[MOD] More Naval AI

So what I'm hearing is that the AI doesn't do well if it's not conquering its neighbors? Fair enough, though it's been my experience that achieving any sort of victory without major military engagements is pretty much impossible in Civ anyway. Is that incorrect? If so, I'd love to hear some specific strategies from anyone who regularly wins without tearing into their neighbors.

Well, there is conquering and then there's falling into blood frenzy, throwing all your resources into building an army and trying to kill everyone you can find. Any AI that tries to play a more balanced game end up wiped out (I'm talking about 1.61 here, the new game with 1.7 seems to be much more sane).

Indiscriminate razing seems to be the major contributor to such behavior - without actually conquering new territories civs never get hit by growing maintenance so never get motivated to stop their rampage. A civ conquering it's neighbor is perfectly fine, but the constant tendency to go to the other side of the world and destroy some civ just for the sake of it ruins the game. Maybe some conditions to AI's raze decisions could be applied, e.g. diplomatic relation should be deep into negative (likes vs dislikes), presence of the conqueror's state or shrine religion in the city or something.
 
My mod is based off of 1.61 ... so it would be easier for me to upload my new Assets folder rather than the entire file :)

(although I will likely do both) ... just easier for people to try my mod methinks. [not yet released]
 
Tholal, any particular reason why the AI is now overvaluing drill promotions? Most units should still bet on combat promotions, in my opinion, especially since one could see the AI using blur more often in the previous patch.
I haven't spotted a unit with the march promotion yet (215 turns played).

Indiscriminate razing seems to be the major contributor to such behavior - without actually conquering new territories civs never get hit by growing maintenance so never get motivated to stop their rampage.

Your experience is the complete opposite from mine. I still say that the AI tends to keep rubbish cities too much.
 
Your experience is the complete opposite from mine. I still say that the AI tends to keep rubbish cities too much.

Well, in my game they don't keep ANY cities, rubbish or not. In my current game (1.7) Illians just razed a large neighbor city with 3 or 4 luxury resources in it's FC. Multiple wars are going on but not a single city has been occupied and I'm not even counting how many were burned down. Doesn't even matter whether it's a neighbor or someone distant.

I wonder though - I play with time, cultural and religious victories off, could this have such an impact on AI behavior? Although I've always played with these settings and never had such problems, 3 games in a row now.
 
Well, in my game they don't keep ANY cities, rubbish or not. In my current game (1.7) Illians just razed a large neighbor city with 3 or 4 luxury resources in it's FC. Multiple wars are going on but not a single city has been occupied and I'm not even counting how many were burned down. Doesn't even matter whether it's a neighbor or someone distant.

I wonder though - I play with time, cultural and religious victories off, could this have such an impact on AI behavior? Although I've always played with these settings and never had such problems, 3 games in a row now.

Strange. I haven't seen this sort of over-razing behavior either. Try a game with different settings and see if it makes a difference.
 
I have seen the AI capture cities (and not raze them)

at least in 1.61
 
Tholal, any particular reason why the AI is now overvaluing drill promotions? Most units should still bet on combat promotions, in my opinion, especially since one could see the AI using blur more often in the previous patch.
I haven't spotted a unit with the march promotion yet (215 turns played).

Okay, I must take this back. In the same game, now that the AI is more active on wars I'm seeing much more units using combat promotions. It was kind of weird seeing at the beginning of the game most units with triple drill promotions, though. Is this really an intended behavior, or was it just a fluke?
 
drill2 and drill3 ain't to bad usually ... yet combat 1 is usually better than drill 1.
 
I'll tone the focus on first strikes down a little. I've always considered first strikes to be a great addition to a defensive unit. Is that not considered optimal?
 
Depends on how good of a defensive unit it is ;), I just typed up a (somewhat lengthy and technical) discussion of this in another thread. Not for the feint of heart, but a good synopsis of how they play into the combat model statistically.

