Game mechanics

[Nelphine]

Impossible, November 1415 (we didnt meet until 1411 or late 1410), patch 2.7.

And I am stopping him from crossing planes but I've only killed one settler this whole time, and I think he's only settled maybe.. 5 cities in this whole time (he may still have been settling his home continent and I couldn't see it, but that's only 15 or 16 cities total.

Oh right, the continent did have 2 breaks I missed - a river turned into an actual 1 tile wide canal, about 9 tiles long. So it was 2 continents plus 2 islands not 1 and 1.

[Seravy]

his casting skill is around 150. So he wants mana income equal to his power maintenance plus 4500, so he can cast overland, pay maintenance, and fight 9 full price battles per turn.

This part the AI already does. If mana reserve<32x casting skill (might be 16 think is 32 now), then all casting skill production is redirected to mana. And it's always active, not just when they are losing.

if trade goods are available for mana income

Define "available". You can set Trade Goods any time so in theory it is always available. But it would not be smart to just set all cities to trade goods whenever we want mana. Currently the AI sets a set percentage of cities to trade goods (not many) at random which is a poor system and could use improvement but I have no ideas how to do it better.

all power that would normally go to mana, should instead go to casting skill

What if the above condition to redirect skill into mana is also true because there is not enough mana produced with trade goods? Which would override the other? Do we push skill when we can't afford the mana needed to use it all, or do we...not.

More importantly, these rules already exist :
If total researched spells*0.75<Player researched spells, then set "research needed". If research needed, then have a 33% chance to turn any other choice into a research choice in power distribution (somewhere in the middle of the execution, so other effects can still override it, such as needed mana)
Likewise do the same for skill is AI skill<1.5* human player skill.

Settlers should increase as time goes on, if there are empty city spots.

This one is impossible. When I call "is there s city spot" from "AI change production" the game crashes. Maybe I should try again though but there is no need to. Current rule is, always produce a settler if number of existing settlers<2. So if there is a spot where the settler can be used up, a new one is produced immediately.
I would like to point out 2.71 and earlier versions had at least 4 different bugs (and base game have several dozens) that caused settlers to stuck in movement in which case the AI becomes permanently unable to settle anything since they are assuming the settler is moving to the spot when it is not. 2.72 should not in theory leave any spots open (unless it's the other plane and the AI has no access to a tower the settlers can reach)

they could have 8 settlers all going at once.

This would be an absolutely horrible strategy. The number of cities has a negative correlation with empty spots left on the map. The more cities they have, the less likely the settler can find a place to go to. Unless there are bugs, the AI tends to fill every single spot before 1410 on arcanus on impossible. After that, only razing opens up places one at a time.

For instance magicians could grant 1 research, priests 1 power.

Don't like this idea, would want people to make units that suck in combat, and it would encourage bad strategies for the human player, such as making magicians against enemies who have a Magic Immunity spell.

In every FFA game with only 1 winner the correct choice when there are only two players left is for them to declare war on each other.

I feel this isn't really right. Of course the best players won't often fall into this trap, but if the human player leaves too much time for the AI before the war (either because they want to get stronger, or just want to have fun with nodes or whatever), the AI might develop a way to win the game. For example if they finish the Spell of Mastery and the human player has no way to take out the capital or spell blast, they lose. But if the AI wards every city they have against the player's realms and they do not have dispelling wave, that might be enough too.
On the other hand the AI surprise attacking the player despite alliances if there is no one else left, would be nice. Except...it happens. Not the "direct" way but if the AI starts casting the Spell of Mastery, that has the same effect. You either go to war immediately or lose. If they also had Time Stop researched, you lose 99% of the time.
And believe me if the human player is already ahead, attacking them hurts more than helps.

Oh, and there is a mechanism to make this happen, although it isn't strong enough to break a stable alliance, but if the human player controls a lot of cities (lot is like one third of the world or more) then they periodically get a random diplomatic penalty just for that.

Impossible, November 1415 (we didnt meet until 1411 or late 1410), patch 2.7.

Then it is one of the bugs found this week. In short the AI settlers were unable to find their way to shores to get loaded into ships, or they went for unreachable tiles blocked by lairs or units. 2.72 should not have this happen. You can easily find out what is the problem, just locate the AI's settlers and see if they are moving or on a shore waiting for a ship to come for them. If they are neither then they are bugged.

Another possible reason could be if the AI randomly selected one of the small islands for their settler destination but that should not have lasted for that many turns, they re-select periodically.

