Waaaaaaa...Back to the Cradle

[EPW]

I was playing a cradle game too on ultra gigantic map. The Ai cant keep up with the human player with the high city cap. I had 45 cities when the strongest computer had only 20, with most civs at the 10-15 range. It made the game very easy even though I was playing on very hard.

[hexagonian]

Re: Me too

Kull, welcome back...I'd strongly urge you to go to my site and use the 1.30b files over the 1.35 files.

Originally posted by Kull
1) Naval: Enemy naval units never attack mine unless they begin their turn adjacent to me. The only exception is that enemy ships will sometimes leave a city and travel 2 squares to hit me....but never more than that.

...One of the reasons why I go for 90-100% land/90-100% Continent. I would love to have a competent Naval AI game, and have hopes that something can be tracked down in the Source Code project, but I'm not holding my breath either.

Naval is something that the civ3 AI does much better, IMO, but the random CTP2 settings allows for the creation of a better landmass map with little oceans. (In my current game, I have 3 Longships and a Dromen as my entire Naval force, because that is all I need)

Originally posted by Kull
2) Cavalry: Same deal with multi-move enemy units. I can safely leave a diplomat 2 squares away from enemy cavalry (whether stacked or alone) and they won't EVER make the easy 2-move kill.

Agreed, though I usually do not create many diplomats either... Still, this movement factor might be tied into the lock stacking feature. (Once locked, the AI will hold all of those units as a group and will not break it apart unless merging them into a larger group)

Originally posted by Kull
3) No-Grow Cities: Some cities with excessive slave populations just stop growing - and never restart. Population is added every turn as usual, but on the turn when it should gain a population point the city just flips back to "zero" (for example 15,999 would gain 2 and next turn the value would read 15,001).

Are you running into the cap? Cities need certain buildings to go beyond certain sizes. (Apothecaries are needed to go past size 12) And slaves do not work against your cap.

Originally posted by Kull
4) Slave Revolts: The one solution is to remove the garrison and let the city revolt, and then reconquer it - now all the slaves are citizens! Lose a lot of improvements, though. One oddity. Two cities within 4 tiles of each other revolted in this fashion on the same turn. One was garrisoned with 12 Praetorians while the other had no units whatsoever (discovered this when reconquering on the next turn)

I think the number of units depends on the number of slaves, although this does not explain why you had no units in the second city.

I would put this in the area of an exploit, although there is the tradeoff of losing all of your buildings. The simple way to stop this exploit is to not use it.

Originally posted by Kull
5) Research Hiatus: For no apparent reason, my civ stopped researching in 330 AD, and never resumed (when you looked at "turns until next discovery", there was a dashed line instead of a number). At first I assumed it had something to do with the Dark Ages advance (there's an idea, eh? Imagine halting all research for 10-15 turns when getting this advance!), but by 380AD it hadn't resumed and further testing revealed it to be an anomaly (I replayed the 330AD turn and research continued normally).

This is a new one - my first thought is that you are in anarchy, but if a reload fixes this problem, then I'm stumped.

Originally posted by EPW
I was playing a cradle game too on ultra gigantic map. The Ai cant keep up with the human player with the high city cap. I had 45 cities when the strongest computer had only 20, with most civs at the 10-15 range. It made the game very easy even though I was playing on very hard.

I favor large/gigantic maps with 6-8 civs. I actually remember that when WesW was working on MedMod, he found that combination of size/number of civs was the best in terms of gameplay, and I fully agree with him, based on my own gameplay.

How many civs? What are your landmass settings? More civs means more competition for land, and if you have higher ocean settings, you are reducing the amount of land even more.

You may need to go into Strategies.txt and bump up the priorities and number of Settlers that the AI will build (locate all of the following entries and bump up the numbers)...

// Number of special units that should be built
SettlerUnitsCount 4 //from 1 april 2002

...although the current settings are working rather well for the current (Normal) Gigantic map setting. In my current game, I have 4 civs, out of the starting 8, with acceptable empire sizes - and the bottom 4 are reasonably competitive in tech (I simply cannot plow through them)

The higher setting might be needed since the AI needs to build even more cities on an Ultra map. And at the same time, the player can, almost always, be more efficient in managing the city cap.

[EPW]

I was playing with 13 other civs and most of them had plenty of room to expand. I cant remember the exact settings of island/continent and ocean/land but I think it was around 70-80%. About five of the civs were invovled in a long lasting war in two civs north of mine and cities were being exchanged nearly every turn. There were other civs though who were nearly isolated and failed to expand further than 15 cities even though the city cap was 30-40 higher than that. I'll try adjusting the file and report back on the results.