First strikes are good for units that are already pretty (if you'll excuse by french) badass, but there is usually a better pick, especially with the more powerful promotions in FfH. If the unit is on the wrong end of a major difference in combat strength (after adding in health and modifiers), no number of first strikes will be very useful.

By the way, really like what you're doing here. Played around with it several versions ago and started building some of my own modifications around it. Once I get DuneWars to where I want it for the next full release I'll take a break and see what kind of progress you've made. Till then the change logs look good :D

Edit: it's been a while, I'd all but forgotten about defensive strikes in FfH. Those at least give some usefulness to the drill line for defensive units. The 1 chance first strike at drill 1 is a joke compared to 20% combat percent for taking combat 1
 
Chris's explanation is very helpful.

Anyway I'll say that right now what seems to happen is that the AI is mostly giving drill promotions to defensive units and combat promotions to conquest units.
But this makes defensive units weaker. When I rushed the AI I lost much less units than normal because the level of strength between my units and the defensive units was not significant. The little defensive strikes that were happening all the time were interesting but the damage was negligible.

Drill is certainly most useful in defensive positions when the unit is already strong. A specific unit may be the strongest on the field, but with only combat promotions it will not survive mobbing. If a unit like a flurry is maximized on combat and drill promotions, then we may see (without the use of blur) that unit doing the ridiculous thing of defending against 100 units without barely taking damage.
With only drill promotions, the unit may not be even a favorite in a single fight.

With this being said, the AI is certainly more active protecting its own improvements and trying to keep out enemy units which is absolutely great.
 
After my first 100 turns, I have to comment on the behavior of barb lizardmen. I'm used to them moving more or less directly toward my cities, unless they could threaten a worker. In this game, they seem to be dancing around my borders, and I've only had a few who actually attacked my units.
 
Tholal, you are slowly creating a monster. The AI still does sometimes some dumb mistakes, but it's now more aggressive than ever. Ridiculously aggressive. I'm amused and I'm not amused...

What do I mean? Well, playing diety level, Alexis just declared war on me on turn 16! The hell...:confused:
What happened? She settled her second city right on my border and instantaneously declared.:eek:

Now that the AI is placing its second city further away from the capital what is happening is that the closest AI will tend to place its second city right next to the human capital if it's a good spot.
In another game, Hannah also placed her second city right on my border. This second city was 12 tiles away from her capital!
Is the AI trolling?...
Inevitably she declared war fairly soon.
Six other games where I retired early day the AI placed its second city right next to me. It didn't always declare war but inevitably I got working tiles stolen from my capital as I was trying to get some settlers going.
Granted, this complaint is only regarding deity.

But the game just became way more tougher and I'm frustrated, lol.
I guess that this is good. Continue with the good work.
 
I have played two naval games with this mod, so far, and I wanted to say how impressed I am at the improvement in AI performance. I particularly like how you have simplified the early naval units and their promotions.

My first game was an emperor one as Hannah on an archipelgo map and as others have reported the AI is now settling offshore islands using galleys and using privateers to pillage seafood and blockade. I had a nice little battle between 4 of my privateers including The Black Wind and 4 other privateers (not sure whose they were but they were stacked together). Now there are plenty of AI ships sailing about the Lanun boarding party unit makes sense and mine captured 2 privateers :lol: I ended that game at about turn 350 when I was 3 times the score of any AI, ahead of everyone in technology and obviously going to win a domination victory if I could be bothered. The AIs were weakened by fighting among themselves and the Sheem had taken a city from the Elohim by an amphibious invasion with Curly (the hill giant) and a stack of adepts casting skeletons and some pyre zombies. The AI were good at claiming wonders and early religions but the power of the maturing pirate ports was soon too much for them. Their economies seemed to be dragged down by overbuilding low level troops but I didn't check carefully what they were doing.

