SORCERY Realm

[Nelphine]

I) recall difference in casting skill based on difficulty, primarily for the human, but also for the ai. Consider your game vs mine - by 1420 I can 4 or 5 times your casting skill. And this is primarily based on style of play. Therefore I dislike casting skill as the basis, unless we do something like 2 spells plus an extra one for every full 600 casting skill the wizard has, or something else equally huge. 

I do like basing it on research value of spells though, as this allows choosing something that someone else has even if they haven't researched very rares. However, basing it on research also means that if you can get one strong very rare, like crusade, you could probably get aura of majesty, just cause, aether binding, doom mastery, survival instinct, and divine order instead. That kind of goes against only getting a few spells. 

Overall I prefer a set number of spells, but the first uncommon or common spell you bind does not count against the limit.

II) I like A - I agree not C, I dislike B (I don't think it should give you instants), I'm hesitant about D (curses still fall here, but even just stealing inspirations or Gaia's blessing is completely cross realm), and even though E lasts entire combats, on an overland scale that's still instant, so I'm against E for the same reason I'm against B.

III) ugh. I dislike B because that could be a giant list, especially if 
part II goes for one of the bigger options . However, I agree with disliking random, as that messes up part I (regardless of whether Sorcery should be random or not). B is also not going to be ai friendly as there could be spells that particular great synergy, but how would you prioritize them all across all 4 realms, especially if you take a larger option from II. Unfortunately due to the ai unfriendliness I think it has to be random.

IV: I think it should be B, but again that leads to the random option for III. Not sure I like that. Without random, you could go for D here, but that forces A for II. Or does it? Well from a coding point of view it probably does, but one can imagine clicking on a city or node or unit stack and then unit, and then choosing one of the spells an opponent currently has cast on that target, and that spell would be added to your spellbook - but how do you track it to remove it from your spellbook if you want to use spell binding on something else (either instead of, or in addition to).

[Seravy]

but how do you track it to remove it from your spellbook if you want to use spell binding on something else (either instead of, or in addition to).

What?

No, the spells you pick are permanent. You can never replace them, you can only pick more if I. allows.

I don't see why IV.B would require random. While it'll be difficult to draw the line where the AI needs to "hold" and wait for a better target, it's doable, probably something like, pick the best spells from the list of all spells available for other players (including unresearched) and only cast this spell if a researched one is among them.

for III. I think the most straigthforward for the priority would be to use the "AI trade value" - higher means it's a spell that benefits the AI more. It does ignore synergy, but I can't really recall much synergy among non-combat Sorcery spells and other realms, especially global enchantments. We probably need to explicitly disable a few choices that assume the wizard plays their own realm, such as Chaos Surge, but that's about it. Synergy between multiple copied spells is only relevant if the formula allows copying a large number of spells (typical wizard getting at least 3), and our choice for IV. is not "b" or "d". I think this is a very unlikely outcome, especially the "3+ spells" part.

Good observation about I. needing to work correctly even for spells that aren't very rare - you could literally take 5 rares for the cost of a good very rare, or 15 uncommons. So it likely needs to be a cost based on spell tier (1 for common, 2 for uncommon, 3 for rare, 4 for very rare) or a flat number (1 spell for each 200 skill).

There are 35 heroes in the game so (at least in theory) the menu should work for up to 35 entries. Meaning for II we need a condition that results in 35 or fewer spells, which means A, B or E.
Alternately we can add "very rare" to the condition, which means anything goes, and also eliminates the problem with I. and different tiers. I think we can agree we don't want C or D either way.

IV.D has a critical problem. If the AI casts the spell and the thing it wants to learn is cancelled or dispelled, it would learn something else and permanently waste their chance.

So...updated :


I. Number of spells you can obtain this way should be limited somehow. Usually more than 1 but always fewer than 4. We'll need some mechanism to control this, and some formula (or a constant number).

II.The type of spells you can obtain should likely be limited to certain spell types :
a. Global enchantments only
b. Global enchantments only (Final Wave and Great Unsummoning included)
c. Overland spells only
d. Noncreature overland spells only.
e. Global enchantments only, but including combat global enchantments (Call lightning, High Prayer, etc)
f. Very rare enchantments only (global or not, would include Pestilence, Regeneration etc but not famine and co)

III. The spell that will be obtained can be
a. Random
b. Chosen from a list of all valid choices

IV. The spell can be
a. Any spell
b. Only spells any opponent already knows
c. Only spells any opponent knows or can research in the future.
d. Only spells any opponent has already cast (only applicable if IIa is picked)
e. other?


