AI

[Seravy]

So, for example, sometimes they'll be able to use the full brunt of their casting prowess in an unimportant battle, but sometimes they'll limit themselves to only 20% skill.

I kinda prefer to keep this based on spell priority since the AI should be much more willing to save mana by not casting Weakness or Holy Armor or Warp Creature than to skip a Disintegrate or Entangle or Web -even in unimportant battles.

The AI, at higher difficulty levels, typically vastly outnumbers the human player in terms of units, and thus the human only has so much fodder to throw at them.

The problem is, if the AI refuses to cast, the human player will not lose this "fodder" -assuming they can afford casting enough spell to win at least the first battle for the turn, if not they won't really be able to afford mana milking either- and can proceed to eliminate another stack as long as it is not killed.
Yes, the AI can afford to lose some units, the keyword here being "some". However the combat casting mechanics pretty much enable the other player to kill an unlimited amount of units as long as they can cast spells, have skill and the AI is passing.

[GermanJoey]

So, for example, sometimes they'll be able to use the full brunt of their casting prowess in an unimportant battle, but sometimes they'll limit themselves to only 20% skill.

I kinda prefer to keep this based on spell priority since the AI should be much more willing to save mana by not casting Weakness or Holy Armor or Warp Creature than to skip a Disintegrate or Entangle or Web -even in unimportant battles.

Sure. I don't think mana milking is that big of a problem, except for WoD. What do you think of my suggestion for that?

[Seravy]

So, for example, sometimes they'll be able to use the full brunt of their casting prowess in an unimportant battle, but sometimes they'll limit themselves to only 20% skill.

I kinda prefer to keep this based on spell priority since the AI should be much more willing to save mana by not casting Weakness or Holy Armor or Warp Creature than to skip a Disintegrate or Entangle or Web -even in unimportant battles.

Sure. I don't think mana milking is that big of a problem, except for WoD. What do you think of my suggestion for that?

Oh I missed the first half of the post, that's what I get for reading forums at 5 am :D

a.) the battle is of average difficulty
b.) the distance is at least 2x
c.) the target is a summoned centaur, a summoned fire elemental, or a non-undead spearmen

would something like that be possible?

Not really, I don't have all that much space left in the combat casting priority calculation - and this is quite a hard list to code. Distance has to be fetched through a far call to another procedure. Not sure what you mean by battle difficulty, if the combat situation, it'll usually be "major advantage" and not "average" since it's 1 spearmen vs the whole army of the AI. If the importance of the battle then I need to save that in a variable from the existing mana saving code somehow. I don't think this works well even if implemented, as it does not cover a lot of cases.

a.) It uses skill too fast. For example, lets say both the player and the AI want to win the battle at all costs, and both have 200 casting skill. The player casts Summon Centaur 4x for 12 skill per cast, while the AI casts Wave of Despair 4x for 50 per cast. By turn 5, the AI will be completely spent, while the player has 152 skill left remaining. At that point, the player can just do whatever - keep casting centaurs, throw nukes, etc.

I agree with this one. It's an independent problem from milking, but true : the player can actually WIN important battles by using this tactic, only undead units, summon 4 centaurs, and then start casting real stuff when the AI has no skill left. The problem, if they do not cast Wave of Despair, the centaurs are actually going to kill their stuff eventually unless everything is invisible or immune to missiles, and at only 12/unit that's a very good deal for the human player. This might be more a problem with Call Centaurs being far too cost effective for what it does. If they kill the Centaurs but not with Wave of Despair, they won't be doing any better in most cases, Black Sleep being a potential exception, a 70% chance of success for 15 is pretty good - but this assumes the centaurs have no resistance bonus which is often not true. At a Nature node success rate would be only 50%, same for Survival Instinct.

I don't think this is as simple as "don't use wave of despair". Even I would use it in many cases because letting the centaurs live is often worse. (Weakness is a better option but the AI can't use that well enough for it, and still many centaurs with weakness dispelled at once = lost battle)

The problem might be the centaurs themselves, a ranged, fast summon for 16 with 6 ammo might just be too good - not in raw power but due to the tactical implications, as it summons a high health threat that is difficult to remove by most available ways.

4.) Yup. That's why I think Wod, Massacre, and Wrack are the only problematic spells.

Massacre has low priority vs a single target so I don't think that is easy to exploit. Wrack is not recastable.

[GermanJoey]

"average" - was how you described lower priority battles:

Right now the AI can only tell apart three cases, "Capital","Important" and "Average" battles.

Hmmm... if that condition is too complex, would it at least be possible to block WoD from casting if the only target is a single battle-summoned creature?