My second game was an immortal one also with Hannah on an archipleago as I wanted to compare the two game levels. I had a good starting city with 4 hills and room for 3 pirate ports but still missed with my attempt at the Heron Throne and Great Lighthouse. The two early religions (OO and RoK) that are good for the Lanun went before I could think about trying to get them. I tried to settle my third city on a nearby island before Decius of the Calabrim. He settled the island soon after and few turns later after I adopted Runes of Kilomorph declared war on me. :(

What followed was some of the most exciting naval warfare I have ever had in civ :D even if it was small scale it was a battle of manouvre with feints and thrusts and retreats. He was using galleys, lots of them with the Buccaneer promotion. He pillaged my seafood and pirate ports and attacked any of my galleys that were damaged. Initially all I had were my 2 galleys to face his 4 and as fast as I killed his and replaced my losses he built more. I was intending get Warfare and so I needed Bronze Working anyway and that gave me triremes (even though I had no metal). That turned the tide but he still inflicted a lot of economic damage to my cities with blockades and pillages and sank 2 of the triremes at the cost of maybe 10 galleys. We were fighting for about 50 turns.

Eventually I resolved the Decius problem and took his 3 cities with a stack of 4 catapults and 10 swordsmen and fortunately a single well promoted archer (who killed a lunatic and later Saverous). I consider myself fortunate that he did not build a training yard and upgrade his bloodpets to moroi and that he didn't have metal or the story could have had a different ending ;)

One comment on game balance; I did find that the Tsunami spell seemed too powerful and overwhelming for both naval and coastal battles. I found that a fleet carrying 2 cultists could reduce the defenders of any city to such an extent that I could attack directly from the ships with ordinary troops without the amphibious promotion and kill the defenders. So basically the way of taking coastal cities is to use the Sirian Doctrine of attacking directly from your ships supported by cultist firepower. I'm not sure quite what to suggest but it did mean that cultists are a one stop solution to naval and coastal warfare and carrying them around on ships makes them fast and flexible in a way the AI can't cope with. The max damage they do to troops sheltering in cities could perhaps be reduced and damage per Tsunami might be toned down so more of them are needed to produce the effect. Perhaps the number of units damaged per cast should be limited (like collateral damage) or like Blinding Light there is a high chance of no effect / spell save. Does anyone else agree that this spell is a bit too strong?
 
Sirian doctrine?
 
Do a search in the Civ 4 Strategy and Tips forum ;)

Spoiler :


Essentially it is a quick way to take cities, by attacking directly from ships. Originally you could use cannons and artillery from ships to provide collateral damage on the defenders but this was patched in BtS.
 
Would you like to play an MP game using Tholal Naval AI mod 1.7?
 
One comment on game balance; I did find that the Tsunami spell seemed too powerful and overwhelming for both naval and coastal battles. I found that a fleet carrying 2 cultists could reduce the defenders of any city to such an extent that I could attack directly from the ships with ordinary troops without the amphibious promotion and kill the defenders. So basically the way of taking coastal cities is to use the Sirian Doctrine of attacking directly from your ships supported by cultist firepower. I'm not sure quite what to suggest but it did mean that cultists are a one stop solution to naval and coastal warfare and carrying them around on ships makes them fast and flexible in a way the AI can't cope with. The max damage they do to troops sheltering in cities could perhaps be reduced and damage per Tsunami might be toned down so more of them are needed to produce the effect. Perhaps the number of units damaged per cast should be limited (like collateral damage) or like Blinding Light there is a high chance of no effect / spell save. Does anyone else agree that this spell is a bit too strong?

Indeed, on an archipelago map you only need cultists and boats, basically.
Any map that there is enough cities near coast or lakes, cultists will make you win games with way less troops. One should need more cultists to achieve the same amount of damage that one can get now with just a few of these units.
 
In .41 n I was able to modify the python code for reqPirateCove(pCaster)
that allowed pirate improvements to be placed on resources, as well as overwrite fishing boats.

I updated to .41p with More Naval AI added.

As far as I can tell the function reqPirateCove(pCaster), no longer exists in the mod folder. I am unable to replicate my early change.

Was the check of work boat's placement of pirate improvement moved to the .DLL ? Or am I completely missing something.

Thank for your time!
 
Back
Top Bottom