SORCERY Realm

[Nelphine]

Another unit I would consider would be draconian warship.. No that would be the same as catapults, no good.

I'd also consider removing halfling magicians IF slinger bullets weren't affected by missile immunity.

I'd be OK removing pegasi (I kid, but I do think the unit is mostly useless by the time you get it. Sadly I dislike high elves too Much to actually play them to test properly.)

I'd also consider removing lizardman halberdiers, as much as I love them. Is there any other race that gets two different units from fighters guild?

I'm sure there are others, but most units definitely seem to have a place.

[Seravy]

More Spell Lock considerations...


I believe we need to set a goal here as well.
I say, with Spell Lock, 3 other enchantments should be still worth it, while Dispelling Wave could be cost effective on 4 or more.

That means breaking even at 4.5 times 9 = 40.5 spells.

To reach that goal, the additional resistance needs to be...a massive +270 on every (big) spell.
Maybe this goal wasn't realistic?

Okay, so without Locks, it's 6 big spells. With locks, 3 big spells plus the lock on the same units is the bare minimal amount I think where it's still worth casting = 24 spells. This one requites only a +100 bonus. This is without AB. Against AB, you need...270 again. It's probably not realistic to include AB here, as it raises the needed amount too much. Unless, 18 spells are still too many? I don't know, maybe it is... Maybe the goal should be 3 buffed units in the stack instead of 6?

Also, the AI's overland casting priority should be set in a way that it doesn't dispel faster than the player can recast on these Spell Locked stacks if they are below the goal. This is important because the AI ignores the cost-effectiveness calculations due to cost reduction and resource bonus. I'm hoping the same casting priority will satisfy both goals, if not, we have trouble.


On raising Spell Binding cost, that has a problem with AI.
When the human player is faced with the decisions to use Disjunction or Spell Binding, they can usually pick the better option - at 1500, Spell Binding will not always be better, for example a 800 mana spell costs 1200 to remove with Disjunction, and if you already own a copy then it's better to do that.
The AI can't do this. It always uses Spell Binding. This was acceptable so far - the AI can afford losing a few hundred in exchange for the added benefit of "I get a new copy of the spell just in case someone dispells it meanwhile" - as the AI can't predict the chance of that happening or look at Detect Magic and see incoming Disjunctions.

As it already has an extra 1.4x cost, this tactic is already questionable, but if we raise the cost even further, it'll be outright awful. On the other hand, even if we teach the AI to pick Disjunction instead which is very hard to do (too many factors to consider, runemaster, specialist, aether binding, the cost of the target, the predicted target of enemy disjunctions - probably over 200 lines of new code just for this), the AI is not aiming for 100% disjunctions against Specialist/runemasters targets to keep those retorts relevant. So in case of using DJ as a replacement of SB, it would even need to consider that.
I'd like to avoid coding this if possible and I don't think letting the AI use 2100 cost Spell Bindings on 800 mana spells it has already stolen instead of 1200 Disjunctions is a good idea. At 1200 Spell Bindings, this was an optimal tactic. At 1680 with the cost multiplier, it's already questionable but it's not too bad. (in fact we might want to remove Spell Binding from this 1.4* cost list? It is expensive enough to Spell Blast it without...)

[Arnuz]

Whoa, the speed of the thread!!

Orc spearmen is not an important unit to me. Go ahead.

Vouched. Please reduce the cost of the orc swordsmen though - I'd go as low as 10, reducing also its values (and eventually name, to goblins? how about a 8 figure swordsman with all values 1? That'd be fun! And as they're supposed to be technological it'd also fit well... Don't know how feasible this is)

I'd be OK removing pegasi (I kid, but I do think the unit is mostly useless by the time you get it. Sadly I dislike high elves too Much to actually play them to test properly.)

Also vouched, totally useless for its cost and time to market

On DW/SL: I understand it can be frustrating but I wouldn't remove its effectiveness against many small spells because of the compounding power of many small spells. I'd make it a multiplier rather than a flat bonus to resistance. OTOH, SL is also sorcery and it makes sense that sorcery would have the counter-weapon to DW, so I'm ok with either. But at least without SL, DW should definitely be a mass small spell nerf.

