Re: Hybrid Spells

[JonnyMonroe]

I would like to get some real discussion of hybrid spells on the table in this thread for a number of reasons; First and foremost, the balance and implimentation of these spells has never been publically discussed. Secondly, the investment into these spells is such that getting a fair sample of data to draw conclusion is difficult without input from multiple players. Without ado:

Hybrid Spells
Hybrid spells are, as the name suggests, spells created by mixing 2 schools of magic. The known hybrid spells are Antipode (fire/water), Defribulate (Lighting/water) and Appocalypse (Earth/Fire).

Skill points
With normal spells 1 new rank is reached every 5 skill points invested into its respective tree. With hybrid spells 1 new rank is reached with 5 points invested into each of its respective trees; thus hybrid spells require a 200% skill investment over an ordinary spell.

Players recieve 2 skill points per level. We assume that a player spends 1 point per level (averaged out) on their 'main attack' school. The most basic aspect of quest is killing mobs; and killing mobs is acheived by dealing damage, hence maintaining maximum rank of your attack for your level is an obvious route for players to take. There are many offensive skill trees however, so it is possible that a player might not keep up top-rank and instead elect to try and get higher damage from mixing in a greater investment in, say, bookworm+precision or precision+contact. With a normal spell those extra options are available while maintiaining perfect rank/level, however with hybrid spells those options can only be pursued if you sacrifice perfect rank/level. We can conclude then that the benefit gained from investing in a normal spell school + offensive stats should -roughly- yield similar results to a pure investment in a hybrid school. We can extrapolate this conclusion and say that a normal spell school without a perfect rank/level investment and more points spent in other offensive skills should also be roughly on par with a hybrid school of imperfect rank/level and some points invested in other offensive skills. With either of the above circumstances, it should follow that spells cannot in fairness be compared rank for rank, as an example;

A rank 5 fireball represents 25 skill invested in fire
A rank 5 Antipode represents 50 skill invested between fire + water

To compare same rank we need to instead view this as:

A rank 5 fireball represents 25 skill invested in fire + 25 in bookworm/precision/contact (offensive stats are presumed as the hybrid spell exists purely as a stronger offensive option)
A rank 5 Antipode represents 50 skill invested between fire + water

If you were to instead compare at same skill investment levels then you would clearly compare rank 10 fireball to rank 5 antipode; I suspect there is a flaw in this comparison however, as skills do not gain power linearly from rank to rank. For that reason this comparison is ignored and the first comparison is used for the rest of this post.

Stop. Math Time.

Firebolt 5:
MP cost: 6
Damage: 137 to 137
+1 INT -> +1.05 damage
base miss %: 5
Ignores 100% of target's armor

Antipode 5:
MP cost: 7
Damage: 162 to 162
+1 INT -> +1.25 damage
base miss %: 5

difference:
17% increased mana cost (probably horribly innacurate due to rounding on low numbers)
18% increased damage
500% increased INT scaling.

Increased mana cost:
For most mages this is not currently an issue. Mana pools are pretty huge right now and only ever run the risk of failing you if you're on a tough boss or simply ignore your mana as you fight.
18% increased damage cap:
Level 25 mobs (the same level as the rank being examined) have around 23% armour reduction. By having 0 armour reduction on this Antipode the higher damage cap is essentially negated.
Increased INT scalling:
At this level INT scalling on both firebolt and Antipode can be fairly safely ignored as the damage cap is easilly reached by the time you have the spell. The high scalling on antipode essentially guarantees you will reach the cap. The low scalling on fireball means you can conceiveably fall below the cap, but with 25 skill points to spend in something like, say, bookworm; its entirely reachable (even if by accident) - this also results in a higher crit% on same level mobs than that provided by antipode. Note that the 'int cap' for this rank can also be reached with some half-decent gear.

(This section of the post is incomplete and I will add more comparisons for more ranks when I have the time).

