Heroes and abilities

[Seravy]

So the Elven Archer already has her unique role : being a bow hero that uses magical ranged.
Also the only hero with native Capacity, at Super level.
And the only magical ranged attacker with natural Blademaster.
So she doesn't really need more unique roles.
She is short by 2 points to reach 16.
So we can give her either a 4th random ability - but magician random abilities are kinda bad (you get more MP, RP, SP, or Power, all of which not that useful on a champion. Capacity is impossible to roll, as it' already is on the hero. So either Arcane Power (can she even get that?) Soul Linker, Lucky, Charmed are the possible good rolls...ok that's not bad at all, the chances are kinda poor though.)...or we can give her a good spell (but then she has one more unique role).
She already has Elemental Armor and Flight which are decent, just not outstanding. Maybe give her Stasis? Banish? idk. Stasis does happen to be a nice combo with an archer..

I'm tempted to do both, as neither really packs a punch on its own.

[Seravy]

And now for the Caster/MP.

Heroes have (1+Caster Level)*(1+Hero level) *4/2+8.

That's this much :
   

Each additional "random ability" roll adds 20MP on demigod level. 10 on heroism level. But these are rolls that "cost" 2 "hero points" each. So that's fine. In fact, I still rather have my 4 agility or 4 attack power, or +1 to hit than 10 extra MP. Or my +9 instead of 20 MP. If anything this seems underpowered to me. No matter how I look at it, being able to cast an extra Doom Bolt (or Prayer or anything for 40 MP) is not equally valuable to having +9 attack and defense both at the same time. Is it? Subjective question, really. Hard to answer.

Looking at item slots, at the very least, it's a trade of 1 slot vs 4 slots currently.

Which implies that
-Either MP on items is 4 times underpriced in slot requirement compared to other abilities - in other words no item slot should be worth more than +5 MP. (Which kinda means we need to get rid of MP items altogether, no one would use a slot for 5 MP I think, ever.)
-Or an ability roll landing on MP is only 1/4 of the value of it landing on something like Might or equivalent.

Either of those show horrible balance here and honestly I don't see how to fix it in either direction.

But I don't think we can even know which direction to look at without being able to judge how much casting a spell is worth vs having the stats provided by other random abilities.

Do note casting power is more available in the mod than originally - heroes no longer use up MP for their normal attacks, have a higher starting MP (old heroes had like 5 on level 1) and it's easier to make +20 caster items.

[Seravy]

Ok, let's try to approach this from a different direction.

The cost of using MP (casting a spell) is using up the unit's turn. The cost of attacking is the same. Even if the unit is melee, it can use one melee attack, and then either use a melee attack again, or use a spell. So the cost of using a spell is still the same as the melee attack, the unit merely gets twice as many "turns" in that scenario than normal.

Since the cost is the same if we measure the impact which can be done through calculating the strategic strength value of the effect, we can judge whether casting a spell or attacking is the more valuable action.

If attacking ends up more valuable, we have proven that MP is a mostly worthless resource, only useful for special strategies. (I really hope this isn't the case though, even if that means I'm right, it also means rolls landing on additional MP will always mean a wasted roll, and MP needs to be counted at zero value instead of whatever we used. Which would be horrible.)
Otherwise, we at least know MP can be used to front loading damage, and can try to compare the strategic strength difference to how much other abilities contribute.

So... I'm going to use an example army of 8 Stone Giants and the hero vs 9 Stone Giants. I'll assume the giants aren't getting a turn on either side, as we want to specifically measure the effect of the hero.

So, the hero is a not-that outstanding but still pretty good hero. Let's assume 30 attack strength and +3 To Hit. No additional abilities like ignoring defense, and the attack will be magical ranged. (Yes, this is a pretty good hero, but far from maxed. That would be 40+ attack at +5 to hit or higher, even on a non-champion. )
I'm going to assume the hero is not taking damage (either attacks at range, has first strike, or high armor.)
Damage before shields is 30*.6=18. Stone Giant blocks 3.6 damage from that, so takes 14.4 damage. That's 57.6% of the unit's total health. So strategic strength drops 57.6%, looking at the precalculated unit data for Stone Giant, that's a change of 145 strategic strength.
If the hero WAS maxed, we'd be dealing a full 252 strategic strength worth damage most likely.

