Early Game Strategy

[Curiosity]

Ok, background to this post; I'm hoping to do work on the CtP2 AI once the source is released (if ever ) so I'm trying to work out how to play the game in advance, with an eye towards what could be implemented in the AI. So, I'd like to put forward my thoughts and pick the brains of the people who are much more experienced than me.

(Oh, this is with the unmodded game, btw. I want to get a feel for it before I move on to the Apolyton mod)

First, on base-placement; it seems that if there is an overlap between bases then the unused tiles are counted as having zero production... this massively reduces productivity. Therefore base-placement seems to need odd-numbered distances to be effective - 3, 5 or 7 tiles apart. Also, it needs to be planned out carefully, as having one base 3 tiles and another 5 tiles away screws everything up nicely.

7 tiles seems way too far apart. 5 is quite decent; you get 18 workers and the rest are specialists. 3 tiles isn't too bad ether, since specialists are pretty nifty in this game, and each base gets 3 free workers.

Generally, the AI seems to place it's bases way too far apart, resulting in it running out of new base sites with 6-7 cities... given how fast it can get to Monarchy, I'd like to see it with around 20-25 cities before 1AD. With all the bonuses the AI gets on Impossible level, I can still have 2-3 times as many cities, a higher population and a larger military at this point.

On that note, do people here respect the limits on empire size or not? In my first game I did, and in my second I built 20 cities under Tyranny, which wasn't too hard... I'm planning to conquer another 5-6 in Monarchy and stop until fascism. I would say that it's more worthwhile to build extra cities in the lower forms of government, given that the penalty affects your whole empire.

Science, commerce and production improvements appear completely useless until there's absolutely nothing you can do to increase your growth. I mean, +10% gold? It seems like a good idea to make the AI focus much more on happiness and growth, and have it play with it's sliders a whole lot more. I think I can do this in the text files, in fact...


Anyway, if anyone has any strategy thoughts on the early game, I'm interested.

Also, does anyone know what the general AI improvements are in the mods? I understand that it's a lot more aggressive, which it certainly needs to be, but does it play any better?

[Maquiladora]

On that note, do people here respect the limits on empire size or not? In my first game I did, and in my second I built 20 cities under Tyranny, which wasn't too hard... I'm planning to conquer another 5-6 in Monarchy and stop until fascism. I would say that it's more worthwhile to build extra cities in the lower forms of government, given that the penalty affects your whole empire.

The only thing that stops me from building more cities is happiness in my least happy cities, which is the most important thing in my new cities in terms of buildings.

Science, commerce and production improvements appear completely useless until there's absolutely nothing you can do to increase your growth. I mean, +10% gold? It seems like a good idea to make the AI focus much more on happiness and growth, and have it play with it's sliders a whole lot more. I think I can do this in the text files, in fact...

I wont build gold buildings for a long time (never caravans) until my cities produce 100 or something gold each then theres not much point, unless it only takes a turn or 2 to build it and im looking for something to fill a hole in a production queue.

I usually get cites growing at around +1000 then i only focus on commerce tile imps around cities, first ports then trading posts. Mostly i dont build any mines until advanced mines because they include commerce.

I max out food and production for the first stage of the game, later on i might swap them around depending on how my science graph and pop % is doing compared to others.


Things that could also help:

Aqueducts - make the AI build them until the city is at least size 9+
Bazaars - only build in cities with plenty of commerce or large cities with empty queues in safe areas.
Academies - dont build in size 2 cities surrounded by plains! build in forested cities by the coast with plenty of ports or large cities.
Granaries - only build when city has at least 100 food or something.
and a whole load more of buildings.

If you can make even these 4 things possible it would free up the AI's build queues to build more important things, like units, as long as it know what to do with them.

7 tiles seems way too far apart. 5 is quite decent; you get 18 workers and the rest are specialists. 3 tiles isn't too bad ether, since specialists are pretty nifty in this game, and each base gets 3 free workers.

There was thread about city placement somewhere ....... I build cities 0-25 or so to expand by 2 borders (<18) then later when cities grow quicker id give 3 border room if theres enough space. cities on rivers first then by coast, forest, grassland, plains or somewhere in that order.

[Peter Triggs]

Curiosity,

If you're interested in CTP2's AI there's a document by Mr Ogre called 'AI modification via SLIC' that should be in Poly's directory.

Also you might find this thread interesting (nobody else did, though ):

http://apolyton.net/forums/showthrea...threadid=86542

[Curiosity]

Maq,

Thanks, sounds like good advice! I wish I could control the AI's build decisions better through the text files, but there seems to be no way to prevent it from building Aquaducts in size-2 cities with low growth.

Peter,

Very interesting posts - I will definitely play around with strategy triggers. I way hoping there was a way to switch to a new strategy when it hits X cities, so this is perfect.

[Tamerlin]

Originally posted by Curiosity
Maq,

Thanks, sounds like good advice! I wish I could control the AI's build decisions better through the text files, but there seems to be no way to prevent it from building Aquaducts in size-2 cities with low growth.

I think Martin has changed a lot of things in the AI Civs' build lists in his GoodMod so that their production is more efficient. You should have a look at the GoodMod files.

[Martin Gühmann]

Originally posted by Curiosity
Maq,

Thanks, sounds like good advice! I wish I could control the AI's build decisions better through the text files, but there seems to be no way to prevent it from building Aquaducts in size-2 cities with low growth.

You could remove the Auaducts from the AI build list. But then the AI wouldn't build in any city aquaducts. But you can look into the strategies.txt for the building priorities. And check which build lists are used there. If you want make it as hard rhule: AI never build Aquaducts in size 2 cities you could use slic to prevent the AI from doing it.

-Martin

[Curiosity]

Martin,

From what I can see it's a question of balance; move the Aquaduct too low in the AI's prorities, and cities will sit at full size building completely useless improvements before getting to Aquaducts. Growth seems to be the most important factor in the game, so that's not desirable either. I want the AI to build Aquaducts at the right time, no sooner or later.

I'm not fully up to speed with slic yet; I assume that you mean there's some way of changing build orders in slic on a city-by-city basis? If so, then it might be a good idea to force the AI to build Aquaducts when a city gets to size X, providing it's not in a bad strategic position.

What I was considering was a fuzzy-type function, MostCitiesAreSize(int size). Once that goes to true for the appropriate number, the AI considers a new strategy where Aquaducts are set to a high priority.


What's really needed is a bunch of sophisticated conditions and priority modifiers in the build list section of the strategy files; something going way beyond ''Top 0.2 ProductionCities'. Once the code is out this may not be that difficult to do, but implementing it in slic...