SORCERY Realm

[Nelphine]

Oh - you only mentioned specialist, not runemaster in the original 9000 comment. Yes, i get where you're coming from now.

Make those additive instead of multiplicative. (So runemaster + specialist = 4500 * 1.5 for disjunction = 6750.)

But I specifically do NOT think that sorcery needs the only spell that counters specialist/runemaster; anyone can face that. Thus my opinion that both need to be toned down so that not only do you not need spell binding, you don't even need aether binding. (Making both 50%, and additive, would mean that charm of life with both would only be 3000. Which a non sorcery wizard at least COULD disjunction. Then make Aether Binding 50% as well, and now sorcery almost has 100% dispel rate against it, and then spell binding is a completely moot point.)

[Seravy]

I do think it should be only one spell that can counter that. The main selling point of picking Runemaster is to make sure your spells don't get dispelled so it should not have more than one counter. Whether it should have ANY counters is a good question though. The human can Spell Blast globals so they don't need it, but the AI is not allowed to do so, meaning they need Spell Binding to do it.

(speaking of which I realize we raised the dispel resistance bonus of it to 150% half a year ago. So in reality it's not even that much but more. 2.5*2*1500*1.5 = 11250. If it was an additive bonus, it would be 7875.)

As we would end up with the AI dropping enchantments and holding back no matter what we pick it seems, we might as well go with the trivial solution of "For each stolen enchantment, <insert drawback here>".
Loss of casting skill, added maintenance (we could try to make this bypass normal AI maintenance discounts), whatever else you can think of. I rather have that than have the AI cancel spells for no reason with the human side being unfixed. Then we can teach the AI to only steal things if they can afford to do some (will be hard but STILL easier than adding the "1 enchantment" limit.).

A few ideas for possible drawbacks :

Slower research for each stolen enchantment (questionable, by this time you might not need research)
Less casting skill (this might be too hard on the AI and would be very unfun though)
Less power income (AI compatible alternative of the "higher maintenance". However we need to consider what happens if power would drop below zero. Can we allow having any number of enchantments in exchange for not gaining any power or not?)
Something else?

This would STILL result in the AI potentially ending up in a "steal then cancel" loop and failing to obey disjunction priorities, though. So it's STILL a very bad solution. Anything that involves the AI cancelling the spell risks the AI trying to steal it again, I don't think I can make it predict what will be kept at a 100% accuracy (or any accuracy at all if the condition to cancel depends on something that can change over time. Diplomacy, AI income, etc...)

[Nelphine]

I disagree that a single realm (even sorcery) should get a counter to a retort. Instead, I'd much rather let a realms have the chance, and the retort just makes it expensive.

Given what you just said, I'd change specialist to +50%, runemaster to +100%, and aether binding to +50%. Make runemaster and specialist additive. 

Then remove the dispel ability of spell binding completely.

Dispel is already a huge 'unfun' issue let alone a balance issue, so reducing the sources and the variance on those sources would be huge.


I do agree that ai shouldn't have to accidentally cancel and switch spells. That's why I don't ever want them to cancel. Don't teach them that. Change spell binding research priority to really late, then most things will be cast already. AI just captures based on priority, and never cancels. If someone does get a new better spell, oh well. That should be rare enough not to matter. And if they're actually taking spells with that really low reearxh priority, they should get strong enough ibes that others will dispel them naturally, allowing them access to the better ones if someone manages to get a better one later.

Then all you need to do is figure out how to show the human which one of their own spells were captured and how many more they still can.

[Seravy]

I disagree that a single realm (even sorcery) should get a counter to a retort.

Spell Blast says Sorcery inherently beats Runemaster. The AI isn't allowed to use Spell Blast for it. So it uses Spell Binding instead. At this point there are three options :
1. Runemaster becomes superior for humans than AI. No retort or realm should work like this if it's avoidable. (Yes, in case of Artificier it's unavaoidable but that's unfortunate)
2. Let AI use Spell Blast (obviously don't want to)
3. Let the AI use Spell Binding.

3 is the least wrong choice of these.
Overhauling retorts is off topic. It's unrelated to Spell Binding. I'm not going to change retorts for Spell Binding, it should be the other way around which means it is unrelated. Changing retorts for a single spell just isn't the right direction of causality.
Considering the AI is holding back more on Disjunction than last year, a weaker Runemaster resistance bonus might be acceptable now but I don't think it's that necessary. There has been absolutely no changes that warrant a weaker dispelling power bonus on anything. I'm strongly against ANYTHING that reduces dispelling power in the game due to buffs. Unless you suggest the dispelling power to be different on disjunction than on other dispelling spells but I prefer to keep it matching because that's easier to remember and makes more sense. Even for Disjunction power we already decided the AI should be handling that through their priorities, not the game mechanic itself, and I have never heard anyone saying Disjunction power is too high for human players. Unless someone actually believes that, there is no reason to nerf it. (and "nerf dispelling wave" is already on the refused suggestions list. Time to add "nerf dispelling power" in general I guess.)

