A pattern governing supply and demand changes

[-Jrabbit]

Outstanding work, solo!!!

Sure makes me wish the game offered duplicate chronology (4000bc = Turn 1, etc.) to make tracking this sort of thing a bit easier.

[William Keenan]

I’m also pretty sure that commodity selection is tied very closely to one’s level of technology, which is already especially noticeable when dealing with coal, oil or uranium. To test the effect of terrain, it might be time to bring out the map editor, and see what can be surmised from cities founded on specially crafted city locales, but this is another big project.

Discovering Industralization causes cities to supply oil. Building Superhyways causes cities to demand oil. Discovering Nuclear Fission causes cities to demand Uranium.

[Scouse Gits]

Originally posted by William Keenan
Discovering Industralization causes cities to supply oil.

Sorry...but that is incorrect. In a recent game I had an oil supply in one city with the discovery of trade...circa 2700BC !

------------------------

SG(2)

[-Jrabbit]

He specified that Superhighways (sic) cause Demand (not Supply) of oil. That's = Automobile, right?

[samson]

I confirm SG's observation. Oil is available as a supply commodity from the start, often found occuring cities with the oil resource or having a large number of desert squares.

In fact, I believe that commodities which have a corresponding resource (Gold, Silk, Spice, Oil, Coal, Wine and Gems) are frequently supplied by cities built near that resource.

[Elephant]

Congrats, Solo. You got the first leg of the "final frontier"...

I was suspecting the order of city starts had something to do with WHICH commodities were supplied, not WHEN they changed. I was looking at ContinentNumber too,rather than specific terrain. I had not been paying attention enough to changes due to terrain in the city radius. I'm going to have to start over again. Sigh.

I was looking at terrain issues but had not logged enough of that in my notes. I did get a few good maps with each civ on a separate island, versus all on the same continent. Seemed like it made a difference.

Practical matter: how will we know the order of settlement of AI cities? Track the commodity changes and work backward? The first seven are probably obvious, but after that...?

[solo]

First, a few comments about other posts:

When a new commodity is being chosen at this stage in the game to replace another on the demand list, there are probably only 6 to choose from, a limited supply. 5 out 16 are taken by the other commodities on the lists. (I have not yet noticed a city supplying and demanding the same commodity at once) 4 out of 16, coal, gold, oil and uranium, do not appear on early demand lists. A commodity replacing itself (if this happens) is one more chance out sixteen that can not be counted as a change. That only leaves 6 candidates to choose from.

-Jrabbit,

It’s not really necessary to keep track of all turn numbers, as long as you know the city numbers. They can be pieced together using the Trade Advisor demand lists. Once known, they never change. When a supply or demand change occurs, just note the city number and how far it lies from the particular city that interests you on the list.

Elephant,

Your use of different continents is another key ingredient, I think in the terrain generated lists. Keep going with that!

City order is easiest determined by the dye demand list, since all AI cities except the very latest will end up on that one. Use other demand lists provided by the Trade Advisor to find the exact positions of your own cities. Also, given the turn #, which can be known for sure, you can use Samson's formula to determine a city number when you notice a supply or demand change taking place.

Oil and terrain,

I have also seen oil early, but concur with William’s other observations.

I’d also bet a lot that the first supply and demand lists for each city are determined entirely by the surrounding terrain. I believe that subsequent lists are determined partly by the “dye phenomena” (discussed below), by the current commodity preferences, and also again by terrain, all of which vary during the course of the game.

Following is the latest test:


Test #3 – Some research about additional questions

1) Does this 16 turn pattern occur in other versions of the game?

My test game, where this 16 turn pattern was first observed, was in an MGE hot seat game without any AI civs. I wanted to see if the pattern could be observed in another game, so thanks to the start provide by SG, where the English began the game with Trade, this question and several others could easily be put to the test.

I used my 2.42 version of Civ II, rather than MGE, this start used a large map, rather than the standard one in my MGE game, and SG’s start was a single player game including AI opponents, rather than my MP test game. Enough difference, I would hope, for testing this question.

With Trade at the start, I was able to use the Trade Advisor supply and demand screens to monitor the demands for the AI capitals, as they were founded. These were checked every turn, for any changes. Added to the AI capitals, was my own, London, founded sixth. I founded York, which came in eighth a turn later, which filled the slot behind Delhi, city #7. No contact was made with any of the AI during this test.

The first change that occurred in the demand lists was for York in the year 3600 BC, exactly 8 turns after the start. Using Samson’s formula, it was encouraging to see that when York’s chronological position as city #8, was added to turn #8 (3600), the total was 16, which when divided by 16 yields a remainder of zero, the same as in my MGE test game. The next change was Delhi, city #7, during turn 3550 BC (turn #9), conforming to the zero remainder pattern. (Actually, this change could not be observed until 3500 BC, because the English have to wait a turn to see any changes taking place for the Indians).