Or... another idea is to give both Fire Elemental and Summoned Centaur Cold Immunity - after, of course, giving Summoned Centaur an additional nerf somehow, as you say. I agree that increasing mana cost seems like the way to go there. Right now, with Conjurer and Specialist (two retorts that are great with Green already), you can summon them for 9 mana a piece. With Archmage and just a few early turns putting power into skill, you can cast 3 of them in a battle. There many uncommon and rare neutral camps that you can kill with just that. You can also survive many otherwise-unwinnable city defenses by saving mana for a centaur until the AI runs outta juice, and then running away with it until the turn-limit arrives. It's a crazy strong spell, especially once you add in stuff like Earth to Mud and Web. Maybe increase the base cost from 16 to 25? The spell would still be very good at that cost, IMHO.

[Seravy]

Hmmm... if that condition is too complex, would it at least be possible to block WoD from casting if the only target is a single battle-summoned creature?

It's easy to do that but I'm afraid I don't even have space remaining to add that right now. I'd need to make some.

I'm not convinced that does anything to improve the AI though.
-there are plenty of legit cases where killing that creature with WoD is the right strategy. Especially if that combat summon is an Air Elemental, but also if doing so enables the AI to end the combat by killing whatever else is left through attacks.
-At the very least "and there are other, immune units" needs to be added otherwise it's doing more harm than helping
-Doesn't help against mana milking much, the AI will cast other spells instead.

I think we should just accept there is no simple solution to this problem.

another idea is to give both Fire Elemental and Summoned Centaur Cold Immunity

That's way too much change for a marginal problem. There is no such unit as "summoned centaur" anyway, so it's not even doable - cold immunity is refreshed every turn on the unit so adding it when summoned is not going to work.

Maybe increase the base cost from 16 to 25? The spell would still be very good at that cost, IMHO.

20 is as high as it can go if we want to retain the ability to summon them using Magicians. Also it should probably stay cheaper than construct catapult.
Running is fortunately not as good as it used to be because the AI will wreck the city if you do that.

[GermanJoey]

Yeah, the AI will wreck the city, but at least you still keep the city.

As far as making the condition summoned units only - yeah, that's not a good idea, I was just wondering what was possible. There's also Earth Elementals.

20 mana for SC is fine I guess. As far as cold immunity not being able to be specific to summoned centaurs - well, what about giving all centaurs cold immunity? How many extra hammers is that worth on the cost of the base unit? Would it be ok if they were 90 hammers? 100? Or what about if all centaurs get dropped down to 3 movement? That unit is already probably too good, could handle a nerf like that, especially if it gets compensated with a minor immunity.

I brought this up because I don't think of mana milking as a marginal problem. I was eventually able to defeat that AI that I was talking about in the original post pretty much soley with this strategy. He was kicking my ass otherwise, forcing me out of many good cities, and I was terrified of sending my remaining heroes against him because of WoD. Starving him down and then engaging him meant I took far fewer losses in reclaiming my territory, and then allowed my heroes to survive a clash at his fortress.

If WoD isn't there, yes, there are other spells, but there are only so many other spells. If I'm going against a single ghoul or skeleton, the AI can only throw so many buffs on that single unit. They will only summon so many zombies. There's only so many global battle enchantments they can cast. I'll say again, while AIs wasting mana and skill is somewhat of a problem, the massive bonuses and discounts they receive make up for it. WoD milking was the first time I felt really abusive against the game (well, sorta the second) and that's why I brought it up.

[Seravy]

I don't think of mana milking as a marginal problem

Mana milking in general is not a marginal problem.
Wave of Despair being used on summoned centaurs is, as it's only one way of mana milking out of, like, a hundred?

and then allowed my heroes to survive a clash at his fortress.

The AI should keep enough mana to always have some left for fighting a battle or two at their capital. Unless you added Mana Leak and Drain Power to the strategy, the capital should not be affected by it.
(which by the way shows the human player will always find a way around the AI, and that's probably how it should be. Most people don't like unbeatable games.)

but there are only so many other spells.

There are like 80 different combat spells and about half of those are reusable. Even those that are not, can be used  X times per spell and a wizard knows like a dozen different spells. Aside from the very early game, I can't imagine a battle where an AI cannot use up all their skill.

 If I'm going against a single ghoul or skeleton, the AI can only throw so many buffs on that single unit. They will only summon so many zombies. There's only so many global battle enchantments they can cast.

Okay, "so many" is way too vague, you should consider the actual numbers. I'm going to use a 10 book Death wizard with spells up to rare for the example.

They can only summon 8 zombies = 160 skill.
They can only cast Wraith Form, Cloak of Fear as buffs, 9 times - once on each zombie and the original unit = 27*9 = 243 skill.
They can only cast Wrack, Black Prayer, Terror, Darkness once and cannot cast Mana Leak (no enemy magical ranged unit)= 40+40+35+25 = 140 skill
They can only cast Weakness, Black Sleep, and 2 Reaper Slashes per centaur summoned = 86 skill.