Then all you need to do is figure out how to show the human which one of their own spells were captured and how many more they still can.

Which is absolutely impossible (and we MUST let the player know the same about all wizards too. You might not care for that information but I do.) This direction is already known to be a dead end, and thus any sort of numeric limitation to the number or value of spells that can be held at a time is also a dead end. We need to go in another direction otherwise nothing will happen.

[Nelphine]

I don't understand. Why would you need to know enemy wizard powers? You don't know what they summon with fairy ring. You don't know what they kill with great unsummoning or final wave. You don't know what unit they buff with any unit buffs. They keep track of their own, you keep track of your own. Regardless, I think you are right, the human can't even keep track if their own.

As for spell blast countering: that's a nice theory on paper. In practice its garbage. The AI on expert or above can instant cast most globals by the time they know very rare spells. Oh never mind, I forgot we specifically nerfed them a bit on that. Still, detect magic plus spell blast comes into play earlier than spell binding, and the list of things that the AI would want to dispel, are also on the list for things it will spell blast. So yes, spell blast still counters runemaster just fine for the AI. Therefore they don't need spell binding to dispel. 'But disjunction detect magic!' Of course. We put this in so that the human COYLD cast and keep things. Giving the ai spell binding dispels directly contradicts why we prevented the AI from using spell blast intelligently in the first place. Therefore, we don't need to give them spell binding dispel. (More importantly, I still have disagree as to why we think ONE realm needs extreme extra dispel ability. If dispelling globals is important, then its important for all realms. Letting one realm do spell blast, enhanced normal dispelling, enhanced disjunction, further enhanced dusjunxtin, and aoe/overland dispelling might SOUND thematic, but gameplay/fun/balance, its awful.)

 But that means if we want to keep the same concept for spell binding at all, its either 'one spell not in your own books, and if you use it again the current one is lost' or 'you can capture as many as you want'.

I don't think we have Anything in between. 

The first option,I think is not strong enough. Why would I research and learn a spell to get one global if someone else happens to get one that I can't AND I care (chaos surge when I have no chaos books? I don't care.) I'd much rather just get my own global to cast on my own.

Therefore, it has to be capture as many as you want. 

Therefore no change is needed on mechanics.

If I take just cause and aura of mastery, I don't want to pay 500 mana per turn for them. 

So I'm against any extra maintenance costs based on number of spells captured.

So, the only thing I think would be reasonable is increasing the up front cost.

[Seravy]

Why would you need to know enemy wizard powers?

I can only judge the "safety" of casting my enchantments if I can know if they want to steal more or not. If I can't, I might as well not cast them. And we are where we started "Do not cast any global enchantment if there is a Sorcery wizard in the game."

Not that letting the human pick an enchantment to cancel on the resolution of Spell Binding can be solved anyway.

Giving the ai spell binding dispels directly contradicts why we prevented the AI from using spell blast intelligently in the first place.

Except Spell Blast stops every spell and costs nothing. It's a lot more powerful, so the AI needs to hold back on it more. Disjunction is still too powerful to let the AI use freely, but it's not powerful enough to completely disable the AI on using it like we did for Spell Blast. The AI is still using dispels, and for high priority targets, fairly often even, just in a way that does not allow the AI's resource advantage to show through it. Basically, the AI is allowed to dispel but isn't allowed to do it better than the human. Your suggestion would mean the AI would be actually worse at dispelling in a SpellBinding vs Runemaster scenario than the human.

Heck, even without retorts, a Charm of Life is 1500. It costs 2250 to dispel. You can't put that much slider into Disjunction so you dispel it with Spell Binding. (Unless you have Aether Binding. But 2 realm AI might not.)

Anyway...

So, the only thing I think would be reasonable is increasing the up front cost.

This has already been deemed impossible. Costs are determined on starting to cast the spell but targeting happens on resolution. So the cost cannot depend on the target unless you have a Time Machine. Which would be very fitting Sorcery but I unfortunately can't program one yet. However the AI can't decide between Disjunction and Spell Binding based on the cost. So it would use "up front 3000" cost Spell Bindings to dispel your Detect Magic, AEther Binding or Just Cause. That's no good. It would also make it even less viable to steal cheap enchantments than 'and it has a lot of maintenance". Which I don't see where you have a problem with, if the maintenance is proportional to the spell's original cost then there is no problem. Just Cause? Instead of 3 you pay 60. Armageddon? Instead of 40 you pay 800. Aura of Majesty? Instead of 12 you pay 240. Ok maybe this has an overpriced upkeep to begin with, idk. But anyway the idea would be to have the cost depend on the "value" of the spell.

[Seravy]

Trying to summarize.