The next change would be expected for London in 3500 BC, but nothing happened, but gaps in the pattern are to be expected. As long as cities stick to the pattern when they do change, things are fine. The next five turns resulted in changes for Beijing (city #5), Madrid (#4), Crewe (#3), Zimbabwe (#2), and Cardiff (#1), in that order with Cardiff’s change coming in 3250 (turn #15).

After this, no changes occurred until the pattern fell over this group of 8 cities again. On the second pass, Delhi had another change in 2750, London had its first change in 2700, and Beijing had another change in 2650, while gaps were observed for the other 5 cities. As the game continued through a few more 16 turn cycles, cities that changed did so in accordance with the pattern.

So, to answer the original question, it appears that the pattern does apply to different versions and ways of playing Civ II, and that the remainder N, as determined by Samson’s formula, will always be zero when a change is enabled by it.

(One side note, of interest, is that the year 3250, the first year that a change is possible for the first city founded in the game, coincides with the first spontaneous appearance of barbarians in the game, too. Anyone who has made tests with the map revealed, may have noticed these spontaneous barbarian “landings” also come in 16 turn cycles. Even though the turns this will happen are very predictable, the location of these appearances is randomized, making it unlikely for individual players to notice the 16 turn cycle while playing the game.)

2) Can supply and demand changes occur at times other than those predicted by the pattern?

Yes, the 16 turn cycle is only one factor governing changes in supply and demand. In my observations so far, which have been early in the game, without many events going on, cities have only been changing on “expected” turns. However, there are independent events that may enable supply and demand changes, too.

Samson has already pointed out that a change occurred for him when he made a terrain change, and that this change was tied to the turn that this event occurred. When he purposely delayed the event for a turn or two, the change in supply and demand was also delayed until the turn the terrain actually changed.

Using SG’s start in another test game, I spent my time building caravans and conducting trade between my two cities, London and York. For awhile, this trading had no effect on the 16 turn cycle governing changes, but finally, the delivery of a caravan from York to London resulted in the appearance of a new supply in York’s supply list. This happened two turns before York was “due” for a change. So I built another caravan using this new supply of gold and delivered it to London. No change! However, on the next turn York was due for a pattern change, one occurred, yielding a new supply of gems, and leading to a situation of alternating supply. During this whole time, nothing was happening to London’s supply and demand lists, although London was also building and using caravans.

Anyways, the question is answered. There are alternative ways of changing supply lists, rather than just waiting every 16 turns and hoping for a change. It appears that there may even be some underlying connection between “independent” events and the cycle, too. A lot more unraveling needs to be done, here!

3) Are there any causes to the gaps appearing in the pattern?

I do not have a definitive answer to this, but have discovered a few things that provide evidence that gaps are not a result of a randomized selection of supply and demand choices that happen to result in no visible changes.

a) More gaps were observed for cities belonging to civs that were isolated (i.e. had no contact with any other civ) and for civs that had not yet learned Trade. An earlier replay of my test game, where Trade was not shared around by the Babylonians, yielded more gaps in the pattern, than occurred in the original test. A consequence of sharing Trade was the initialization of contact with a number of the isolated civs, too. Another thing I did when sharing Trade was share maps, so more geographical knowledge may be another factor entering into this equation.

b) During my test, trying to answer question #1 above, I also kept track of all the demand lists of each civ. Of special interest was the demand for dye, because in my MGE test game, sooner or later dye had become fixed in position #1 every city’s demand list, the exceptions being the cities most recently added to the game. This was also accomplished for every new city, too, as soon as permitted by the 16 turn cycle. Why?

My test using SG’s start provided a clue. Of the 8 cities being observed, only York, one of my own, started out demanding dye. During York’s first change in 3600, dye was dropped from its demand list. Oh well, I thought, so much for the “everyone wants dye” theory! I continued observing the action with dye, though. The next 6 changes involved AI cities, all occurring during this first 16-turn cycle:

(Note: the ordering of these lists could not be determined using just the Trade Advisor)

Delhi demands wool, dye, wine
Beijing demands wool, beads, dye
Madrid demands cloth, dye, gems
Crewe demands beads, dye, gems
Zimbabwe demands hides, dye, wine
Cardiff demands hides, beads, dye