Now spells.
Casting Doom Bolt would deal 12 damage. That's less, so this one is not worth it.
Casting Flame Strike would deal 62.1 damage. Definitely worth it, roughly four times as much as attacking. Would still be twice as much even on maxed hero.
Caster High Prayer. For simplicity I'll assume the hero is a 9th stone giant (our assumed stats are similar, albeit slightly higher). So attack goes up by 3, Hit goes up by 1. To Def by 1.
Our Stone Giants increase from 252 to 478 power if I calculated this well, so worth it. (we get this bonus 9 times)

So I guess it's safe to say most global, high end spells are worth it. Most single target spells are not.

In the above examples, each Flame Strike was able to front load 2 Stone Giants worth of strategic power. High Prayer was worth like 7. High Prayer is limited to once a battle meaning you could do it yourself, and nothing else even gets remotely close to the obscene boost it provides. So let's consider Flame Strike instead (it's still a far above average spell). That would mean 60 MP = 626 strategic strength, or 1 MP = 10 strategic strength of value at most. At least it has to be 145, otherwise you don't use the Caster ability, attacking is more effective. Which would be 1 MP = 2.41 strategic strength.

So we at least know 1 MP is worth between 2.410 strategic strength and

[Seravy]

Ok, let's try to approach this from a different direction.

The cost of using MP (casting a spell) is using up the unit's turn. The cost of attacking is the same. Even if the unit is melee, it can use one melee attack, and then either use a melee attack again, or use a spell. So the cost of using a spell is still the same as the melee attack, the unit merely gets twice as many "turns" in that scenario than normal.

Since the cost is the same if we measure the impact which can be done through calculating the strategic strength value of the effect, we can judge whether casting a spell or attacking is the more valuable action.

If attacking ends up more valuable, we have proven that MP is a mostly worthless resource, only useful for special strategies. (I really hope this isn't the case though, even if that means I'm right, it also means rolls landing on additional MP will always mean a wasted roll, and MP needs to be counted at zero value instead of whatever we used. Which would be horrible.)
Otherwise, we at least know MP can be used to front loading damage, and can try to compare the strategic strength difference to how much other abilities contribute.

So... I'm going to use an example army of 8 Stone Giants and the hero vs 9 Stone Giants. I'll assume the giants aren't getting a turn on either side, as we want to specifically measure the effect of the hero.

So, the hero is a not-that outstanding but still pretty good hero. Let's assume 30 attack strength and +3 To Hit. No additional abilities like ignoring defense, and the attack will be magical ranged. (Yes, this is a pretty good hero, but far from maxed. That would be 40+ attack at +5 to hit or higher, even on a non-champion. )
I'm going to assume the hero is not taking damage (either attacks at range, has first strike, or high armor.)
Damage before shields is 30*.6=18. Stone Giant blocks 3.6 damage from that, so takes 14.4 damage. That's 57.6% of the unit's total health. So strategic strength drops 57.6%, looking at the precalculated unit data for Stone Giant, that's a change of 145 strategic strength.
If the hero WAS maxed, we'd be dealing a full 252 strategic strength worth damage most likely.

Now spells.
Casting Doom Bolt would deal 12 damage. That's less, so this one is not worth it.
Casting Flame Strike would deal 62.1 damage. Definitely worth it, roughly four times as much as attacking. Would still be twice as much even on maxed hero.
Caster High Prayer. For simplicity I'll assume the hero is a 9th stone giant (our assumed stats are similar, albeit slightly higher). So attack goes up by 3, Hit goes up by 1. To Def by 1.
Our Stone Giants increase from 252 to 478 power if I calculated this well, so worth it. (we get this bonus 9 times)

So I guess it's safe to say most global, high end spells are worth it. Most single target spells are not.

In the above examples, each Flame Strike was able to front load 2 Stone Giants worth of strategic power. High Prayer was worth like 7. High Prayer is limited to once a battle meaning you could do it yourself, and nothing else even gets remotely close to the obscene boost it provides. So let's consider Flame Strike instead (it's still a far above average spell). That would mean 60 MP = 626 strategic strength, or 1 MP = 10 strategic strength of value at most. At least it has to be 145, otherwise you don't use the Caster ability, attacking is more effective. Which would be 1 MP = 2.41 strategic strength.

