A.S. Repeat Trade?

[Hermann the Lombard]

This is loosely a question for Adam Smith. At SG(2)'s suggestion, I'm trying an ICS Trade Sleaze, and I'm starting to generate caravans at a great rate. I had read the "Repeated Commodity Trade Strategy" and thought I'd make that a part of my strategy for the game. So the key to that is identifying AI cities whose demands do not change (or go in parens) when goods are delivered. According to the A.S. document, 1/3 to 1/4 of cities have this behavior.

Here's the problem: I've looked at 34 AI cities. Of those, 30 have changed demands since I first looked, and 2 of the remaining 4 now have demands in (parens). The remaining two have never received a caravan from me. Statistically, it would appear that there are NO cities with this property among the AI. Can this be so (in 2.42?)

Meanwhile, my capital seems to have this property, so I'm making out like a bandit locally, but enduring the usual PITA of having caravans arrive at cities that no longer want my cargoes (or more annoying, a galleon arrives with two cargos for (say) Moscow, and delivery of the first cancels the demand for the second. We all know how bad the trade interface is, and this just makes it more so.

-- Hermann

[George Garrett]

Good news Hermann! I'm currently playing on the large word map,(Chieftain in 2.42), and San Francisco(size17) is taking my caravans from near and far, without changing demands, or going brackets on any commodity...moreover, it pays off in the hundreds for each consignment - as much as 700+gold. All in all, I've found "America" a jolly good civ to trade with (No wonder they have a trade deficit!)
[This message has been edited by George Garrett (edited March 08, 2001).]

[Adam Smith]

Hermann:

Three points to mention here.

First, how long has it been since you first looked at the AI cities? There are global factors in the game which change the commodites available (eg, industrializaiton makes cities supply oil), so the same commodities do not repeat indefinitely.

Second, in part because commodities do change, and in part because big science bonuses can get wasted, it is better to have lower payoff routes that you can complete consistently every one or two or three turns, rather than a huge payoff route that you can complete every 15 turns or so.

Third, even if some of the cities have a commodity or two in parens, all is not lost. Try sending pair of caravans to such cities, one with a commodity that is in parens and one with a commodity that is demanded but is not in parens. You can usually make the parens flip from one commodity to the other and back again, making a repeated commodity city out of one that has parens.

I hope this helps. Get out there and go trading.

------------------
Old posters never die.
They justfadeaway

[This message has been edited by Adam Smith (edited March 08, 2001).]

[Hermann the Lombard]

<<
First, how long has it been since you first looked at the AI cities? There are global factors in the game which change the commodites available (eg, industrializaiton makes cities supply oil), so the same commodities do not repeat indefinitely.
>>

Understood. It seems like they changed on a short time frame, but in fact I didn't note the year when I first checked (it was around the time I got MPE in Caravel times; I'm up to Galleons now. Looking at patterns, it seems like certain cities get Dye or Hides as every 4th or 5th demand.

<<
Second, in part because commodities do change, and in part because big science bonuses can get wasted, it is better to have lower payoff routes that you can complete consistently every one or two or three turns, rather than a huge payoff route that you can complete every 15 turns or so.
>>

Part of the problem is the large map. It takes many, many turns for a caravan to reach an AI city. The exception has been Japan, but they are on the same continent (not that I can reach them overland). Still, I have three "+10" routes from my capital to Kyoto. The Vikings are also on the same continent, and sneak attack too often for profitable trading. Of course, now I'm poised to settle *their* bacon, but that's a digression from trade. Fun, though. The nearest AI not on my continent is the Sioux. Most of their cities are only four or five turns from my capital by galleon and overland. (You'd think these AI know about ironclads! Damned few coastal cities!) The nearest Greek cities are 6 or 7 turns out (three from the end of my land mass, so once railroads are in, they'll be 3 turns from my capital). The Russians are 2 or 3 turns farther away (8-10) and the Spanish are beyond that, just about exactly around the world...which puts them right across the water from "Alaska" one of my major colonies. So I will try to have "Alaska" trade with the Spanish and Russians while the homeland trades with the Sioux and the Greeks. Maybe the heartland with the Sioux and the west with the Greeks and Russians.

As for big science payoffs, I'm getting tech every two turns *without* trade, thanks to the SSC at the heart of my ICS. So I'm not exactly hurting. I just didn't feel like I understood "repeated commodities."

<<
Third, even if some of the cities have a commodity or two in parens, all is not lost. Try sending pair of caravans to such cities, one with a commodity that is in parens and one with a commodity that is demanded but is not in parens. You can usually make the parens flip from one commodity to the other and back again, making a repeated commodity city out of one that has parens.
>>

Aha! I'll try that, with concentration on the Sioux and the Greeks (and Spain from Alaska).

Thanks for the help!

-- Ed

[Scouse Gits]

Have you thought of reducing your transit times by using a ship chain?