So they can only use so many skill per combat, where so many equals 543+ 86 for each centaur you summon, if they play pure death. And that's already far more than what they will even have.
Notice how this example wizard does not know any expensive reusable spells.. Casting Invulnerability, Lionheart, Call to Arms instead of zombies, etc would triple the spending.

PS : aside from centaurs and fire elementals, summoning catapults or phantom warriors also works. Summoning zombies also works but in a different way, they can't kill them at all so they'll have to remain in combat and keep casting for the full 25 turns, you only need to make sure you attack something that's not too fast and takes time to reach your zombies.

[GermanJoey]

I don't think of mana milking as a marginal problem

Mana milking in general is not a marginal problem.
Wave of Despair being used on summoned centaurs is, as it's only one way of mana milking out of, like, a hundred?

and then allowed my heroes to survive a clash at his fortress.

The AI should keep enough mana to always have some left for fighting a battle or two at their capital. Unless you added Mana Leak and Drain Power to the strategy, the capital should not be affected by it.
(which by the way shows the human player will always find a way around the AI, and that's probably how it should be. Most people don't like unbeatable games.)

but there are only so many other spells.

There are like 80 different combat spells and about half of those are reusable. Even those that are not, can be used  X times per spell and a wizard knows like a dozen different spells. Aside from the very early game, I can't imagine a battle where an AI cannot use up all their skill.

 If I'm going against a single ghoul or skeleton, the AI can only throw so many buffs on that single unit. They will only summon so many zombies. There's only so many global battle enchantments they can cast.

Okay, "so many" is way too vague, you should consider the actual numbers. I'm going to use a 10 book Death wizard with spells up to rare for the example.

They can only summon 8 zombies = 160 skill.
They can only cast Wraith Form, Cloak of Fear as buffs, 9 times - once on each zombie and the original unit = 27*9 = 243 skill.
They can only cast Wrack, Black Prayer, Terror, Darkness once and cannot cast Mana Leak (no enemy magical ranged unit)= 40+40+35+25 = 140 skill
They can only cast Weakness, Black Sleep, and 2 Reaper Slashes per centaur summoned = 86 skill.

So they can only use so many skill per combat, where so many equals 543+ 86 for each centaur you summon, if they play pure death. And that's already far more than what they will even have.
Notice how this example wizard does not know any expensive reusable spells.. Casting Invulnerability, Lionheart, Call to Arms instead of zombies, etc would triple the spending.

The last time I tried using specifics in a balance post, when we were talking about death magic, you got mad at me and said it was ridiculous to use specifics to talk about a strategy game. 

Here, I note that there's only 25 turns in a battle, so that a wizard casting 8 zombies, 18 buffs, 4 global enchantments, and then 4 more attack spells per centaur, is actually completely impossible. (A wizard can cast at most 25 spells a battle!) Furthermore, that predicates my single unit somehow managing to survive for 25 turns against all this stuff, much less against 8 invulnerable, lionhearted paladins, especially without spending mana of my own. (Normally the ghoul would kill my undead swordsmen after 2-3 attacks). This sort of "slow milking" you suggest also would take forever and be extraordinarily tedious, so I don't know why anyone would ever do it. In contrast, these WoD battles would take me about 30 seconds a piece.

What the hell, man? I don't understand why you're warping the issue so much.

edit: The AI did keep mana for its capital, of course, but it was all the battles before that, where it was drained of mana, let me recover enough to even make an assault on the capital to begin with.

[Seravy]

Here, I note that there's only 25 turns in a battle, so that a wizard casting 8 zombies, 18 buffs, 4 global enchantments, and then 4 more attack spells per centaur, is actually completely impossible.

Good catch. Still even 25*~15 mana is a more than enough in most cases.

much less against 8 invulnerable, lionhearted paladins

Cast flight on turn 1 and you can, easily. Or use swordmen affected by doom mastery that already fly.

This sort of "slow milking" you suggest also would take forever and be extraordinarily tedious, so I don't know why anyone would ever do it.

To win the game, obviously.

I already stated my opinion so I kinda feel we are going in circles.

-Mana Milking is a widespread generic unfixable problem not related to WoD.
-Using WoD on a single summoned unit is often a bad move for the AI but not always and I don't think I can come up with better AI rules for that spell, or other spells that could be used instead. Expensive, yes, but still the most cost-effective way the AI can use to kill the summoned unit. (I don't want to change units for a minor AI problem so no cold immunity to centaurs, phantoms, and elementals)

[GermanJoey]

The only one going in circles is yourself, twisting the issue like a wet towel and then whip-slashing men of straw. But, well, it's your game, dude, you don't need to answer to me. If you're not even interested in acknowledging such an extreme exploit as this, I won't bother you anymore.