So we at least know 1 MP is worth between 2.41 and 10 strategic strength. So 20 MP (demigod) is worth 48 to 200 strength.

At the same time we could have a random ability. For simplicity I will only consider abilities that we rated a "2" value. We could for example have Might for +9 attack. Assuming the hero has stats equal to a stone giant defensively, that would raise the overall strategic strength by 25*14*5*9*20/8196 = 25.6 points.
So the Might is only worth half of the MP. However if the hero had 28 defense instead of 14, which is more likely, then the two would be evenly matched. Ultimately how powerful these abilities are depend on the existing hero stats - the higher the better, the effect multiplies. This is fundamentally different from the value of MP which is the same regardless of how much of it, or other abilities the hero has - you get the same value (assuming you have enough to even cast the spell) but you require to have the spell yourself.

Conclusion, MP seems fairly evenly matched with other random abilities, +20 MP is a fair trade for losing Might.

Which, unfortunately, implies that :
"MP on items is 4 times underpriced in slot requirement compared to other abilities"
as well as
-Each level of MP pool is worth 2 hero points. Which we did count as such on random MP rolls, but only at 0.5 on existing MP pool. So I was underrating MP.

However, MP has a serious disadvantage. It's vulnerable. Mana Leak and AEther Sparks can destroy it before it takes effect, and the strategic strength boost it creates doesn't take effect until the MP is used.  You also need good spells, otherwise the MP is literally worse than using normal attacks so it has zero value. Meanwhile, any gains from other hero abilities happen immediately, regardless of situation, before the enemy can take a turn. So they contribute to the hero surviving turns, while Caster does not at least until it is used up. So the Caster hero might die with MP still unused, while the noncaster's benefit already applied and can't go to waste this way.
Due to these differences I rather say MP is only worth half as much as other hero abilities.
In which case our conclusions change as well :

-Each MP level is worth 1 hero points. (I calculated with 0.5)
-Items should be limited at +10 MP to be balanced.
-Random rolls that land on MP are below average. (but I rather not make them worth double MP anyway - bad rolls is a thing which is while we pushed for more random rolls usually.)

However due to the fact that MP ends up being worth zero if the hero's stats raise high enough to outdamage the value of your best spell (or if your best spell is weak) I think I'm fine with leaving the items unchanged, maybe at a higher cost of creating, but still at 20 MP per slot. The better the hero, the more likely this will happen... (which unfortunately also means that equipping a weak hero with +60 MP is by far the best move, if you already have very rares. How you keep such a hero alive though, at that stage of the game, is a question...)

Instead of recalculating and readjusting all heroes, maybe changing their default mana pools to lower might be better?

[Seravy]

Here is an alternate formula, with each "caster level" being worth half as much MP. To use this we'd change random MP rolls to be worth 2. Doing that means it stays the same, but it allows us to adjust the base MP level of heroes in a more fine grained way (and possibly reduce it).
   

In this table, each "1" is as good as "2" in the old. However we'd still need to think about how much MP we want for the heroes. If we make it too low, it'll be a disappointing "my caster can't cast anything useful" situation.
In general, lower MP pools are worth nothing, while higher MP pools can be highly valuable - but there is not much in-between.

One more possible formula, this one raises MP more when the hero levels, but the bonus for having Caster levels is the same.
   

[Nelphine]

Right, that's why I was valuing caster differently by tier.

A Fame 0 hero will usually be found before you have access to either high casting skill items OR high impact spells, so caster simply isn't that meaningful. And they'll probably die or get replaced before the game reaches the stage that they'll see the potential the champions below have.

On the other hand, with current pricing, I can generally give all 6 heroes at least +30 skill from items, and at least 2 have the full +60 - without crafting a single item.

At the same time, by the time I'm using champions, I have rares, and sometimes very rare spells. No matter what realm I am, there's a high chance of at least one powerful spell, that I otherwise could not cast on turn 1 (because I know at least 2 such powerful spells).

Therefore, regardless of mana pool, caster on champions is extraordinarily strong, and caster on fame 0s is generally weak - or at least not as good as some other hero abilities.

For me, that means, if item costs stay the same, we have to value caster as if we can give any given hero +60 skill in the champion tier - which means the extra levels are completely meaningless.