Let's try to narrow down II-IV to one choice each, as that makes it much easier to design a formula for I.

[Nelphine]

What menu displays up to 35 entries? But yes that would help.

II) firmly on A of those choices, but I can understand other options. However, would E even meet the 35 options requirement? I really dislike E in general. I don't much like F, as it seems super unintuitive since you can easily imagine utustions where lower tier enchantments are better.

IV) Firmly on B. Main reason is replayability (choosing A, you pick the same spells every time, in which case why call it Spell Binding instead of just calling it the spell that is determined to be the best?) and in game balance (in my opinion, not fair at all to get a spell based on someone else.. before they get it.)

[Seravy]

What menu displays up to 35 entries? But yes that would help.

Resurrection's should be able to, as you can have that many dead heroes. Well, I guess 34, as Chosen is not included.

However, would E even meet the 35 options requirement? I really dislike E in general.

Probably. Considering we specifically excluded dispelling combat globals from Sorcery, maybe we shouldn't let it copy them instead. It's not the same thing but...way too close.
So I guess we can exclude E here.

Also agreed F is way too counterintuitive.

So that leaves only A or B (tiny difference but it's there)

IV.A, you have a point, I would always pick the same thing, too. (probably Armageddon every game, unless I can't afford the diplomacy penalty - wait, with Sorcery that can't happen lol.)

So, assuming it'll be

Spell Binding
Cost X
When cast, choose an overland Global Enchantment any other wizard already knows. You can cast the spell as though you researched it. You can't gain more than a total of Y Spells this way.


So now we'll need an X and a formula for Y.
...and decide between II A and B. I think A is better. Great Unsummoning is in the same realm, and Final Wave is something Sorcery is specifically unable to do (kill nonsummoned units instantly).


...problem. What if the human player intentionally does not research any global enchantment? Wait, that's the same as the current Spell Binding, it doesn't matter. I think this should be a solution then, if we can get a good definition for Y. Something that's reasonably intuitive, doesn't break on different rarity tiers and guarantees at least 1, but at most 3 (I guess maybe 4 in extreme cases like 800-1000 skill AIs) spells, of the "most valuable" kind.

...btw this will have a new functionality old Spell Binding did not have : It will be able to "steal" Time Stop.

...we might also need a new name for the spell, as it's not "binding" an enchantment, in the sense that it's neither taking away one from another player, nor does that player even need to cast it for stealing to work.

[zitro1987]

I like how this new spell is turning out to be. I would probably put the limit around 3 and the cost around 800-1000.
*Lower than 3 would nerf the original uber-powerful spell too much
*I like II A now that it was decided to be a simple 'choose a global' approach.
*I agree that the other wizard must already know the enchantment


As for name: Here's my attempt 'Pansophic Learning' as it covers all the following themes:
-Pansophy roughly translates to omniscience (word already used in retort)
-learning (you are learning something new, in a way that is less bounded to your realm, hence the omniscence)
-Combining the two implies 'all-knowing'
-We are talking about global effects, so the omniscient theme carries.
-Word attached to multi-century ago, kind of fitting the 1400s of game

[Nelphine]

I still like Spell Binding as a name - you're just not binding a precast spell, instead you're binding the very magic that allows you to cast the spell at all.

Casting Cost 750

Number of spells: 1 for every 300 skill or fraction thereof? (so 1-300 skill is 1, 301-600 skill is 2, etc) That should put most humans at 1 or 2 to start, and very high end humans would get 3. AI would probably have 2 to start, and on higher difficulties get up to 3. Anyone who reaches 4 is completely crazy anyway.

If you want to get fancy, give the player 3 spell points for every 300 skill or fraction thereof. A common or uncommon spell takes 1 spell point (I'm purposefully lumping commons and uncommons together; if for some ungodly reason you really want to get a common, you shouldn't get a further discount past the uncommon discount as commons really aren't relevant at that stage and don't need to be encouraged), a rare spell takes 2 spell points, a very rare spell takes 3 spell points.

Alternatively, 1 spell point for every 75 skill or fraction thereof, commons or uncommons take 2, rares take 3, very rares take 4. However, you may end up not able to spell bind a very rare spell with this, so I'm not sure I like it as much.