So at rank 5 Firebolt will do around the same damage as Antipode 5 and allow you points to spend making it either crit more frequently, hit more frequently or if you prefer, pick up some healing in holy. I am of the opinion that damage caps are a flawed system and I intend to go into it more deeply elsewhere on this board.

Q&A:
But what about the other spells you have access to as a Hybrid caster?
While its true that as a hybrid mage you essentially have access to 3 spell schools; it is not an advantage. You can't use more than 1 school at a time and the current spell system means that the veriety in your options lends no advantage. Its like having 3 different cars you could drive to the store in - you can only take one of them.

-Please post more questions below-

Edit - 5*5=25!

[JonnyMonroe]

I'd like to jump to an early conclusion here and suggest (it is a suggestions forum afterall) that hybrid spell not be diminished by armour. Its too early to say that that is an ideal solution but it certainly seems to be a stem in the right direction.

[fingolfin]

Correct me if I am wrong, but don't the damage caps come off once you get to 50? Also, I think you need to look long term as well, all wep/magic skills are capped at 500, so looking into the future I would rather have a capped combo spell then a regular one. After lvl 500 you'll need something else to put points in.

[Dlorak]

brilliant post.. great great insight in a readable format. Awsome.

To add some raw information that may help your analysis, here are some of the top-level specs of the three synergy trees.


(very standard. the flat return on investment for 2x skillpoints)
Antipod 100
fireRqr = 500
WaterRqr = 500
102mp   
intdmg 6
5% base miss
3 second recovery time


(High miss and high variance. Highest single-attack damages possible in Quest will come from Defibrilate.. When it lands.)
Defibrilate 100
LightningRqr = 500
WaterRqr = 500
102mp   
intdmg 8
29% base miss
+/- 20% damage 'variance' (if INT*intdmg = 100, then the resulting damage after variance the resulting damage can be as low as 80 and as high as 120)
3 second recovery time


(self-damage in exchange for consistently devastating attack damages.)
Apocalypse 100
fireRqr = 500
EarthRqr = 500
103mp   
intdmg 8
2% base miss
5% damage to self. (-0.4 INT = self damage)  (The self-damage is tweaked for monster-fighting, and is not a balanced for PK/PVP/Boss fights. This is a known "possible imbalance" and has not seen extensive play-testing)
3 second recovery time



Some other facts:
intdmg scaling is linear for all spell trees.
MP differences are indeed negligible as levels increase, synergy spell MP costs are essentially the same as non-synergy spell MP costs.

So at rank 5 Firebolt will do around the same damage as Antipode 5 and allow you points to spend making it either crit more frequently, hit more frequently or if you prefer, pick up some healing in holy. I am of the opinion that damage caps are a flawed system and I intend to go into it more deeply elsewhere on this board.

Very good observation. The math is skewed at the early stages because both trees start at an intDmg of 1, instead of the more mathematically elgant starting point of zero.
if y = ax + b,
b=1
x=level of spell
a=intScaling
y=intDmg

then we have
FireboltIntDmg = (0.01)x + 1
AntipodeIntDmg = (0.05)x + 1

the end points are way different (6 for antipode, v 2 for firebolt) but early on, 1.25intdmg is very very similar to 1.05 creating the near-equivalence early on.

Again, excellent post. I look forward to your comments on the int-cap system, it is a sour-point in the magic system. Before you  delve into it, I'd like to give you a little info on how int-caps became real: int-caps are an ad-hoc fix for balancing MP costs to damages, preventing firebolt 1 from being king-kong in damage to MP ratio.
firebolt2: 200INT @ 1.01 intdmg = 202 dmg, 2MP cost.  202/2 = 101dmg/MP
firebolt3: 200INT @ 1.02 intdmg = 204 dmg, 3MP cost. 204/3 = 68dmg/MP
there is some dependence on INT here, and some inflection point where the ratio stops moving around but its a bit of a moving target and I never sunk my teeth into a real analysis of how best to deal with it.