As you've proven, the single target damage potential of casting a spell does not ever match a reasonably equipped hero.

Therefore, the only spells that matter are the big ones. Casting something like prayer and blur on round one, can easily mean you prevent dozens of damage that you would therwise take - that's far far better than demigod divine barrier. So it doesn't matter that the rest of the mana is never used because the hero always makes range attacks - a single 25 cost spell on turn 1 is better than demigod divine barrier.

Those turn 1 global combat spells (and turn 1 aoe spells like flame strike or massacre or holy word) are similarly so unbelievably powerful, that they also far surpass any other option, even if the rest of the mp never gets used.


The entire value of caster comes from that turn 1 optimized spell use. Some builds absolutely won't have this, particularly if you're unlucky with what spells you get. But we can't build around those, not when every mono realm CAN have this potential.

Thus for me, it's important to increase the cost of skill on items so that +30 is no longer expected simply through finding treasure. Getting +60 should be as rare as getting a wraithform item. This is the only way to make levels of caster relevant.

[Seravy]

Right, that's why I was valuing caster differently by tier.

A Fame 0 hero will usually be found before you have access to either high casting skill items OR high impact spells, so caster simply isn't that meaningful.

Yeah but that's actually the wrong approach.

Caster is meaningful if the caster benefit is higher than the noncaster benefit. With hero tier and and game progress, both increases - the hero gets higher stats, and your spells do more effect/MP.
So overall it's not that simple. I'd risk the statement that both increase nonlinearly, and that due to the spell system being balanced as is, spell strength increases at the same rate as stats on units.

Now heroes are a bit tricky because due to items, they get ahead of typical units, so in the midgame, they are ahead in stats, while you aren't ahead in spells, so in the midgame specifically, caster on heroes is worthless. In the early and late game, it can be (sometimes very) valuable.
Problem is, heroes are really designed for being used in the midgame, so that devalues caster in my eyes quite a bit. I can get ahead in hero stats (through buffs, retorts, equipment, enchant item, etc) to benefit more from noncaster than caster. (and defensive stats in particular not only are multiplicative in strategic value between each other but also have increasing returns in actual damage reduction within the same stat, which is a boost the strategic strength doesn't even include)

That said, heroes do have too much caster ability. They can generally cast more than your wizard in the early game, and with enough heroes, in every phase of the game. That's shows that there is a problem. While Caster (the ability) is fair or even weak compared to other random abilities, heroes still get far too much of it, and I likely devalued it more than I should have (mainly due to the influence of "worth nothing" binary logic of being able to deal more damage by attacking in the most common game scenarios with my own gameplay style.).

Items also have too much, but I don't really see how to reduce it. Items with caster are a bad pick for strong heroes in most cases, and on weak heroes, the hero dies and you don't get to use it. So it really is exclusive to early game (enemy still can't nuke the hero) or late game (you have awesome spells and the hero, even if weak by endgame standards, can expect to at least survive to cast the spells anyway - there aren't much rare and very rare single target direct damage spells in the game, so spell damage output on the hero isn't growing much past uncommon (except against Chaos wizards)).

Since they are a bad pick to begin with (useless for half the game, or on good heroes), no one would sacrifice an item slot for a smaller amount than the full 20. It's just not worth it. But if we allow the 20, we end up with 60 which is atrociously high for cheap, easy to make, early items.

That's pretty much impossible to balance - it's both too weak and too much at the same time.

We also have to keep create artifact in mind - if the cost is too high, you'll never want to create such an item. Spell Charges will become strictly superior.

I think we first should decide on the hero MP pool formula to use, and revise Caster levels on heroes to reduce base MP (without that most heroes have enough to cast those big spells, regardless of items).
Then we can try to do something with items (I think that'll be the hardest, right now I see no real way to do it other than completely dropping the "Spell Skill +X" from items and replace it with some other stat (which we can't, we already have every relevant stat in items... or would anyone want an ammo+3 item? I think not.)).

[Seravy]

   
I think this might be the perfect table for the MP formula.

No hero should start below 10 MP - you can't cast anything relevant for that much.
However wizards start at 18-24 skill (unless going low on books), so even high starting MP heroes should not go over that 18 at the first level.