[stankarp]

I usually play a modified size of ultra gig map, was 105 by 210 but now 90 by 180. 60% land and 70% continent. But even with 12 civs I find I can cope because I am rarely threatened by more than one ai at a time.

I remember a succession game with Hex where the map was smaller again (the original gigantic size) and more land and that was a lot harder because you were closer to the ai and therefore more likely to be attacked by more than one at a time. The ai wont cross water to attack you with land units except for small attacks. More land means more access for the ai.

I have introduced some new techs and new units that slow progress down a bit and noticed that with two types of chariots, the AI sometimes combines stacks of pure chariots that can move 2, and have been caught out by that a few times. Also hypaspists that move 2, the AI sometimes has stacks of only hypapsists and mobile archers which you have to watch.

[Protra3211]

Seems like a great mod

Just some small points i want to point out. I dont how to make my own mods yet but have been checking thur the files of the all the mods to see how they compare.Like many in here tried my own hand at tweaking files this way and that to see how the ai works .You could lower the amout of cities from 65 to 55.My thinking here is if I wanted to expand more to take out a 1st place civ I would have to give up a city or two to another civ.
In the early part of the game the ai expands to the goverment city cap with no problems. I tweak the early gov. city cap by 5.
Its in the modern age as the human play begins to go over 45 to 50 cities -the ai begins to fall behind. Most players are at 1st place and will never look back.
This may be due to the way the scoring is done .
In the Player1 mod he made changes in the DiffDB text file lowing the Wonder factor score to 250 and others Adance factor from 10 to 20 maybe he felt it gave the game more balance.
One last thing in the Personalities text file in the line for Expansion: I didnt set all to max I like to see the more agressive civs gain more land . In my games some civs would have as much as 40 to 45 cities not much more.

[hexagonian]

Originally posted by Protra3211
You could lower the amout of cities from 65 to 55.My thinking here is if I wanted to expand more to take out a 1st place civ I would have to give up a city or two to another civ.

In the early part of the game the ai expands to the goverment city cap with no problems. I tweak the early gov. city cap by 5.
Its in the modern age as the human play begins to go over 45 to 50 cities -the ai begins to fall behind. Most players are at 1st place and will never look back.

The cap is a two-edged sword for the AI because on the one hand, the player is limited in expansion (if he does not raze cities) but on the other hand, the AI cannot manage the cap for itself.

I actually had to boost the cap in Monarchy to 50 in my current game because the Romans were above the normal Cradle cap of 40 for that government.

Unfortunately, by severely limiting the cap in the late game, you end up limiting the player in his gameplay style. Expansionistic warmongoring is out the window. For some players, that is not acceptable. However, I'm finding that the gameplay is just as interesting because I'm more of a builder-type, and the AI still maintains an aggressive stance with its military. Expansion has to be handled with a lot of forethought too. Certain cities are not worth the effort to take because there are better ones out there for the taking.

I really like the city cap as a means to halt military expansion. It's somewhat realistic to expect that as you expand past the means of your government to support your empire, you have a greater chance of riots, etc.

But rather than an all-or-nothing approach of riots, I'd like to see crime take away a large chunk of production/food/commerce but still have the city function to some degree.

And you are right - the problem lies in the Late Medieval/Modern age because the AI does start to fall behind once the cap gets to 40+ cities. (Still, in terms of gameplay, that is 600 turns - and Cradle was orginally designed to end during the Renissance)

I'd take 600 turns of CTP2 any day, over a full game of 500+ turns of civ3...

Originally posted by Protra3211
This may be due to the way the scoring is done .
In the Player1 mod he made changes in the DiffDB text file lowing the Wonder factor score to 250 and others Adance factor from 10 to 20 maybe he felt it gave the game more balance.

Reducing the score gains from Wonder construction would probably help slant the scores more toward the AI, since once the player gets tech superiority, he can get a lot of Wonders.

It's just that in Cradle, it can take a long time for the player to get that tech superiority...

[stankarp]

Hex,

What about reducing the riot chance when under the happiness level and the the revolt chance as well? That might help the ai.

[Protra3211]

I think your right.