[Arnuz]

...and now back to watching that Youtube video. Someone was playing on Lunatic! I should be sleeping already...
https://www.youtube.com/watch?v=vg6SAH2Hik4

Watched it too, pretty nifty, lunatic without ever casting - unless I missed some casting but I don't think so, it seemed that the mana was always converted to gold. Looks like I'm not the only one to think that omni is too front loaded : )

[Juffos]

...and now back to watching that Youtube video. Someone was playing on Lunatic! I should be sleeping already...
https://www.youtube.com/watch?v=vg6SAH2Hik4

Watched it too, pretty nifty, lunatic without ever casting - unless I missed some casting but I don't think so, it seemed that the mana was always converted to gold. Looks like I'm not the only one to think that omni is too front loaded : )

Zero mana spent casting spells entire game. 8 skill at the endgame. The project leader won't dirty her hands with magic, she delegates.

[Nibiru]

...and now back to watching that Youtube video. Someone was playing on Lunatic! I should be sleeping already...
https://www.youtube.com/watch?v=vg6SAH2Hik4

Watched it too, pretty nifty, lunatic without ever casting - unless I missed some casting but I don't think so, it seemed that the mana was always converted to gold. Looks like I'm not the only one to think that omni is too front loaded : )

More like flying is overpowered.

[Arnuz]

Zero mana spent casting spells entire game. 8 skill at the endgame. The project leader won't dirty her hands with magic, she delegates.

Grats 

Watched it too, pretty nifty, lunatic without ever casting - unless I missed some casting but I don't think so, it seemed that the mana was always converted to gold. Looks like I'm not the only one to think that omni is too front loaded : )

More like flying is overpowered.

He he Juffos looks like you're being challenged to repeat that feat without starting draconian!!

[Nibiru]

Zero mana spent casting spells entire game. 8 skill at the endgame. The project leader won't dirty her hands with magic, she delegates.

Grats 

Watched it too, pretty nifty, lunatic without ever casting - unless I missed some casting but I don't think so, it seemed that the mana was always converted to gold. Looks like I'm not the only one to think that omni is too front loaded : )

More like flying is overpowered.

He he Juffos looks like you're being challenged to repeat that feat without starting draconian!!

Without using flying to be more specific.

[Seravy]

As Disenchant Area uses the same formula to dispel city curses as Dispelling Wave dispells city buffs, I believe it would make sense if it granted the same +66% dispel bonus to both, how about it? Does Sorcery really needs to be that much more effective against city curses (rare spells) on Uncommon level? At very rare, Spell Ward completely prevents curses anyway.

Dispel Magic and Disjuction could stay at +100%.
Or....actually, to quote Monty Python "And now for something completely different."

I'm worried as I had time to think today. Sure, we calculated cost-effectiveness but...this is a game, not a math exam.
Let's say you face a stack with 6 units that have a buff on it - we defined this to be the "break-even" point.
You pay 125 to cast Dispelling Wave. It has strength 25 (with AB), so against average, cost 100 enchantments, that's a 20% chance of success per spell. On 6 spells that means you'll dispel one, sometimes two - makes sense, 1.25 of those = 125 mana.
The problem? You might break even on resources but wasting 125 of your combat skill to remove a singe enchantment is an absolutely horrible "I want to lose this battle" level move. It gets you nowhere, especially as in this situation you often need all of them to be gone. If you can't damage a unit with Invulnerability, it doesn't matter if there are 5 or one. If you fail to remove all 6, even removing one is a wasted move.

Also, combat has a distance modifier. So it won't even cost 125 but often 250 or 375 mana.

We messed the calculations up by forgetting that, and probably it was a very bad idea to even calculate cost efficiency. MP cost is completely irrelevant in combat on the side of the defender. What matters is the capability of removing enough enchantments to win the battle, or at least, severely damage the stack so it stops being a threat.  Every time that doesn't happen, the stack gets to conquer or destroy an additional city. At 6 or even 9 or 14 spells, that fails. Yes, the enemy might have lost a good amount of overland casting skill, but their stack is still good to go and the defender lost their army and city. Then said stack attacks the next target...so "break even or not", using the spell isn't helping. Combat is binary. You either severely damage the enemy, or you lose the entire game as they are free to keep attacking and taking away your stuff.

