The OCN Problem - and a suggestion to Firaxis

[Lt. 'Killer' M.]

OCN = Optimum City Number, the number that determines, how many cities an AI will try to build, the number that influences corruption, too. And one of the biggest gameplay problems!


Why?
- Because I never see an AI try for domination (or conquest), because even if they did try, they would never grow strong enough, since they are prone to raze cities once they get over OCN.

- Because as long as an Ai is under the OCN for the map size, it will try to reach it, and that means settler diarrhea. I found that a decidedly lowered OCN will greatly imporve on this, and on gameplay in general, but this makes the AI raze everything they take, thus they never grow as strong as a human who simply keeps them all! Every city will be able to produce enought to defend itself eventually, even with rampant corruption, but the Ai will judge you immensely strong because you control so much territory.

For an example, take DS2 SG, we keep all we take, and the AI keeps fighting each other in no-mans-land, because they never take, just raze. Then, the loser rebuilds, and the thing starts over again!


Solution:

Firaxis should change the code for the way the AI decides whether to take or raze a city! If the AI would learn not to raze everything, and would learn to keep cities in WLTKD (thus control corruption effectively - can't be hard to program!), then the AI could create huge sprawling empires as I currently tend to do. Then, it would be a real challenge!

what do you think?????? I'm especially looking forward to comments from mike B., Dan M. and Speedbump and so on from Firaxis!

 

[Gastric ReFlux]

Hmmmm. You have some good suggestions here.

Now if the AI could learn how to use artillery too...

 

[Anglophile]

To a significant extent, just increasing the OCN in the editor will help. Setting it very high, i.e above the absolute limit of 512 cities divided by the number of civs will make available area the main limiter of expansion, as it should be. I have done this but note that most civs stop well short of the OCN (which I have at 70+ for huge and bigger maps). Expansionist civs are definitely more prone to keep expanding in my games but some civs stop well short (halfway +/-) to the OCN. They also don't manage sprawling civs very well, too few roads/rails/improvements/forts along borders/defence in depth/ etc. IMHO, better AI management of larger civs is the real issue as tweaking the editor will significantly raise the number of cities per civ, but they seem to hit diminishing returns quicker than humans, even on high levels with the built-in AI advantages. You will be faced with many more attacking units due to the increased number of AI cities but AI defence is well-known to be the more difficult to program, so the extra size of AI empires tends to mean easier prey for a human, once the initial wave of attackers is defeated.

 

[alpha wolf 64]

I went with a much larger OCN also and noticed that most AIs run out of land before that. Its another one of those wacky variables that is used for 2 purposes and those always cause problems. As much as I hate early game settler diahrea, its a necessity for the AI late in the game. The way I slowed early diahrea was to raise the pop cost of settlers to 3 and make them wheeled. That way jungles and mountains slow them down. I've seen the AI build roads out to jungle sites in order to build a town. One downside to this is the AI seems hard coded to imediately build a settler even if it doesnt have the pop to finish it. Many times i've seen poorly placed cities building settlers with a pop of 1 so that the settler never gets finsihed. I wished the AI would check for the availability of the needed pop before it starts a settler.

 

[Zouave]

Originally posted by alpha wolf 64
. . . The way I slowed early diahrea was to raise the pop cost of settlers to 3 and make them wheeled. That way jungles and mountains slow them down. I've seen the AI build roads out to jungle sites in order to build a town. One downside to this is the AI seems hard coded to imediately build a settler even if it doesnt have the pop to finish it. Many times i've seen poorly placed cities building settlers with a pop of 1 so that the settler never gets finsihed. I wished the AI would check for the availability of the needed pop before it starts a settler.

Yes, that stupid AI.

Your idea is good, but why should we have to constantly tweak the Editor (as much as we can) to fix something Firaxis got wrong in the first place?

And what you suggest still will not prevent rival settler/spearman combos fom wandering into my civ, refusing to leave, and on the THIRD turn ("Declare War or get out!") TELEPORT themselves magically to the OTHER side of my civ and start a town on open tiles they should have no way of knowing even exist. That is such a LOAD OF CRAP it never should have been even considered during R&D., and is the single most annoying pasrt of the Settler Diarrhea problem.


BTW, I doubt if Firaxis cares enougb about the game to change code.

 

[kring]

The problem with raising it to 3 pop cost, is that the AI will try building it as if it were still 2 pop count, and the town will stay stuck building it until the 4 pop count. This causes the AI to get even further behind. Unless this was fixed in 1.21, and I haven't read where it was.

 

[Lt. 'Killer' M.]

All of you are so right!!!!!! I tend to play certain map and terrain settings these days, those with which it works fine, but any time I change them, the Ai sucks


Especially, I'm torn between uping and lowering the OCN, exactly because of the reasons named....... We need a totally refurbished (yes i know de english ferry fell ) AI, fully reprogrammed - and I do know this isn't in the cards, especially since it would never pay......

 

[kring]

I have modified my OCN for all levels; increasing at a higher multiple for the larger sizes of maps.

I will say I haven't noticed the REX so bad after early game with some unclaimed space still being left open by the AI on some occasions; nothing like 1.16 REX.

 

[Lt. 'Killer' M.]

kring: REX???? WTH???

 

[kring]

Rapid Expansion. It's in an abbreviation list somewhere, either CFC or Poly. I hadn't heard of it until recently; it took a while for me to track down what it meant.

But it was extreme in 1.16; other civs would plop a city on a one square opening.

 

[Lt. 'Killer' M.]

ah, OK, same here, with 1.21 it's better, but it still was too bad for my liking. With halved OCN I can play a lot more civs on the same size map and still it's OK.

 

[Lt. 'Killer' M.]

I did some testing to see how strongly the OCn influences the AIs decision on razing.

For this, I used a map where settlers have 100 move on chieftain, smothered the place with towns, then waited what the AI would do.

Sure enough, a city soon flipped. it is in the middle of egypts territory - and they refused it!!!!

OCN is 8, they have 9 cities at that time........

So it seems the OCN is an extremely strong factor.

 

[Anglophile]

I think REX originally stood for Rapid Early eXpansion (but Rapid EXpansion works just as well). A concept from Vel's Strategy Thread over on Apolyton that was pretty good on the early version of the game, especially for aggressive game play on smaller maps (that is actually my interpretation as map size was seldom mentioned in the thread). It ran to dozens of pages and several renewed threads. Haven't read it since 1.16 so don't know the current status/evolution but would recommend it for newer players interested in rapidly advancing their strategic/ tactical approach to the game.