We discussed :

1. Remove the spell completely
Rather not, would be a huge loss to the game and might even make Sorcery slightly underpowered. (They don't have that much globals of their own, actually.)
2. Make it not remove the target (Copy Enchantment)
Does not fix the core problem (too many things stolen) and would even escalate it in cases (AI stealing from other AI)
3. Pay a higher cost up front.
A high cost would make the spell work very poorly against low/medium cost targets. So a variable cost depending on target is needed, but that's impossible because it would require knowing the future.
4. Make the AI hold back on using it
Still leaves it overpowered for human players. Would cause the AI to need losing the ability to use it for dispel which I don't want, or would require the AI to cancel excess spells which looks stupid and I can't justify it. But in either case does not solve it being overpowered for the human and it sure does because the AI is doing only a fair amount of dispels so the player gets to keep what they stole.
5. Make it have a hard limit on the number it can "hold" at a time
Can't do the interface for human players so this is completely in the realm of impossible.
6. Make it have a cost for holding spells.
AI would need to learn to cancel spells. It will mess up and go back and forth between stealing something then cancelling it, as it can't see the future. Which results in potentially losing the "AI doesn't use excessive dispelling" functionality.

...I don't think any of these are a working solution. Which isn't looking too promising. Anyone has an idea for a 7?
If not, so far 6 is the best but it's still awful.

Edit : or we can even have 6b. Same as 6 but the AI will not cancel and would just try to survive through the cost it cannot afford. But this might even be worse.

[Nelphine]

3 is fine. just don't let spell binding be used for dispelling. disjunction is plenty. aether binding already makes sorcery literally twice as good at disjunction as anyone else, they don't need a further tier to the spell.

6b is also fine. the costs we're talking about our far too low to bother an ai except on the lowest difficulty. (Unless you're seriously considering a maintenance of 300 for just cause. But assuming you 'only' multiply by 10 or 20 the ai would never have a problem with it. The bigger problem is that the HUMAN could never afford to steal a very rare global unless the multiplier is less than about 5.)

1 I completely agree with you on, but sorcery already has 2 of the most game changing globals in the game, AND they get both at uncommon. The particular argument that they don't have globals is simply not true. However, I do like spell binding being in the game and it should stay there.

[Seravy]

I guess I should have added "I don't think a higher casting cost really helps for the AI" to 3 as well. Or does it? How much higher are we exactly talking about? Past a certain point it's just unusuable for the human but still worth it for AI...

6b I think as a human I would be ok with up to 10x maintenance albeit probably not on every spell. Since normal maintenance on enchantments is in most cases just for show (who cares about 40Mp when having 2500 income) we can tweak them as needed for spell Binding. However "maintenance" would only work if it's not a maintenance but a power income penalty, otherwise the AI is not paying it. Even then it probably can easily afford from its 150-200% power income. If we do want it this way it might be better if it was a percentage of power income instead, like 1% for each 2 maintenance the spell has. That would matter equally for AI.

I guess if no one can come up with a better "option 7", we'll need to consider 3 or 6b but I prefer to wait a little, right now we need to work on other things anyway. Both 3 and 6b might work out if we really try for them but getting the numbers right will be a nightmare with no guarantee of success. At least 3 is easy to implement, 6b as well. In both cases the AI will likely need to learn to use Disjunction instead of Spell Binding vs low cost spells, or ignore them completely but that should be doable.

At the very least we at least agree that 1,2,4 and 5 are out which is a step forward.

[namad]

I guess I should have added "I don't think a higher casting cost really helps for the AI" to 3 as well. Or does it? How much higher are we exactly talking about? Past a certain point it's just unusuable for the human but still worth it for AI...

6b I think as a human I would be ok with up to 10x maintenance albeit probably not on every spell. Since normal maintenance on enchantments is in most cases just for show (who cares about 40Mp when having 2500 income) we can tweak them as needed for spell Binding. However "maintenance" would only work if it's not a maintenance but a power income penalty, otherwise the AI is not paying it. Even then it probably can easily afford from its 150-200% power income. If we do want it this way it might be better if it was a percentage of power income instead, like 1% for each 2 maintenance the spell has. That would matter equally for AI.

I guess if no one can come up with a better "option 7", we'll need to consider 3 or 6b but I prefer to wait a little, right now we need to work on other things anyway. Both 3 and 6b might work out if we really try for them but getting the numbers right will be a nightmare with no guarantee of success. At least 3 is easy to implement, 6b as well. In both cases the AI will likely need to learn to use Disjunction instead of Spell Binding vs low cost spells, or ignore them completely but that should be doable.

At the very least we at least agree that 1,2,4 and 5 are out which is a step forward.

maybe this cannot be coded and is not possible but what if 2) copy enchantment wasn't copyable? so if wizard A copies a global from wizard b, and then wizard a dispels the global from wizard b, then wizard c tries to copy the global from player a, it's simply not possible, not a valid target, and doesn't work? and perhaps the global is still shown and displayed as the text spell binding? using spell binding's upkeep?


also 5) isn't this problem entirely solved if the limit is 1 (no need for any UI changes to explain a 1/1)? Yes limit 1 sucks, it's pretty weak, and might be disappointing, but looking at your list it looks like the only other ideas left that can work are either to leave things alone, or remove it entirely. Why not test out limit 1 spell stolen for a few weeks before giving up? also if you can't do anything to show if it's 1/1 or 0/1... well? so what? if the spell is cast a second time, simply remove the first global stolen and replace it with the second global?