Leader Personalities

I'd guess that option does override any mod-added settings. The goal of this work with personalities was to make leaders use their unique traits/units/buildings better, so the AI's probably stronger with the mod settings. If you like unpredictability though, stick with random personalities.

 

I've done another round of changes attempting to make leaders with military-focused traits more warlike. If you play with beta build 21 or above, feedback on how leaders seem to react in games would be very helpful!

 

Edit: on the actual download page, it says Beta 20 is available - not 21. But b21 is what actually downloads.

 

Yeah just had a typo.

I've since updated to beta 22... really trying to give each leader a distinct personality. I've finally accomplished this with citystates, yay! Each one pursues a recognizable goal with those now. I'm having more difficulty with how the AIs interact with one another though, because they have literally hundreds of variables and most of them are ambiguous.

Right now I'm just trying to enforce that leaders with militaristic traits attack their neighbors, and leaders with peaceful traits don't. It's surprisingly hard to do this!

 

Are you saying that you worked on AI civs' attitudes toward CS? Because in b21, Greece was up to his old self-destructive antics, attacking CS like crazy from fairly early on.

Otherwise I haven't noticed anything different. A lot of denouncements, but the Aztecs, Greeks, Japanese and Russians are all in the same continent with Arabia, so...

 

In b23 I've given Alexander a colossal negative modifier for MINOR_CIV_APPROACH_CONQUEST. He was previously at 0, which I'd assumed meant he wouldn't do it but apparently was not the case. If these modifiers work as I now believe they do (after analyzing some AI files more), hopefully this will block him from wasting his trait by going to war with citystates.

I'm rather happy with how militaristic leaders now declare war on their neighbors early, often, and with large armies. The problem is they tend to then make peace 4 turns later!

I've made some additional changes which should hopefully alter this in b23. In particular... if they haven't suffered any damage in the war yet, they should be less likely to ask for peace. Key word "should!" A lot of this stuff is ambiguous so I'm making educated guesses, then just seeing what happens.

 

Out of curiosity, did you do something to Greece that would have made them more aggressive toward CS in a recent beta? Because I hadn't seen Alexander go nuts like that in along, long time.

Making peace after four turns? Hilarious. They really are such mama's boys (and girls).

I'm looking forward to my next game being b23. It will probably be with the new patch, though, so that may throw some things off.

 

I altered the general AI variables concerning CSs. There was a problem with the way it's set up in vanilla... completing quests (mainly just killing a barbarian camp) makes a leader friendly with that CS, and if they're friendly they got a -150 "war approach" modifier, which basically meant they'd never go to war with that CS. Get a few of those mainly-random events and leaders like Genghis would have no viable CS targets. I reduced this modifier for friendliness to -50 and militaristic leaders are now attacking CSs regularly.

They also favor attacking CSs more than vanilla if their "grand strategy" is Conquest. This means that if they have no good major-civ targets nearby but have built a large army, they can use that army on a citystate instead. Otherwise it's just wasting all those units.

The side-effect was since Alexander is a militaristic leader who commonly picks a conquest strategy, he was going for citystates even though he had a leader-specific "war approach" mod of zero. I've now made this -100, which should hopefully block this option from him.

 

Fascinating. Did you get any value out of those "diplomacy explained" threads?

 

Not really... that thread only went into some basics, and it's more complex than what's described there. The AI has targeting priorities, cooperation and peacemaking levels, flavors for what each leader builds, and how those flavors influence their diplomatic choices, and so on.

 

I just finished a b24 game with Germany, and had no problems capturing barbs throughout. For the first time also, swordsmen upgraded to longswordsmen, etc, rather than landsknechts. More importantly, the civ balance seems strong, but not OP. I played for a science victory via military expansion, and achieved results comparable to a second-tier Science win using builder tactics. Since it was my first attempt at this, I could probably whittle 10-20 turns off another win, thus making both approaches approximately equal.