------------------
Scouse Git[1]

"Staring at your screen in horror and disbelief when you open a saved game is one of the fun things of a succession game " - Hueij
"The Great Library must be built!"
"A short cut has to be challenging,
were it not so it would be 'the way'." - Paul Craven

[La Fayette]

HtL,
Welcome to the club of greedy 'trademongers'!
1) Repeated commodity
I don't think any definite pattern has been discovered yet (though my own experience is that it happens mainly with Hides). I would say "if you encounter it, use it".
2) 2 continents
There is a long and interesting discussion between Xin Yu and Adam Smith on that topic (that you can find in the GL). The conclusion is that in the very late game '2 continents' is more performing than 'repeated commodity'. My own experience is that '2 continents' also can be very performing much earlier (if you manage to "rule the seas" or at least achieve a reasonable degree of safety along your main trading line).
3) Chaining ships
Of course sound advice (hello SG!). Difficult to have in use early (main question: "shall I build many caravans and few ships, or the reverse?")
4) Internal trade
Some strong players (like Ming himself) advocate that speed and low risks make it a good choice, especially for early trade.
5) Conclusion (in short)
Foreign trade is risky but can be VERY rewarding.
If you manage to rule the seas, most risks disappear.
If you manage to build a strong and safe '2 continents' trade route, you have won (but 'repeated commodity' can be very useful meanwhile).

------------------
aux bords mystérieux du monde occidental

[Xin Yu]

quote: Originally posted by Adam Smith on 03-08-2001 07:21 PM Third, even if some of the cities have a commodity or two in parens, all is not lost. Try sending pair of caravans to such cities, one with a commodity that is in parens and one with a commodity that is demanded but is not in parens. You can usually make the parens flip from one commodity to the other and back again, making a repeated commodity city out of one that has parens.

Generally, the new trade route needs to be more valuable than the old ones for that to happen (the city must already have 3 trade routes, and it will re-calculate supply and demand when a new trade route comes and replaces one of the 3). In two continental trading, I have full control over the trade routes, thus I can blow up the parens of any particular supply and demand of a city. I lower the city's existing trade routes' values by hiring all of the city's trading partner cities' citizens as specialists (temporarily, they can come back to work after the trading in the same turn). Then I send a couple of caravans from non-partner cities to that city, making sure that the non-partner cities do not demand or supply the particular commodity I want to deal with. The new trade routes will almost surely be established and the supply and demand for the city will be renewed. You can also do it to a foreign city, as far as you keep a record on which cities have generated the trade routes.

[Adam Smith]

I am not sure that the new trade route has to be more valuable than the old one, and I know that the destination city does not have to have three trade routes to make this work.
Suppose the destination city reads Silver, Wine, (Cloth)
Send a cloth and a wine caravan.
Land the wine first and the destination city reads Silver, (Wine), Cloth
Then land the cloth caravan and it reads Silver, Wine, (Cloth) again.
This does not work for every city, but it happens often enough to turn a tidy profit. And, unlike the Two Continents Strategy, you do not have to travel half way around the world to turn a profit.

------------------
Old posters never die.
They justfadeaway

[Edward]

Xin Yu,

I don't fully understand trade routes, but you postulate something I had suspected.

If a destination city (say City A) has three trade routes and you send in a new caravan to try and set up a fourth route, that the fourth route will bump off and replace one of City A's current three if and only if that new trade route would generate more permanent trade arrows per turn. A city only get permanent trade arrows from its three best routes. I think everyone agrees with this.

Then you say something that I've been too lazy to prove. You imply that if City A has one of its demands in parentheses, the reason is that that demand is being met by one of its permanent trade routes. You're saying that supplies and demands being tied up are not mysterious - but are always associated with a given city's three permanent trade routes.

This leads to the breakthrough that if a new trade route would give more permanent trade arrows than City A's worst permanent route, it bumps off the worst route which then releases a tied up demand if (and only if) that worst route was meeting a demand.

Now - and here's your sheerly brilliant step - you realize that you yourself can artificially determine the "value" (as seen by the game in terms of permanent arrows at the instant the new caravan arrives) of City A's trade routes by cleverly shifting workers on or off of trade squares in cities that City A has permanent routes with. With this "power" you can force an old demand-meeting route to be replaced by the new one (demand-meeting or not). Voila! That Babylonian capital is again demanding Gold from your Super Science City! Or perhaps the one and only AI demanding Copper is again open for business! Well, I guess you only have total control, as you note, over your own cities - but this is still good. This could really change trading (for those who don't mind a little micromanagement and record keeping).

(Of course techs or passing time can also change demands and thus free up routes as well.)

[Adam Smith]

The theory seems to be that demand will be freed up in a destination city if and only if an existing trade route is replaced by a more profitable one. If that is the case, how can there be three commodities delivered to a destination city, each commodity meeting a demand, and still have the destination city demand three commodities, in most cases the same three commodities as it did to begin with, with no other trade routes than the three that were established?? The theory does not appear to fit the data in this case.

------------------
Old posters never die.
They justfadeaway

[This message has been edited by Adam Smith (edited March 09, 2001).]

