AI

[Seravy]

Well, we could use mana instead but mana is basically casting skill*32. So then we are using casting skill.
It's easier to special case Alchemy and tell the AI to ignore this rule if they have it (since gold makes sure they never run low on MP).

True, it would probably make the game harder. But fighting 200 stacks of garbage units when both of us has no MP is kinda not very fun.

Maybe we should instead change how much mana the AI is saving up somehow. Currently it's 32*skill which is enough for like 10-15 battles, after that it's used up and then the AI can only use their income which is like, 2-3 battles a turn. (These numbers assume a 3x distance multiplier)
That's usually enough but clearly not enough for the myrran/arcanus war. Since we can't do anything about the income part, we can only make them save more mana. Under typical conditions (300-400 skill), they already save 10-15k so the best we can do is tell them to double the amount if they are in a "unused units piling up" situation. In other words, if the AI is not hostile towards anyone, double the required MP.

How about this?
It's not really a help for the long term and the AI would still run out of mana faster than the human, but it should mean they get to do an additional 10 battles before running out.

(to be honest, doing interplanar wars is kinda too expensive for AI and human both as is. Spending 1k+ mana each battle when you have an income of 3k and usual turns see more than 3 battles is far from optimal...)

...nah it wouldn't help much. 20k is the max they can store to avoid risk of turning it into gold... And with the level of casting skill they had, they most definitely already use 12k-15k of that...


The greatest problem with this is, they might not have enough to cast in actually relevant battles. also this is a huge advantage for Death wizards who can turn all enemies to undead of the enemy can't cast spells, and then use those undead to further make them waste MP in battles. Zombie Mastery is like guaranteed "you win against the myrran wizard".

[Nelphine]

I'd prefer that yes. Can increase the max to 25k before converting to gold. It still won't be enough, but even without garbage units it still won't be enough. And those garbage units usually aren't strong enough to initiate an attack anyway unless they already meet the criteria you were suggesting; and similarly they're too weak for the human to waste time attacking. Those 3 attacks a turn won't usually be the units you're trying to stop in the first place.

[Seravy]

Yes, but Zombie Mastery entirely breaks that. It adds respawning targets that the AI can and will attack, each and every turn. (to some extent all Death spells that make undead in battle do it, but that's the biggest offender)

Also they sure can attack. I mean if I can't afford a bigger stack, I block towers with a single Night Stalker because why not? And then it gets hit by 5 stacks a turn.

I don't think I can safely increase gold since Alchemy is used based on gold to mana ratios, not preset values. I rather not risk it. (It would be nice if we could hold more than 30k mana to be honest, that's what the AI really needs. But we can't do that, even if we make the value unsigned it's only 60k but even doing that is not realistic to try.)
Unless....if we had some sort of an overflow variable and made the game automatically take away all your mana over 20k and put it into that variable, and then readd it if you drop below 20k and for all displays for mana, show the sum of the two numbers... then we could have like unlimited amounts stored. (would need new AI alchemy rules, ofc)
Ok, that's actually an interesting idea... I don't think the "display number" calls are even capable of displaying numbers over 65k though, assuming the input is an integer type... no the game doesn't have a procedure to display a non-integer as string.

[Nelphine]

Still, even up to 65k would solve most games. And if you did go higher, you could just display the max.

[Seravy]

Well I can add it to a "todo" list but trying to convert all signed references of gold and mana into unsigned would be time consuming to do so it's something for later.

(more importantly I'm not sure if the function that displays the number uses signed or unsigned. If it uses signed, we'd see negative mana and gold...)

[Seravy]

Now that I think about it, saving an extra 30k mana is equivalent to missing the research of two very rare spells. That's an amount that impacts even the AI. I don't think it's realistic to expect anyone to need that much mana. I have to think the number of units is the problem afterall.
Let's say you own 30 cities each at 50 production. That can produce 10 cost 150 units. A full stack, so one battle worth a turn. So you need to save 1000 mana a turn to be able to fuel the spellcasting if you're an AI (but probably at least half of that even if you're a human). Well, I suppose that's about the mana your cities are producing so it's balanced, but there definitely is nowhere to store the amount, and honestly that means for each 15 turns you could have a very rare spell if you haven't built and used that stack.
Problem is while the AI does have double the power, they also produce double units, and spend double in combat (since they are bad at holding back on spending), so overall the AI does use up more mana than it produces if they go full unit and mana production. Other problem is, they do produce these units even during peace, so basically they are always in "military spending" mode on the troop side but not on the mana side (since they can store troops but not mana. But if they could store mana, they'd miss out on research and raising skill so that's not really good either.)