With this, heroes will have roughly identical MP early, but the gap grows more and more with levels.
At the highest level, we can pick the amount of MP the hero can have in steps of 9. (lowest we can do without doing fractions/level)

So what's left is to pick which of these columns should each hero be. (and then if we feel extra masochist, recalculate the hero's overall value. With this table, each level is worth 0.5 points, but the first is worth 1 extra (26 = almost 3*9).

Do we have any particular goal on how much MP we want on a bad, average, good caster hero? (At either max level, or level 4 or any level of your choice in between you feel relevant?)

...

Ok, encountered a problem right on the first hero, the Sage.
Dispelling Wave is useless if the hero doesn't have a large MP pool to cast it. Casting cost starts at 25 and maxes at 125. So I would want the hero to at least have column 7, but the "hero points" would only permit column 2. Dropping the Sage ability here would be stupid (it's the Sage hero...) and dropping Dispelling Wave makes the whole thing meaningless. So what that means is the hero needs to move up a tier to be fair, but we can't do that either (nothing to move down, I still don't want tier 0 huntress). We don't even want to - Dispelling Wave being a common thing in AI fortresses is desired, so falls in the same category as the Beastmaster (albeit not that important, it's still important enough)
The only real way out is to say we can't consider the spell worth extra points when we also consider the MP needed to cast it extra points. You can't have one without the other, MP without good spells is, as demonstrated, useless, and spells without MP, well, obviously.
...but the MP can be spent on other spells. (spamming Phantom Warriors is pretty OP on its own for the early game)

So I'm kinda stuck, can't really incorporate caster MP into the points system... I guess I can still try to figure out how much MP the heroes should have, maybe this is the only problematic one...

[Nelphine]

I'd risk the statement that both increase nonlinearly, and that due to the spell system being balanced as is, spell strength increases at the same rate as stats on units.

I completely disagree. Spells by default have one absolutely huge limiter - you can only cast 1 spell per combat round. This is why cloud of shadow is so powerful. It's one of the big issues that came up with wrack (and to a lesser extent, call lightning).

Heroes specifically break that rule. So do other caster units, but all other caster units are carefully balanced to cast certain expected spells, often with a mono realm wizard which limits the combination potentials. 

THAT is what makes caster so powerful, and that's why I don't care about warp lightning being a problem for efreets - single target damage ISN'T a turn 1 combo.

And yes, the really does mean that caster is a binary ability. AND it means that you can saturate the number of heroes you should have in one combat. Splitting up heroes is almost always a better choice, because it only takes 1 or 2 to cast an invincible spell combo on turn 1. Obviously there are exceptions in all realms, but there are diminishing returns.


This means, I think you should price caster skill around this. 140 per 5 skill means you can add 10 skill cheaply, with enchant item. That lets you use it to boost a very weak caster like beastnaster up to casting blur. Totally worth an item slot.

+20 would then be rare. No one would invest in it, unless they had one of those game winning conbos, and weak casters.

Then drop caster levels down so that no hero, without rolling high (at least 2) on random abilities, can get 50 caster skill before level 7.

Then the best game winning Cobos would virtually require those +15 or +20 caster skill items, making them worth the extra cost.

You could take it further, and make skill on items either +5 (enchant item), or +10 (create artifact), and then you'd require 3 +10s to get the best turn 1 combos. But that might be pushing it too far.

[Seravy]

140 per 5 skill means you can add 10 skill cheaply, with enchant item. That lets you use it to boost a very weak caster like beastnaster up to casting blur. Totally worth an item slot.

Not really. I mean, I'll need to throw that item away after like two battles when the hero levels up. And I paid 280 - three full Gargoyles/Spiders/etc worth of my own casting skill for it.
Heck, for 280 MP at this phase of the game I can raise my own skill by 7! (why give +10 to a hero when I can have +7 on myself?)

No, +5 or +10 items are garbage. If I know Blur and want it cast, I do it myself. The hero can be in charge of casting the 1x confusion instead.
There aren't really that many high impact early spells in the game to have multiples.

Also, I can have as many 20s from Magicians as I want, so if any of those important spells are below 21, that takes care of it. (If I have Sorcery for Focus Magic, up to 35!)

But let's do the heroes first, items later please. (in fact, items tomorrow. It's 3 am...)