Civ Respawn ends in 1500

[Zenon]

I have tested now where respawns occur. It is always the colour of the civilization destroyed that reappears in a random spot. Units do not effect reappearance. The closest distance to an existing city I saw was 7. It is usually 9 or more (simple distance measured in steps). I then tried to narrow the possibilities by founding cities at places where the restart appeared and repeated the process. I thought I could then easier determine the conditions for a respawn. The surprising effect: the probability of a respawn became lower during the process. The only explanation I see is that the program randomly tries a certain number of places. If there is none suitable there is no respawn. If there are cities even in remote places respawns stop completely even before 1500.

This also means I have to question my former findings. I didn´t take this into account. The only certain rules so far: there is a random factor for respawns to occur (or better to NOT occur). The probability is lower if the available free space is reduced by other cities. Respawns stop after 175 turns (deity).

At least this gives an idea how to improve conquest with restarts on: found cities all over the map!


Duke, I have never seen a restart that did not occur immediately after destroying a civ. As for the split: I think this is a different process. If you destroy the capital of a civilization superior in power rating to your own you just need a free colour for the rebel faction. IIRC splits occur before 1500.

Zenon

[Grigor]

Zenon, there is an anecdote that respawning civs occur precisely in places where there is an unpopped hut. Do your findings confirm that?

[Bloody Monk]

Zenon,

Very good!! Does this new finding indicate that in a game where you specified 7 civs but only 6 appeared that the probability of respawn is reduced?? (Assumption: the purple civ failed to generate due to lack of "good" site, and respawn uses the same decision rule.)

Thanks for all your work.

Monk

[Zenon]

Monk, a very interesting idea. I have done some more tests and from what I have seen I´d say there is enough evidence to assume, that indeed starting and restarting conditions are based on the same process. The number of civs that make it also has a random factor. If you start a game on a very small premade map you may get 5 civs in one and 6 civs in the second start. Also, on very small maps where less than 7 civs make it there is a high probability of having no respawn even early in the game, as you assumed, whereas on bigger maps with enough suitable terrain it is almost impossible to prevent a respawn early on. I covered the map with cities and gave up, when the respawning civs appeared even within city limits. This never happened later in the game and would certainly never happen under normal conditions. I think that the neccessary distance to existing cities is one of the factors that changes during the game.

The availability of suitable terrain obviously is a factor. The fact that the human player sometimes gets an almost impossible starting position (like on one of the poles) probably is an exception. Obviously the human player has to be placed while an AI can be left out.

Another factor is the distance to other civilizations. Under normal conditions the civs are distributed more or less equally on the map. On a test map with a big continent consisting of mountains and a small continent with good terrain all 7 civs always started (and respawned) on the small continent. Whatever the program considers to be suitable terrain it seems to check this for each continent. On a continent that has suitable terrain the (re)start can happen on a mountain in an area that we would consider as very difficult. I think that the program first checks a random continent and - if the terrain is suitable in general - then places the civ with a good distance to other civs. If a certain number of checks fail there is no (re)start.

I think I have reached a dead end with this little research. I am sure there is a random factor in the restarting process. But to test the probabilities you need to know what factors play a role. We don´t even know how exactly the terrain affects a (re)starting position. Also there are to many possible factors that play a role (and may influence each other): map size, number of civs, distance, difficulty level, turns played etc. However, for those interested I think I can give this conclusion:

1. Starts and restarts are determined in a process that includes a random factor.
2. The probability of starts and restarts depends on the availability of suitable terrain. Existing cities narrow the availability of terrain.
3. The probability of a restart becomes lower in the process of a game.
4. Restarts end after a certain number of turns (175 for deity).

Not very much, but at least some indications how to deal with „restarts on“.

Grigor, respawns certainly do not regularly appear on fields with huts. This seems indeed to be one of the civ legends. On the contrary, I never noticed this.

Zenon

[Bloody Monk]

Zenon,

This is another in a long line of thoughtful well written posts you have submitted, recently. Thanks, much!!

Monk

[Grigor]

The CFC Game of the Month #33 shows clearly that on Emperor level the 1500AD cutoff does not apply. In 1760 I wiped out all the AI on the same turn and they immediately respawned. However, I found all the settlers and on the next turn I simultaneously bribed every AI settler, but the game did not end! On the next turn all civs respawned.

The map size is custom between small and medium, flat, and there are only 5 civs.

I will attach a save, but nobody can read my saves on Apolyton, so if somebody wants to check I will email them.

[Zenon]

Grigor, thats interesting and weird at the same time. Why would the designer make a conquest more difficult on a lower level? Looks like another bug to me.

I would test the (possible) end date on king and emperor but I have absolutely no material, not even an OCC. And I don´t want to play a whole game just to get a test situation. If someone would care to post or send a sav just before 1750 (king or emperor, restarts on) I´d volunteer to test it. (Can´t use yours Grigor, at least not yet :-) )

Zenon

[La Fayette]

I am currently playing the same game as Grigor
... with full respawning (5 civs) in 1590.
I'll send you a save in 1750 (or slightly sooner, if I manage to win sooner )

[duke o' york]

Has anyone ever won a game by bribing a respawned AI's settler?

Grigor, there is also no respawn if you wipe out all of the AI on the same turn. Since the respawns occur after the turn in which the AI was destroyed, then if you can take them all out in one go they will not all come back. This is the only way to get a quick conquest when respawns are on. I have tested this on all levels.

[La Fayette]

Originally posted by duke o' york
Grigor, there is also no respawn if you wipe out all of the AI on the same turn.

Sorry, duke
Many of us thought this was true, but Zenon has just tested that it isn't + I am playing GOTM 33 (same game as Grigor) and I just wiped out all civs on the same turn in 1590 and all 5 civs respawned at once

[Grigor]

Sorry, duke 'o york. I tried exactly your suggestion, and in that GOTM 33 it was not successful. All the AI were taken out in 1756 on one turn. On the next turn, I found all the new settlers and parked a military unit next to them blocking them from settling. Then the next turn I bribed all the AI settlers. No dice. The next turn the tribes all respawned. The terrain was pretty bleak for the AI by that time, too, and some came back on islands with only hills.

I am sure that you are accurately reporting the results of your tests, but it seems that you have only proved that under some conditions civs will not respawn after 1500 AD. There seem to be other factors which may come into play.

This was emperor level, and an odd-sized map with some terrain modification done by the moderator. Otherwise, it was a normal game, not a scenario.

I will be away for the next 2 weeks and unable to test, but this seems like a typically knotty Civ II problem and I hope somebody can get a handle on it.

[Zenon]

I think I have found the end date on Emperor: respawns end in 1770 (after 210 turns), at least on a normal medium sized map, 7 civs, 2.42. They still happen in 1768.

I also found that respawns always happened in this test, even on tiny islands and when I settled all continents/islands even within city limits. This is definitely different from what I have seen on deity and also on king level, were a random factor applies.

Zenon

[Scouse Gits]

Could the GOTM33 results be due to there being only 5 civs?

Stu

[Zenon]

I don´t think so. I did not test on the GOTM 33 map but on a normal medium sized one with 7 civs and the results are so far consistent with what Grigor and La Fayette found (or should we say endured )

Zenon

[duke o' york]

Well my tests are shonky anyway because they were on the PS version, and I haven't bothered to faff about with MGE. I would not be surprised if various things were changed to make Civ 2 better compatible with the PS engine.