 

Yeah, barbarian archers now also upgrade to crossbows again. The only special thing now is brutes upgrade to landsknecht. After playing with it for a while, I realized losing the range of archers->lansknecht was disadvantageous so I changed them, and changed the swordsman upgrade since it's somewhat of a "sidegrade" anyway.

 

To me, it's perfect. I wouldn't want brutes to upgrade to swords, given their numbers. And the landsknecht upgrade is cheap. Too bad barbs don't have siege units - that's just about all I've built as Germany.

 

That's the great part about the brute/spearman upgrades... ties in well with the already-inexpensive landsknecht so the upgrade cost is (I think) just 10. It's good synergy, one of those ideas where ya wonder why we didn't think of it earlier.

 

After running five games to midway on prince and king difficulties, I have to say that the AI is not nearly aggressive enough across the board. Maybe that's part of the difficulty level, but it's not just AI personalities. Individual units do dumb things like stand there and take ranged damage until they die. AI cities under siege are not responding well to ranged attacks either, such as training up fast units to hit trebs and such. It's pretty good at garrisoning a ranged unit for defense, but otherwise their land units seem to avoid my ranged units like the plague. Several times I've pushed back a carpet of doom with two or three siege units when they could have easily charged and trashed them. Ranged attacks from sea units also seem to baffle them.

Barbarians aggression is pretty weak too, and I always play with raging barbarians just to make things more interesting. Sea pirates aren't plundering work boats. Injured units tend to run away or wander around like they're drunk, and they can't heal. I don't think run for safety = 0 is actually working in the included Barbarians! mod. One of my cities sitting at a choke point had barbarians lined up four tiles behind it and they wouldn't attack. They just crowded up to the city walls and died one by one, apparently trying to get to something on the other side. I kinda felt sorry for them.

So I'm not sure if it's a bug or just how the AI is calculating attack/retreat when ranged attacks are involved. I disabled mods and played a vanilla game just to check unit AI aggro. On warlord difficulty it made a few mistakes but otherwise responded well enough.

 

None of this has anything to do with a bug, except not moving when being hit by ranged units (especially naval ones). The AI isn't programmed well enough.

I have noticed that since the addition of Denmark, the AI is trying a lot of amphibious attacks. Sometimes they're suicidal, but often they catch a ranged unit unaware.

 

Fair enough. I've been playing a vanilla game most of the day and yep, they're not reacting to naval ranged attacks. Has it always been that way? I haven't been playing V much through all the updates since release, but I don't recall the unit AI being that ********. Maybe because I've just recently been running coastal and naval strategies again. When I posted earlier I was running combined land/sea attacks and the AI units were at least shuffling around the city. Now they're not when I only roll up ships, though they will at least fire ranged.

 

I think it has always been this way.

 

@Erilaz
The problems you describe have always existed in Civ and are part of the tactical AI (small scale decisions), which we have no direct control over as modders. This mod affects the strategic AI (large-scale decisions). The strategic AI involves things like deciding to declare war or ask for peace, what to prioritize in general, and diplomatic relations.

 

Has anyone else been having problems with the AI not expanding? Ever since release in vanilla, I've always seen one or two just sit around with 2-3 cities and never expand, but in the game I'm currently playing it's almost everyone. Large map, PW3, sparse resources; of all the AI opponents, only 3 have more than 3 cities. Spain expanded how I'd expect: filling in all the good city spots around them pretty quickly, ending up with 10 cities. Greece did about the same, with I think 7 or 8 cities. Egypt stopped at I think 5, but is filling out huge populations to take on a typical "tall" empire formation. Every other civ has exactly 3 cities, and Japan even stayed at 1 city until around the renaissance. They all have plenty of room to expand if they wanted to, with prime city spots that even have luxury resources that their empires don't otherwise contain. Their populations for the most part aren't particularly large, and I can't see any reason why they wouldn't want to expand as soon as possible. It's almost the modern era and only India has more than two policy trees completed (and he only has three), so I don't think any of them are even going for a culture victory.