[Seravy]

An idea. What if the AI's "player has too many cities" diplomatic warning and penalty would also including this : If turn>250 then for each wizard already eliminated from the game, penalty value is increased by 0.5 for every 50 turns elapsed?
Alternatively, the bonus relation for having a treaty could be reduced to zero when turns>300? Then natural gravitation would eventually draw it to the starting value and if it's low enough war might start. (though if they play the same books as you it still won't)

Or we can just let the player hold the Alliance and the AI research their way into Spell of Mastery and pack all their cities full of very rare creatures, plus building several stacks of 4-9 of those waiting for the upcoming war..I kinda like this better, with the kind of advantage the AI has on higher difficulty, the timing for that war is not that trivial for the player. Too early and they lose for being too weak, too late and they lose because they AI got too strong. And ofc waiting is not an option at all if the AI has stuff like Armageddon in effect.

[namad]

An idea. What if the AI's "player has too many cities" diplomatic warning and penalty would also including this : If turn>250 then for each wizard already eliminated from the game, penalty value is increased by 0.5 for every 50 turns elapsed?
Alternatively, the bonus relation for having a treaty could be reduced to zero when turns>300? Then natural gravitation would eventually draw it to the starting value and if it's low enough war might start. (though if they play the same books as you it still won't)

Or we can just let the player hold the Alliance and the AI research their way into Spell of Mastery and pack all their cities full of very rare creatures, plus building several stacks of 4-9 of those waiting for the upcoming war..I kinda like this better, with the kind of advantage the AI has on higher difficulty, the timing for that war is not that trivial for the player. Too early and they lose for being too weak, too late and they lose because they AI got too strong. And ofc waiting is not an option at all if the AI has stuff like Armageddon in effect.

Hmm, maybe just make the ai declare war instantly on impossible? Since we can safely assume anyone playing impossible is not going to get "surprised" by time stop Armageddon or spell of mastery? Plenty of those spells can be cast after the ai declares war. Or maybe on hard or extreme if the ai has less cities than the player it declares war (when 3 ai wizards are dead)? Since with less cities peace is unlikely to yield a faster spell of mastery? (or some score here other than cities?)


Also, I think in older versions of CoM if you literally had an army stack on every wizard tower in the game, it would seriously harm the ai's ability to figure out how to use settlers, I noticed it last a long time ago in 2.4x. I would assume 2.7 fixes that issue as well as all the boat issues. I think once the settler's ai is entirely bug free we'll see the AI as a lot stronger, and if not.... well there's always my idea about making ai settlers have double the mapmovement score. This wouldn't have the same flaw as them building double the settlers (settlers cannot find anywhere to go). With double the movement they'll build twice as many settlers, but only if those fast settlers are turning into outposts, otherwise they'll stick with just 2 that are stuck.

[Nelphine]

I'll be trying my next impossible game with 2.72 soon. I'm sad I didn't get to clear nodes with my death game but you're right, the AI is just no threat in that game and getting.. even more very Rares wouldn't change that.

[Nelphine]

So, another strange thing is happening in this game. My opponents are very strong on the other plane (only two of them over there). I'm.. strongish on my plane. Realistically, I don't think I could stop a full blown attack. (Got lucky with personalities though, the sorcery nut is Lawful, and the life wizardress is much weaker, and I think I can hold my own/defeat her relatively easily, especially given the unsummoning discussed elsewhere.)

But of the 6 towers, the AI only cracked one, in, around 1409. Maybe early 1410. It's now early 1412, and the AI have literally never sent a unit through that tower. Nothing is guarding it, not even anything that was here long enough to meet me (I have a city 4 squares away.) At least one of the two wizards has a city adjacent to the tower, and they both have sea traffic near it (I've finally ventured through), but nothing is coming through it.

[namad]

I presume the life wizard who is weak owns the city? and the sorcery wizard who is strong isn't willing to trespass?

[Nelphine]

nope, strong wizard controls it. But it's a hamlet, so I doubt it's been there as long as the tower has been open either way.

[Nelphine]

This is also the first game I've seen behemoths roaming around as rampaging monsters. That was a bit of a shock. Awesome though.

[Seravy]

At least one of the two wizards has a city adjacent to the tower, and they both have sea traffic near it (I've finally ventured through), but nothing is coming through it.

I can see a few reasons for that :
-If the source of the tower is a place where the AI has no much presence, for example an uncolonized continent
-If their homeworld is on your side of the tower then they won't be moving from that plane to yours, only the other way, but from the sound of it you own the continent where the tower is on this side.
-The continent the tower leads to on your side is smaller than 10 tiles and contain no enemy cities so they consider the tower not worth using.