Dyes jumped onto the demand list of every AI capital! After this cycle, dyes were gone from the demand lists of my own cities. AI cities only began showing gaps after they started demanding dye! The fact that the human player is the only non-participant in this scheme, suggests more purpose than chance behind commodity change selections. Further evidence of this is provided by the display below, showing the demand for dye plotted vs. game turns in my MGE game, where an “X” means dyes are currently on the list:

XXXXXXXXXXXXXXXXXXXXXX Delhi
XXXXXXXXXXXXXXXXXXXXXX Berlin
XXXXXXXXXXXXXXXXXXXXXX Leipzig
OOOOOOOOOXXXXXXXXXXXXX Rome
XXXXXXXXXXXXXXXXXXXXXX Washington
XXXXXXXXXXXXXXXXXXXXXX Athens
XXXXXXXXXXXXXXXXXXXXXX Thebes
XXXXXXXXXXXXXXXXXXXXXX Memphis
XXXXXXXXXXXXXXXXXXXXXX New York
XXXXXXXXXXXXXXXXXXXXXX Babylon
XXXXXXXXXXXXXXXXXXXXXX Bombay
OXXXXXXXXXXXXXXXXXXXXX Veii
XXXXXXXXXXXXXXXXXXXXXX Sparta
XXXXXXXXXXXXXXXXXXXXXX Ur
XXXXXXXXXXXXXXXXXXXXXX Nineveh
XXXXXXXXXXXXXXXXXXXXXX Ashur
XXXXXXXXXXXXXXXXXXXXXX Hamburg
XXXXXXXXXXXXXXXXXXXXXX Heliopolis
XXXXXXXXXXXXXXXXXXXXXX Thermopylae
OOOOOOOOOXXXXXXXXXXXXX Madras
OOOOOOOOXXXXXXXXXXXXXX Ellipi
OOOOOOOXXXXXXXXXXXXXXX Boston
OOOOOOXXXXXXXXXXXXXXXX Antium
OOOOOXXXXXXXXXXXXXXXXX Konigsburg
OOOOXXXXXXXXXXXXXXXXXX Elephantine
OOOXXXXXXXXXXXXXXXXXXX Bangalore
OOXXXXXXXXXXXXXXXXXXXX Corinth
OXXXXXXXXXXXXXXXXXXXXX Frankfurt
XXXXXXXXXXXXXXXXXXXXXX Alexandria
XXXXXXXXXXXXXXXXXXXXXX Delphi
OOOOOOOOOOOOOOXXXXXXXX Akkad
OOOOOOOOOOOOOXXXXXXXXX Pi-Rameses
OOOOOOOOOOOOXXXXXXXXXX Calcutta
OOOOOOOOOOOXXXXXXXXXXX Uruk
OOOOOOOOOOXXXXXXXXXXXX Pharsalos
OOOOOOOOOOOOOOOOOOOOOO Cumae
OOOOOOOOOOOOOOOOOOOOOO Philadelphia
OOOOOOOOOOOOOOOOOOOOOO Knossus
OOOOOOOOOOOOOOOOOOOOOO Munich
OOOOOOOOOOOOOOOOOOOOOX Argos
OOOOOOOOOOOOOOOOOOOOXX Neapolis

(I'm sorry to see that O's are fatter than X's. But the pattern for dyes can still be ooobserved fairly well)

In this game there were not any AI. However the Romans, perhaps because they start the game as the “host”, have the only early cities not demanding dye, Rome and Veii, when my observations began. However, in this MP game, things were equalized later, as even the Romans succumbed to this craving for dyes. Come to think of it, when checking lists while playing Civ II, the dye demand list has always been very crowded.

The patterns on the dye list are undeniable. Other lists show a hint of being more random, but many of them still follow trends that continue for a long stretch of game turns. I believe that when supplies and demands change, these trends are more responsible for new choices than randomization.

[samson]

coal, gold, oil and uranium, do not appear on early demand lists.

Not true for Gold. In my 16AD, the Mongol capital demanded Gold right at first contact. The other three I agree with, I think demand for them is triggered by tech advances.

I have not yet noticed a city supplying and demanding the same commodity at once

This definitely happens, though possibly it is a late game phenomenon. This snap is from the post-Automobile period.

[solo]

Samson,

Thanks for those corrections.

I was hesitant to add gold to the list. It would have been more accurate to say this was true for gold during my test game. One game hardly establishes a happening as a general rule!

The other 3, coal, oil and uranium, also share something else in common, in that they are all used as fuels.

I stand corrected on dual supply and demand, and am again confounded by an exception to what seems to be more of a valid observation, made from the experience of playing many games!

[Straybow]

Is the city order is reflected accurately in the "Find City" command? As long as you have an embassy you should know all city names for that civ. Having all embassies then gives you a complete list in order.

