Raise Volcano

[zitro1987]

Volcano's 17 rate of return (a bit higher, closer to 20 because volcanos revert back) is on paper rather impressive, but you have to consider the research cost and the lost skill usage that could be used for other spells and summons. You will be inclined or forced to eventually use other spells, limiting your volcano spell casting to about 25-50 volcanos based on game factors.

So I still think 50-60 mana is the most appropriate spell cost.

[Nelphine]

Zitro at 0.5% chance to turn back into a mountain, you expect to have it for 200 turns - the return on investment is still less than 18 turns even accounting for that.

The BIG reason I consider this spell a problem is spammable nature + ai overland casting skill bonuses. Even with expert, the ai only spends 25 skill to cast this spell. This makes the spell into an economic powerhouse. Unlike summons (that have a maximum per stack therefore a maximum per target, and that the AI must use for garrisons in many cases) and city curses and buffs (which have a limited number of targets), raise volcano can be spammed.

This not only can result in an entire islands covered in volcano in 1405 or 1406 (which is a problem discussed elsewhere) whereas the human probably can't afford to set aside enough skill for that until 1412+, the AI also gets extremely good economy from the spell that is only limited by the number of tiles around enemy cities the AI has visited - on advanced plus, this is effectively unlimited. (And powerwise, the so can easily create a node per enemy city, on max power; on lower power settings, it's much better than a node) which is easily enough to sustain the AI even if you banish them and take away all their nodes/developed cities.

Thus where my suggestion for 65 cost comes from.

However, just the fact that this spell can be used to decimate entire islands whether by human or ai, also leads me to believe it is undercosted.

[Seravy]

spammable nature + ai overland casting skill bonuses.

As stated in the initial post the AI won't cast the spell less if it costs more. Likewise it won't cast it more if it costs less. Not by a relevant amount, anyway. The percentage chance of picking the spell only depends on the personality.

The AI does not use curses freely, you know that.

So the only reason for raising the cost, if we have one, is the human player.

[Nelphine]

A maniacal mono chaos without corruption or uncommon Summons (which sounds strange but the last three chaos I've played against missed all the uncommon summons), who does have raise volcano, will throws huge number of volcanoes. They do in fact throw enough to cover islands, and to still have tons of them even if they get banished 150 turns later.

Changing the casting cost doesn't change how often they throw them compared to any other spell, but if it's one of the only overland spells they know, it makes a significant difference in how many spells total they can throw.

For the human, it's still so much better than corruption as to be ridiculous. (I'd like corruption to go to 35 at the same time as raise volcano goes to 65 except that for the same maniacal chaos wizard, that's just obscene. It's bad enough seeing 40+ tiles corrupted in 1405, it can't be even cheaper.)

[Domon]

vulcano is not an unlimited growth strategy. at some point, you start losing volcanoes and have to keep spamming them to stay even.

[zitro1987]

Then the alleged issue is a high AI priority to cast volcanos, not it being overpowered for humans.

[Nelphine]

No, the issue is economic power per cast, compared to cost of casting it. Given that each one averages 200 turns, it's effectively unlimited in terms of duration for practical purposes (yes of my ~250 cast, 4 have reverted. Not significant.)

The second issue is overpowered compared to corruption.any other similar nature spells (famine vs plague, firebolt vs lightning bolt, holy armor vs iron skin, chimera vs great drake), the more powerful spell costs more than the weaker spell, and the cost is comparative to the power difference, although the stronger one gets a reduction (10-20%) due to being higher tier.

Raise volcano, without any economic bonus, is at least 50% better than corruption. It's irreversible. With the economic bonus, it's at least 100%, probably closer to 150% better than corruption. 

Yet raise volcano is only 25% more costly than corruption. The price difference is simply too low. Corruption is reasonably priced. Therefore, raise volcano is underpriced.

[teelaurila]

Alternatively, why the 200 turn (average) longevity? Make it 50 turns, and you got 150 power back from 50 invested. On par with Drain Power (~200-400 from 100 invested).

[Nelphine]

For me, I actually thought the same. However, I think the longevity and the economic power is important, specifically to differentiate from corruption, and to justify the higher tier, higher research, and higher casting cost.

While you could reduce the casting cost (more accurately, leave it where it is now) if the economic power or longevity was reduced, its harder to justify why it's different from corruption in that case. So I think leaving the economic power and longevity where it is, while increasing the casting cost, is actually a better choice.

[Seravy]

Let's do the math.

You pay X mana and X casting skill to cast a volcano where X is the casting cost.
We want it to have a "return of investment" of 50 turns, past that it should be generating profit.

X mana costs X, obviously.
Casting skill however is a renewable resource, X casting skill, if obtained during a 50 turn time period, means skill has to be raised by X/50. Assume the casting skill of the player is Y, it costs 2Y to raise it by one, so XY/25 for the required amount.

So the total cost is X+XY/25, this is how much power we want to be generated in 50 turns. Each volcano generates 3 power.

Meaning we want X+XY/25=150.
X(1+0.04*Y)=150 , X=150/(1+0.04*Y)

Which means that, the cost actually depends on which casting skill we want this to be this profitable at.
I'd say we want it to be profitable up to 100 casting skill, past that it's far too spammable and the economic impact on enemies should outweight the own economic boost there.

Which would require X= 150/(1+4) = 150/5 = 30 casting cost. Eh.

Okay, maybe the return of investment at 50 turns is a bit too good? Past that point we do have 1 higher casting skill permanently, and a power source of 3, which is quite amazing from a spammable uncommon. Let's assume 100 turns instead.

Then we get : 300/(1+0.02*Y) = 100

So assuming our acceptable RoI point is between 50 and 100, the fair cost is between 30 and 100. Well, that didn't help much in making the decision, it's no different from saying "anything goes if it's reasonable".

At least we have proven that whichever we pick (in our planned 50-75 range), it won't break the game.

Since the "main" effect is fine no matter what, we might want to look at the "side" effects instead.
Destroying an enemy mineral sounds impressive but it is pretty horrible compared to most other city curses. Even a +10 power crystal pales in comparison to an Earthquake or Famine or Evil Presence. So this part of the effect isn't worth all that much. Except for one. Destroying Mithril/Adamant/Orihalcon. That is an effect to counter Transmute, but it's a hard counter as the destroyed ore cannot be transmuted again. So while counters generally need to cost less to be worth it, in this case that isn't true. So this part isn't giving us new information either.
Creating a new own mineral is also a thing. The chance is only 20% though - you need to cast the spell 5 times to guarantee getting one. Probably no matter how low the cost, it won't really be worth doing that, even for Dwarves.
Sadly we can't raise the mineral chance - the volcano is more often placed in enemy territory than not, especially through Armageddon.

tl; dr, we are free to pick any cost we want between 50-75 and it would work fine.