Craddle seems like a great mod. Have to give it a try -battles between other civs(from threads I have seen)takes planing on the level of chess.Have to think ahead more than one turn. Most mods I have played take between 75 to 80 percent of the game before im ahead.I play 8 to 10 civs on a map mostly land 90 to 180 size map --need alot of room for each civ.very hard.
The Ai in ctp2 puts up a good fight for most of the game now. Needs something for end game play. Like chess it only takes a small lead in the end game like a passed pawn and its over. Something like a merging of civs could be added to Craddle -Peter Triggs has a program like this in CTC. Last thing I think you others have done alot of work to make this a great game. Now if only Civ3 was put together with this same passion it might have been the best empire game ever made.

[hexagonian]

Originally posted by stankarp
Hex,
What about reducing the riot chance when under the happiness level and the the revolt chance as well? That might help the ai.

I have though about that, but this also works in the human's favor, because it becomes an easy workaround for the city cap.

And the crime effect is too small for each point of happiness lost, and I haven't been able to track down how to control the amount of lost production/commerce/food due to crime per happiness point lost. If I can get this figured out, I can give the cap a little more of a margin without having it fall into an all-or-nothing effect that a riot creates.

Sidenote - I was able to pillage my own Pasture.

Originally posted by Protra3211
Something like a merging of civs could be added to Craddle -Peter Triggs has a program like this in CTC.

I did have some questions about how effective this is for gameplay. Does it really work without causing civs to inadvertantly bypass the cap due to a merge.

[stankarp]

Correction, I thought there was a setting for the % chance of riot once below the happiness level but I could not find it. Sure I saw something once. Found the revolt level.

Is there a setting for the % chance of riot when below happiness threshold?

[J Bytheway]

Originally posted by hexagonian
And the crime effect is too small for each point of happiness lost, and I haven't been able to track down how to control the amount of lost production/commerce/food due to crime per happiness point lost. If I can get this figured out, I can give the cap a little more of a margin without having it fall into an all-or-nothing effect that a riot creates.

Would you like me to go hunting through the source code and try to figure it out...

[hexagonian]

Originally posted by J Bytheway
Would you like me to go hunting through the source code and try to figure it out...

I have several things that are my holy grails to be fixed in CTP2...

1. A replacement for Frenzy that will help the AI garrison newly-captured cities with strong units, whille maintaining an aggressive stance.

2. A means to help the AI effectively manage it's city cap - something that I do not think can be done as I want it to be. (This would involve the AI switching to a mindset not to take or build additional cities if it is drastically over the cap - possibly become more of a pillager and focusing on taking out field armies, and possibly disbanding/razing captured cities or its small cities to get the most efficient empire.)

But if I can use happiness loss to adjust crime levels in a drastic manner without having the city fall into a riot state, then this will be a viable compromise solution - at least it would be better than what is currently in place.

[stankarp]

As an ineritm, Hex, why not change the revolt level.

My observation is that once a ai civ starts breaking up, it keeps happening. By delaying it it gives the ai a bit more time to get to the next level of government. In addition, the AI seems to go after its delinquent cities quickly, distracting it from attacking other players. If its a slave revolt, the city will have a garrison which ties up the ai more.

Think its worth a try.

[stankarp]

Also, the ai has production, pw,science and gold boosts. A few cities rioting will hurt the AI less than the Human.

The point I was making is that once an ai's cities start revolting, it cripples an AI most of the time. Changing the revolt cap helps the ai more than the human.

[hexagonian]

Well in my current game, I boosted Monarchy to 50 (from 40). I was in Theocracy, and had no plans to switch to Monarchy. I had done this to benefit Rome, as they were exceeding the 40-city cap. (I was at 34 at the time)

No problems with the AIs - in fact the AI civs that are my main rivals (Rome and Persia) are both nicely tucked below the cap, and are managing to maintain a safe level of happiness - in fact with a rather healthy margin. (I had boosted entertainer useage for the AI for 1.30b in strategies.txt, as well as added some happiness buildings for the second level Settler unit)

But Protra is right - they both seem to plateau and are not expanding as much as they were earlier. (Rome is still in the low 40s and Persia is at the high 30s - though they both have a cap of 50. I'm at 52, with a cap of 50 - however, I made a huge priority to have happiness buildings and Wonders, so my empire is very stable, and I have additional room to grow even more)

There is not a lot of free land left, so growth will have to come via conquest. So the problem now seems to be to get them to continue to expand. There is a lot of back and forth on cities - they switch from civ to civ like a hot potato. The bottom line is that no true growth is occuring post Medieval, at least not to the extent that I want...