balance).

Enemies around a city should cut the city off the import/export grid.

The change would slightly reduce the potential for piracy/raiding, but I don't think by much. The Europe route is when that happens anyway.

I know it won't be an easy implementation, but maybe not impossible? I know my way around generic coding, and have some very limited experience from modding civ4 years ago (I have forgotten most, though), but nothing at this depth. Unfortunately currently I have very limited time to put to hobbies like the game. Anyways, if people think the idea could fit some of the existing mod projects, maybe I could try and start to learn...

Has somebody tried something of this sort?

 

[Nightinggale]

You aren't the first to propose such a system and there are reasons why it haven't been coded. One is the concept of the game playing itself rather than putting you in charge. The other is a question of implementation. How do you determine the distance between two colonies? If you mean the distance count you get from dragging the goto cursor around, then it's not an easy solution because it works on an instance of a unit class. This means it assumes an actual unit to be present in colony A and it wants to go to colony B when calculating the travel time and the question if it is by land or sea is determined by the unit data. In fact I think the pathfinder kind of ignores the issue of land/sea and tries to move on both. It doesn't really care for the reason why it can't move on a plot and will ignore it if it has water, mountains or something else, which makes the movement illegal for the unit in question.

The answer lies in better management of transport units, not avoiding using transport units.

 

[orlanth]

Well I'm a fan of the concept of PlotGroups, which AFAIK has been implemented in Medieval Conquests and is somewhat similar to what you describe, but without dummy transporter units.

One of the main benefits could be freeing the AI from getting badly stuck/crippled from the fact that it can't effectively plan to have x Yields in a given city to complete a unit/building/profession. Players who hate transport micromanagement (such as me ) could also benefit, and this is great for the Domestic Market sales feature in MC Perhaps Handicapinfos.xml settings could be used to let this feature be customized to be on or off for humans or AI as desired?

 

[tlaurila]

Thanks for the replys!

As for the "what is fun" philosophy, of course to each his own. Mine is that I want to build the economy machine, and then make the strategic decisions on where to invest the output: Build more of the machine, wage colonial wars, or ramp up for independence. Having to manually operate the machine, doing the same things over and over, is what I consider tedious micro-management.

Therefore, I generally prefer to manually move around the "strategic" goods: tools, weapons, stone, lumber. Manually moving the big ships from harbor hub(s) to europe is still ok - also allows to keep a check that the economy keeps on outputting as it should.

The vanilla-like produce-process-port transport chains could still be doable even by manually defining simple routes of one transport unit moving one yield from A to B. It's the added domestic market feature that blows the transport route needs to a mind-boggling number of source-destination pairs. The domestic market makes the custom house seem outright ironic: You can abstract-automate the simple one-sink shipping to Europe, but the branching network of domestic demand you have to manage with some kind of interactive, time-delayed, N-agent, traveling salesman-ish problem?

Didn't mean to rant, though, so far loving the domestic market on my large map game with 7 towns. So far fully automated transports work fine, but I suppose it is only because of two small-enough points:

- Total production and number of cities is low enough that I can make do with a single hub/port (storage space is sufficient). Products and excess produce is all imported there, then shipped to Europe manually or distributed to domestic markets by fully automated wagons. Thus the hub is the only town that both imports and exports yields.

- All cities are close enough that almost all trips are done in 1 turn, even the longest distance is just 2 turns.

Reading up on PlotGroup, it seems to already be sufficient for an abstracted continental trade routine. Distances between cities would be a nice thouch, but not must have - the custom house ignores it too.

Another option might be to allow the domestic market to source from all of your colonies instead of locally (maybe at, say 80% of profit)? Source-to-factory-to-Europe routes are simple enough to automate the simple way. Suppose this wouldn't help the AI though.

A related issue is warehouse space. It is rather unrealistic that one city can only hold so many warehouse buildings. Doubly so because this can be micro-managed by parking transport units and manually using their transport slots as storage. So some mechanism to keep on building warehouses to keep a central hub operating would really be nice.

 

[M07]

Hello tlaurila,

For your idea, I think about it before doing the trade route system and the main default for me is that you cannot intercept wagon trains on the land during wars.
Resupply is a key point in a war. For instance, if you succeed to isolate an enemy colony with soldiers, and with the new system your enemy send a lot of muskets to this "isolated" colony; AI could save his city by this kind of cheat.
I also agree with what Nightinggale said. Distances between cities is hard to handle. If you have two colonies one in north and the second one in south and both linked by a route. How do you know the road is still existing everywhere (costly algorithm to use), and how many time you should wait before having the goods?
It was for me more complex to answer to those question than doing the automatic trade route system in DoaNE.

Btw, I am interested by your point view, you have said that you played to DoaNE. Do you still think that we are doing micro-management with trade routes system we have implemented?
Thank you for your answer.