In playing our recent succession game I noticed that a new city appeared in an odd place in the F1 city list. Most of the cities shown were founded early in the game, (eg, Hoover Sucks before 225 BC) except VotesForWomen, founded by popping a hut shortly after 1650. So why does it appear amongst cities that were built before 1500 BC?

I hadn't noticed new cities in other games ever appearing anywhere but the end of my F1 list. The game in focus is played on MGE, and I have only been MGEing for less than a year, so maybe it is a post-2.42 change made in coding?

[solo]

Straybow,

Yes, the order is the same when using the Find City option, as I have checked it with a game using the 2.42 version of Civ II. I would expect the MGE find city to work the same way.

As long as you have recently exchanged maps with all civs, the list should be complete, too, and would be the quickest way to determine the correct city order.

Thanks for suggesting this better idea of using find city to get the complete chronological city list.

As for your other question about a new city being inserted among the older ones, a possible explanation might be that one of the your cities was destroyed or captured, creating a vacancy on your personal city list that was filled by VotesForWomen.

[Scouse Gits]

solo - thanks for another good read! Dye does seem to dominate demand lists in some games ... but not in every game.

Two commodities which bring equal trade bonuses are Silk and Spice, yet the former is always more common than the latter. This raises the subject of major and minor commodities. 16 commodities are available but only half seem to be widely used. The following seem to be traded more frequently:
Hides, Beads, Salt, Dye, Wine, Silk, Gems and Oil with its demand coming in the later game.

-----------------------

SG(2)

[solo]

SG,

Thanks for that post, and for reminding me that some games do not feature so much activity in dyes, as I was almost ready to conclude that this may be a routine activity in the early part of each game.

I agree that those commodities on your list get most of the activity, too, and have got to wonder what it is that makes things turn out that way for the ones not tied to certain eras. If commodity selections are mostly tied to the type of surrounding terrain, it could be that our natural tendency to avoid having many city tiles with “poor” terrain could be limiting the activity of commodities linked to inhospitable terrain.

When I provided information about dyes in my last post, I also mentioned that the other commodities seemed to be following trends of their own, although not as clearly defined as that of dyes.

Just to illustrate this, I will provide charts below of the changes in supply or demand for each commodity. A plus will indicate an addition of that commodity on a list, and a minus will indicate a commodity that was replaced. If either are in boldface type, this will mean there were either two additions or deletions on that game turn. Turns where no changes occurred will be marked with a dot and turns where both an addition and deletion occurred will be marked with b for both. This will probably look crummy, so be warned in advance! Here goes:

Changes on supply lists are followed by changes on demand lists for each commodity:

….+…+ +--.-….. hides -….-..--..--++.+.--.+
……….+…..- wool .++++-..-…..-…..-.-
.-…..-…….+…+++. beads ..---…+.++…..-…-
..+…..+…+.+……. cloth .--..+.-…-.b….+…
….-.-…..b……..+ salt ….+……………..
…-……….+..+…-- coal ………………….
…+….-…..-….-.. copper ..+……..+………..
..-………-……..- dye ++++++++++.++++……+
……+…….+.b….. wine .-..b--++---.--.-…..
.+……….-…+….- silk …………..-….+..
..+….+………..--.. silver ……………….+..
…………………. Spice ..--…..-……..+….
…-……-……..+.+ gold ………………….

Oil and uranium have not appeared yet. There you have it, for what it's worth. Perhaps there is enough evidence of trends to rule out randomization of changing supplies and demands. There really isn’t enough going on to try and detect a method at work behind it, either. Perhaps we can rely on Samson’s brainpower for figuring out this one, as I bet the solution to all of this is as tricky as his key civ discovery.

[samson]

Two commodities which bring equal trade bonuses are Silk and Spice, yet the former is always more common than the latter. This raises the subject of major and minor commodities. 16 commodities are available but only half seem to be widely used. The following seem to be traded more frequently:
Hides, Beads, Salt, Dye, Wine, Silk, Gems and Oil with its demand coming in the later game.

I think terrain is the key here. I believe it is one of the determinents of supply anyway. Cities are not randomly founded, locations with productive terrain are selected. Thus, commodities appearing with greater frequency may be those associated with the terrains more suitable for cities.

Perhaps a study should be done in which the frequency of commodites appearing in supply and demand lists (over hundreds of cities) is compared with the frequency of the various terrain types found within the city radii. A statistical correlation might prove insightful.

[Straybow]

SGs can confirm or correct me since it was on their turns that Votes4W was built. Some German cities were captured afterwards, but no Mongol or German cities were destroyed. The other civs would take some cheat menu to find out.