Note this is just guessing and not actual calculations. Maybe cities are balanced in their output, maybe not, idk.

So let's try to figure it out. Assuming 9*150 = 1350 hammers is a stack of units and a combat costs 1000 mana. Means 1.35 production to power ratio needed to have 100% funding.


Pop 20 city, 85% production bonus, 8 farmers, 12 workers, buildings is 70.3 hammers.
Same city produces 6+4+2+10+6=28 power.

Even if I assume there are nodes and other power sources, the ratio is a lot worse, 2.51 normally, probably around 2 considering other sources of power. (power and production multipliers for AI are similar, production is a bit lower but not by much)
Only through resupplying it by alchemy, or having channeller, can the AI reasonably keep up, unless they have some really big extra power spell (rare or very rare).

So basically cities produce more units than the mana you'd need to use them up in combat with full spellcasting, and even if they didn't there is nowhere to store this mana and it's probably better worth it to not produce as many troops and spend on research and skill anyway, at least before finishing very rares. 

Now this is of course a worst case scenario - if battle cost less, as they end quickly with either side's victory, we are good. If they don't, or the human forces the AI to play extra battles through, say, zombie mastery, the AI is screwed either way.

So I don't think we can do anything, but we could maybe still tweak the AI to produce less idle troops during peace situations. (which makes their research and skill higher which might actually help them more than the useless troops. Albeit like half the "useless" troops I encounter are adamant Minotaurs and manticores, so maybe not really. Those are actual top tier units...I just play death so they feel like garbage to me.

heck, I think the problem is not the AI, Death is just too powerful against it. Or just too powerful in general? I mean, I can kill an entire stack for like, 80 mp and 2 zombies, by casting Possession 3 times...
I often notice I could use Wrack to kill an AI stack but I'm like, "sure, but 3x possession costs less than wrack+black prayer+a way to stall for turns so why bother" and that's the main problem of Wrack - it needs low resistance and it is supposed to be the "cheap, kills all" option, but if the enemy has low resist, Possession kills them all cheaper...
)

[Nelphine]

Sounds right, re: death.

Of course, I consider death the second weakest realm despite all that.

So I think it's fine.

It's not that death I'd overpowered, its that death is designed for that playstyle - overwhelm with cheap units. That's almost exactly what your description of the death realm is in the first post.

Play sorcery and it's even worse with spells like invisibility or magic immunity. Play nature or life and he mana they spend literally doesn't do anything. Play chaos, and even if they spend that much mana, you're designed to be even more effective.

Any realm at that level destroys things and soaks up AI mana like a sponge.

[Nelphine]

What's the priority of guardian wind based on?

Specifically, when facing things like slingers, I've been seeing djinn casting invisibility before guardian wind. While invisibility protects against offensive combat spells, and provides defense even if something has illusions immunity, guardian wind would completely stop the slingers from doing anything.

Against a combination of slingers and heroes, the djinn actually attacked as the first action (attacking the weakest hero), which resulted in the slingers killing the djinn on turn 1.

Can guardian wind priority be changed to be based on the strength of the missile attacks it would stop? 

In these cases, slingers are far more dangerous than potential combat spells, or heroes.

[Seravy]

Invisibility is enemy ranged/4 + type B + Capped at 80.
Guardian Wind is enemy missile/4 + type B + Capped at 66.

Invisibility blocks missile, magic ranged, and reduces melee (and sometimes prevents spells) as well as enables stalling and is harder to dispel (or target with dispels).
Guardian Wind only blocks Missile and has a dispel resistance of a mere 10, so it's strictly inferior in most cases.

Attacking or using MP is a random choice, the weight depends on attack strength vs MP amount.

[Nelphine]

Hum. I suppose since I know exactly how to counter, it seems off to me. Ugh.
I don't know, slingers are supposed to be blocked by guardian wind, but I'm just trashing everything. I'd go so far as to say it's actually stronger than bezerkers. Then again, maybe efreet would be better against them (even though usually it's djinn that are stronger).
Ai can't prioritize based on my spellbooks right? (I have true sight which is why invisibility is throwing me off. But at least invisibility I can accept. Attacking/psionic blast when faced with lots of adamantium slingers? That was literally saying 'kill me before I act again'.)