[Seravy]

There aren't that many globals so we can look at them case-by-case to decide on a cost. This is how much I would prefer to spend on each (assuming 3 = 1 very rare spell)

Eternal Night - 3
Evil Omens - 3
Zombie Mastery - 3
Aura of Majesty - any (by then diplomacy should not be an issue, but if it is, even 3 is worth it)
AEther Binding - 2-3 (double dispel power is worth a lot, free skill allows getting more spells)
Suppress Magic - 3
Time Stop - 3-4
Fairy Ring - 3
Seismic Mastery - 3
Herb Mastery - 2-3?
Chaos Surge - 0 (not worth stealing, works if cast by enemy)
Doom Mastery - 2
Doomsday - 3
Meteor Storm - 3
Armageddon - 3-4
Enligtenment - 2-3
Life Force - 3
Crusade - 3
Just Cause - 1
Holy Arms - 2
Divine Order - 0 (not really worth stealing, albeit it stacks, it does provide the same effect already in play)
Charm of Life - 3
Detect Magic - 0 (you have it already)
Survival Instinct - 2-3

So I think making every spell have the same cost is fine, we'd just not target Just Cause with it ever...the rest are usually worth the price with Holy Arms and Doom Mastery being somewhat subpar (but still worth it if there isn't anything better to get and you don't want to wait fr enemies to learn something better).

My preferred system is thus : 1 spell for each X casting skill of your wizard, rounded up. X should be a number in the 200-300 range, I'm leaning towards 200-250.
300 feels excessive - I never even go above 500 skill unless I'm playing on a Huge map so 600 seems pretty much unreachable in most games. (yes, Spellweavers might have an advantage here but that's a special case)

So, only one very important thing left to decide.
What algorithm the AI should use to decide to cast or not cast the spell, what priority for casting it, and what target to select.

[Seravy]

Looking at the code.
The Spell Binding procedure is in the same segment as the one that calls the Resurrection menu so it should be safe to use.

For the AI...


1. Never steal
Detect Magic
Just Cause
Chaos Surge
Divine Order

2. Steal or wait for a better spell
Everything else?

3. Always steal immediately when available
Time Stop
Meteor Storm
Armageddon
Charm of Life

So...the hard part is this :
The AI should wait for a better spell, if another player is expected to research a significantly better spell in a reasonable amount of time.

Problem : How do we define "significantly better spell", and "a reasonable amount of time".
I guess for the former we can make a list of priorities, and say "if priority is at least X higher than the other spell", so that part is easy. But how can the AI know if it's worth waiting for another spell to be researched or not?

[Nelphine]

As a note: chaos surge, divine order,holy arms, and doom mastery can all be amazing to steal for dual realm wizards, although realistically you should try to trade for them in those situations. So I am convinced, yes they should all be the same.

I'm assuming that it would be based on base casting skill so that neither hero skill nor spellweaver nor AI overland cost reduction would affect it (making archmage a little more attractive to sorcery wizards). I'm also assuming when you say every 200 or 250, you mean round up (so 251 is 2 spells).

Part of why I'm suggesting 300 is because it's reasonable to assume that a sorcery wizard will have Uranus blessing. That means on fair, assuming 3 wizards survive to very rares, and each controls ~2/3 of one plane, we can expect the casting wizard to have ~25 cities. That's 275 casting skill just from amplifying towers. We should also assume aether binding. That means the human can reasonably expect to have 400 casting skill, and an AI can reasonably expect to have 400-600 depending on difficulty. 

Even at 250, an AI would often start with 3 spells, which seems too high. At 200, a human would start with or be almost at 3 (only needing 401 casting skill for 3).

With only 2 wizards left, that number goes up so that at 250 but high difficulty an AI could start with 4. Moving to huge, a human, at 250, would be pretty close to starting with 4 (550 just from amp towers). On huge, an AI who controlled one whole plane, could reach 5, which is absurdly powerful.

That seems too many spells to me. The only time anyone ever has 1 is if their empire is small given they've  reached very rares AND they don't have Uranus blessing. I don't think you should balance the spell around that.

[zitro1987]

Cost: 700. 1 spell for each *200* casting skill of your wizard, rounded up.

1 Never Steal
Detect Magic
Just Cause
Chaos Surge
Divine Order
Aura of Majesty
Aether Binding? (most likely has this spell for research)

2. Steal or wait for a better spell
Doom Mastery
Eternal Night
Zombie Mastery
Suppress Magic
Seismic Mastery
Herb Mastery
Enlightment
Holy Arms
Survival Instinct
Evil Omens / Doomsday (may be 'always steal')

3. Always steal immediately when available
Fairy Ring (yes, you have fewer summons to choose from, but you will have lots of storm giants/djinns/sky drakes ... and getting air elementals is interesting)
Time Stop
Meteor Storm
Armageddon
Charm of Life
Life Force (extreme increase in power)
Crusade (self explanatory .. surprised you didn't list it)