How Supply and Demand Lists Are Determined

[samson]

Just did a quick test. Continent size does seem to be a determinant of Spice Demand. Thanks, Solo and DrFell, you may have found the missing factor. Let me do a bit more testing and see if this is it.

[solo]

Wildcards eliminated some of the observations first posted here.

Further investigation indicates spice on the largest continent, often clustered close together, with excellent access to water.

[samson]

Well, it's pretty tough to get a demand for Spice under testing conditions. It has a pretty low demand quotient whatever you do. But it does indeed seem that the strongest factor is continent size. And you need a darn big one, too. I've posted a tentative DQ formula for Spice above, let me know if it works in your cases.

I've also found a correction to the Silk demand formula in the process of testing for Spice. A really strange bonus based on City#. If this is an indication of the kinds of subtle modifiers at work, we may never get this whole thing nailed down.

[solo]

Samson,

The adjustment to silver also helped another Roman city, Antium, pass muster with the commodity formulas. The next city, Cumae, also had spice early at size 1. Same continent (#13) and also a seaport.

Two other problems with Cumae(59,51), wildcards (salt,gems), (Pottery and iron working known):

forest: 5 + pheasant = 9
mountain: gold = 4
ocean: 4 + whale = 8
hills: wine = 4
jungle =1
grass = 3
plains = 4

1150 hides,salt,coal dye,gems,wool size 1 techs: 17
600 (coal),salt,hides dye,gems,cloth size 3 techs: 23

600 BC checks out okay, but for 1150 I am getting:

hides,salt,silk dye,gems,cloth

On the supply side, I'm suspecting that continent 13 may not be acting as a multiplier for coal, since if this step is left out, coal will show up in slot 3 as required in 1150 and still be strong enough for slot #1 in 600, leaving silk under the wildcard both times.

On the demand side, it is a very close call with cloth inching out wool by a score of 36 to 35, so this city may provide some help in tweaking the cloth formula for when techs are not below 10. Gems cover beads on both of these dates.

Just saw your latest, and will check this out. Problem may not be with coal. Could be silk. I'll check out the new spice deamnd formula.

[samson]

You know, continent size is not an easy thing to determine. In my testing it wasn't too bad because I used edited maps with continents of 25, 50, 100, 200, 500, 1000 squares. But in a game, nobody's going to count up all those squares!

[solo]

Nope on silk, as I was dealing with supplies.

However, that latest modifier for silk demand is hard to believe (continent #1 AND city#/2 is ODD!!!!!!!!!).

There must be more to spice demand, which sure does come and go in mysterious ways. I was starting to think that it may only come around to visit at the EVEN city sizes.

In spite of this weirdness, the formulas are holding up pretty well so far in my tests, and perhaps very good approximations are the best we can ever hope for.

I wouldn't expect too many exceptions to your general rules, because these would add a lot of programming overhead in making the commodity calculations. In addition, we can never rule out the occasional spin of the dice, too, just to put some more "spice" into these mysteries.

[DrFell]

Having a big continent for spice demand might make sense. After all, wasn't the spice trade in Asia quite a lucrative business in ancient times? (I might be wrong though ). Basing supply/demand on continent numbers is odd though, although higher numbers are more likely to be out of the way islands, so I suppose that also makes some sense. But basing it on city numbers seems inexplicable...

[samson]

Solo,

Another good catch on Coal. The supply formula did indeed have an error in the Continent# bonus. A transcription error from my testing notes to my formula writeup. The bonus should be x 3/2 (50%) not 2x. That should fix up things in Cumae.

Re: Spice. I. too, noticed that I could only get Spice demand to appear in coastal cities, those with lots of Ocean. But I believe this is due not to a dependency of Spice on Ocean, but to the fact that Ocean is demand-neutral terrain. Ocean does not figuire in any demand formulas. So more Ocean reduces the DQs of other commodities, allowing Spice to make the list. Spice's DQ is very low, so it is only in these coastal cities with few commodities having any demand that it appears.
And yes, there's probably more to the formula. We need to find cases where it doesn't predict correctly.

DrFell & Solo,

The City# in the Silk demand formula is weird and disheartening. Like continent size, it is not easy to determine and thus reduces the utility of the formulas as a whole. However, it is not entirely unexpected. After all, City# is a factor in the solo-cycle which is where this all got started.

[samson]

Update: another mod to the Wool Supply formula based on another geographic concept: Polar Circles.

I think I understand now what happens when there aren't enough commodities with non-zero SQs or DQS to fill out a list. It simply takes the highest numbered commodity (Uranium=15, Hides=0) with a 0 quotient. At the start of the game this is would be Oil (14) for supply and Gold (13) for demand. After Industrialization, Oil is available for demand as well. After Nuclear Fission, Uranium is available for both.

[solo]

Yes, the fix for coal now works for Cumae.

Your explanation for why oil appears on some early supply lists makes a lot of sense, and in a way it also supports William's contention that oil should not be available for supply until later in the game. Ordinarily, its chances of making a supply list early on are quite low, and when it does, it's probably because of lack of other candidates.

I am looking into spice a little, using the giga-sized map we played an OCC game on where the largest continent produced a demand quotient of roughly 2000 for spice. (No, I didn't dare count them all up!) You would think that this would put spice on every demand list. Not so, but I have some more checking to do on that.

[samson]

Correction.
Oil is not available for supply list fill-in until after Industrialization. I checked on a saved game where I saw Gems appear on both the Supply and Demand lists. The demand was legit, but the supply of Gems was a fill-in. Oil was not used (no Industrialization) and Gold was already being supplied, so Gems (12) was next in line.

[DrFell]

It's possible that for spice there is some kind of scaling factor based on map size, or an upper limit on the continent size count.

[samson]

Yes, there has to be some mitigating factor, otherwise the largest continents would be overrun with spice-demanders.

[solo]

Then William is right about oil, but only when it's being used as a fill in!

It looks like a scaling factor of some sort for spice, but the scale is not based on the total number of tiles. Spice does well on the giga map, but not as well as predicted by its formula.

This brings up another point. Many of the formulae depend on map distances, and it is possible that they may have to be scaled when applied to maps that are not the standard size.

[samson]

Then William is right about oil, but only when it's being used as a fill in!

Not really. Oil is just like any other terrain-dependant commodity. Build a city near a lot of deserts, especially with specials nearby, and it will supply Oil. The big increase in Oil supply comes from the discovery of Combustion.

It looks like a scaling factor of some sort for spice, but the scale is not based on the total number of tiles. Spice does well on the giga map, but not as well as predicted by its formula.

Can you expand on this a bit more? Or post a save of your Giga map test with Spice-demanding cities?

BTW, I'm looking at Spice supply, and it seems to be affected by Continent size, too ... in reverse. Smaller continents produce more Spice.

This brings up another point. Many of the formulae depend on map distances, and it is possible that they may have to be scaled when applied to maps that are not the standard size.

Or perhaps formulas with DistanceToSomething will produce higher supply/demand quotients on larger maps where the absolute distances are greater.