Another viewpoint...I cast 2 doom bolts to kill your buffed unit. For 80 mana I erased all the spells and all the enchantments on it. Compared to 125 having a 20% chance to remove buffs, which is the equivalent of...removing all buffs from about 2 units in a full stack of 9 enemies all evenly buffed, so 1 and a half for 80 mana. Not bad, 1 is more than 1.5 right? Wrong! The unit itself is still unhurt! Doom Bolt killed it. So Doom Bolt is better. And Doom Bolt is the least cost-effective direct damage spell in exchange to being Doom type damage. Also the enemy stack will usually not be evenly buffed - half the units will be buffed heavily, the rest not at all, making Doom Bolt twice as good! (and even Dispel Magic better...)

Or look at Death Spell of Holy Word. Since when were those, or any other combat spell "evenly cost effective"? Even commons...Black Sleep (nearly) kills a unit. It costs 15 and the save modifier is -2. So for example against a Great Lizard, you pay 30 to deal 225 MP worth of damage. That's not even at all!
In fact, most combat spells deal much more damage than their cost. So a "break even" Dispelling Wave fails compared to those. Yes, it's still good against, say, 3 times as many enchantments, but we set the break even goal with the intention that it should be worth it if it's one or two more than that. "Worth", not economically but as part of the greater picture of the combat and overall game. And this isn't the case, far from it.

(that said, cost effectiveness might be valid for the overland map. Except for two problems :
One, the AI is massively ahead in resources and rarely megabuffs so cost efficiency it doesn't matter when the human player casts it. (in fact, as it targets one of the 4 enemies, it should be 4 times more efficient to be worth doing, even if ignoring the AI bonus)
Two, the AI only spends a tiny fraction of its resources on casting Dispelling Waves (even if the target is high priority) - so it doesn't care either. How many enchantments can be kept up on a stack if the player is recasting the lost spells only depends on how often the AI casts DW. Even if it is half as effective, if the AI casts it twice as often, it works, and as the player is pouring their resources into buffing, the AI won't be threatened by anything else so it's safe to do so. On the other hand, even if it's more effective, if the player can keep just enough buffs on the stack to keep destroying AI cities without losses, then the player wins and DW isn't doing their job, and again this only depends on how often it is cast. (or more like, the product of AI casting priority and dispelling effectiveness per cast, but the former is the more relevant part as it's independent of all other considerations such as human use of DW against AI.)

...so why were we treating this as an economic problem in the first place? I feel like I completely missed the point here.

Time to start over.

Please post the following goals instead and ignore everything in the last X pages, I'm quite sure our premise was very wrong :

1. Assuming an enemy stack with 27 buffs (Endurance, Lionheart, Holy Armor), if you are the defender, how many of those should be dispelled for each 125 casting skill and 375 mana you spend on DW to make it guaranteed that you are better off than casting any of Doom Boltx3, Phantom Beastx3, or Prayer+Black Prayer+Blur+Mana Leak, or 3x Possession or 6x Psionic Blast or 5x Ice Bolt? All of these cost 125. The enemy units are something generic, like 3 berserkers, 3 great lizards 2 nomad pikemen and one high level hero.

2. Assuming the enemy doomstack has 27 buffs - Magic Immunity, Invulnerability, and Lionheart - imagine generic doomstack units, such as idk, 9 Angels - how many of those do you want to remove per casting to be sure you are not losing the game - you have 12 cities, each with a full garrison of 9 Magicians and a casting skill of 20.

(For a moment, please imagine you ARE playing a game where these happen and don't bulldoze the entire map with Berserkers in the first 5 years.)

[Seravy]

Also one more thing I forgot. Buff stacking is one thing but...
The human player can use buffs in a very effective way. They can, for example, say "hey that's a city of 9 Magicians"...I put Elemental Armor on these 6 random units and now that city is mine. Oh wait, you're an AI, so most of your cities have magicians in them? Good, then I win the game. You play Sorcery and could dispell them to stop me? Too bad, I only used 6 enchantments which are too few so you aren't allowed to dispel them even though you totally should.

That should never happen. Not a very high change to dispel because it's a small stack with few buffs? Fine, if the AI is significantly more powerful, they'll do it every once in a while anyway, but it might crush a weaker AI. That's ok. But not being able to ever do it...no I think that's not ok.

...so at this point I just don't know anymore.