[Edward]

quote: Originally posted by Adam Smith on 03-09-2001 02:11 PM I am not sure that the new trade route has to be more valuable than the old one, and I know that the destination city does not have to have three trade routes to make this work. Suppose the destination city reads Silver, Wine, (Cloth) Send a cloth and a wine caravan. Land the wine first and the destination city reads Silver, (Wine), Cloth Then land the cloth caravan and it reads Silver, Wine, (Cloth) again. This does not work for every city, but it happens often enough to turn a tidy profit. And, unlike the Two Continents Strategy, you do not have to travel half way around the world to turn a profit.

Perhaps in your example the city (City A) DID have 3 routes. It's just that two of its best routes were routes that IT established (and therefore are tying up two of its supplies which generally can't be seen for an opponent's city).

Your new Wine route comes in and is better than City A's worst route - the one where it's getting Cloth. Now City A's three best routes are 1. Wine from you to City A; 2. something from City A to some other city; 3. yet another thing from city A to some other city. The demanded Wine and two of City A's supplies are in parentheses, but Cloth is now an unmet demand. (Note that City A didn't even need to be demanding Wine to free up the Cloth route. Any caravan coming in that would give City A more permanent trade arrows than the Cloth route was giving would have bumped the Cloth route out of the top three - thus freeing it up as a demand.)

Now your new Cloth caravan comes in. It's better than the your recent Wine route. Now City A's three best routes are 1. Cloth from you to City A; 2. something from City A to some other city; 3. yet another thing from city A to some other city. The demanded Cloth and two of City A's supplies are in parentheses, but Wine is now an unmet demand again.

Next time you deliver a caravan to City A, implement Xin Yu's strategy by moving all your workers in your city supplying Cloth off of their trade squares. Your new caravan is going to look pretty good to City A. So good in fact that City A will ditch that "terrible" Cloth route in favor of your new route. Voila! City A wants Cloth again. Now reassign workers in your Cloth supplying city and get that next caravan ready!


(Note: When you said that it "does not work for every city," that's because that one "visible" permanent route needn't be City A's worst route. If it wasn't, then your Wine caravan would have met City A's demand, but left City A with: Silver, (Wine), (Cloth). In that case either City A had less than three routes, or City A's worst route was one of the two "invisible" routes where it's supplying the goods.)

[Edward]

Adam Smith,

You're right. Three delivered caravans (each with a different demanded commodity) not tying up any demands in destination City A sounds unlikely under this theory. The only way for it to happen is if City A made three routes early on. All three of its best routes are ones in which it is doing the supplying. Apparently your three caravans weren't good enough to knock out any of City A's three permanent routes.

Xin Yu's theory says that a foriegn city with mysterious repeated demands is actually just a city with one or more really good permanent routes in which it's the supplier.

[Xin Yu]

Recently I noticed that a permanent trade supply/demand commodity oftenly is the 'hide'. This makes me think about the following:

Maybe the commodities are 'paired'? Hide is paired with Uranium, for example. When a supply/demand is tied up, the computer searches for the commodity's buddy and frees that commodity up. At the beginning, since Uranium is not available, Hide is sometimes paired with itself, and will be permanently available.

[Blaupanzer]

Worthy of note here is that the first delivery (meant to free the commodity in parens) can change the demand pattern for the receiving city to something other than the item formerly in parens. In our example the demands were Silver, Wine, and (Cloth). Delivering the Wine can cause the demands to become Silver, Copper, (Wine). I say stick with Ming's internal trading strategy. Doesn't seem worth it to try to deliver demanded commodities to faraway places until you have flight (specifically, airports). Then you can send the freight to a "scout city" (another thread calls island scout cities "Cubas"), and go in from there.

Ship chains do work, but require serious management, tactically and strategically. Tactically, you have to keep them all in motion, in the proper timing and get them back to the proper intercept points. Strategically, you have to keep the chain pointed at the appropriate target and properly protected with ships and treaties. A lot of work and resources to move caravans that can move on their own to nearer targets.

[Marquis de Sodaq]

Has anybody looked into how demand shifts around the map? If you've paid good attention, certain patches of the map (and thus the cities within) demand certain commodities. When they change, another area of the map demands the commodity. Clearly, demand is linked to tech/time as well - thus parts of the middle game during which everybody and his uncle demands dye.

Is it time to cheat up a world with evenly spread cities and test demand across time? I'd suspect that some patterns could be found:

1. predictable time of demand change (game year, population threshold, or tech)
2. weighted importance of commodities in different conditions

Many posters grumble about the demand vanishing just before delivery. If we know how many turns to expect demand to remain, or where it might shift next, planning deliveries could be made less stressful. And more profitable.

Trying to exhaustively account for just these 3 variables would give me a headache. However, I'd be willing to test out a generic map with uniform cities to start. Is there any interest in this info?

------------------
"There is no fortress impregnable to an ass laden with gold."
-Philip of Macedon