[JonnyMonroe]

I personally feel like damage caps are a carrot-on-a-stick for mages, baiting them into playing long enough to reach that golden rank 51 where all that INT finally matters. I won't debate the value of carrots and sticks in online games, I think that value is regularly proven, but some arbitrary magic spell rank seems a bad way to go about it. If mana efficiency is the concern there are other ways to address that (assuming of course that it is a concern, considering that players make a substantial damage sacrifice in order to achieve the improved efficiency). I conceed that there may be ranks I've not yet examined in which the damage/efficiency trade-offs become blurred and ranks become more interchangeable. Helen suggested to me that an increasing mana cost might be a possible solution. If I make myself an intdmg/mana chart and look at the places where the line blurs, I can probably come up with some ideas.

--------------------------------------------------------------------------------------------------------------------------------------

Whilst it is a good point about the initial scaling of the early spells, it is still moot as a result of damage caps that you basically hit as soon as you have the spell. I understand the purpose of damage caps even if I don't agree with it. What I think would be a better solution would be a 'softcap' wherein after the damage cap is reached the int--> damage ratio begins to deplete. So lets say we have a spell of 2int --> damage and a cap of 100 (purely hypothetical numbers),

At 10 int the spell hits 20
at 50int the spell hits 100 - at this point the cap is reached and any INT gained beyond this has a lower contribution
51-75 int - 3/4 contirbution (1.5 int-->damage) - 101 to 138 damge
76++ int - 1/2 contribution (1int --> damage) - 139 onwards

This basically causes a softcap on int for spells. Maybe a more formulaic solution would look like:
I = Int
X = INT/Damage Ratio
Y = Final Damage
Z = Damage Cap

Y=I*X;  */ (where I*X <= Z) /*
Y=((Z*0.5)+(I*(X*0.5))

Just to see what kind of numbers something like this would produce, I threw the formulas into a spreadsheet on Gdocs - http://spreadsheets.google.com/ccc?key=pM5t1o77FOFTzqEElsJ03WA

Thats publically editable so feel free to change the 'cap' and 'scale' values.

------------------------------------------------------------------------------------------------------

Wild idea out of the blue - How about 1 rank of spell for every school, and the intdmg amount is simply based off the number of skillpoints invested in that school (and the MP cost is simply a % of the skill invested also)? - It removes the problem of downranking for efficiency, it removes the *huge* list of spells a hybrid mage has in their spell book. For those of us that often forget to update our hotkeys, it saves a small job. You simply make the mana cost, intdmg and base damage ratios the same as they are now (6intdmg on anti @ 500 skill, 5intdmg @ 400 skill, 1intdmg @ 0skill). Obviously for hybrid spells it would need to take the skill point value for the tree with the lower amount of the 2 (at 100 fire + 200 water your Antipode and Firebolt would effectively be what is now rank 20, your hydro would be what is currently rank 40).

[Mighty_Mite]

Very Nice work Jonny, I guess having done it before, I look at this in a long term strategy
and don't mind the nerfing that marks the early levels (below 51) for the power that comes
at the later levels. I find that leveling with just fire although weaker than the fighters at
the same levels, is made up for by leveling Holy and having the great healing ability high
INT provides. After 255 ( rank 51) all skill points can go to leveling up the second school,
damage and healing will continue to grow with int both from leveling and from stronger mods
as equipment is upgraded. At level 376 the second school is up to the first and leveling
both schools begins For Apoc you are at this point getting 4 x int not too shabby.
It would be nice to have a more level field but I got into it knowing the slant. 

I guess I really don't understand the caps, having little experience in the game before them.
It does seem to me with mods being low due to level needed requirements, it would be hard
to get really high int and unbalance the game. But then again I am sure there is more to it
that I am not seeing. Again thanks for the good work and bringing this issue forward in such
an understandable and thorough way.

[Dlorak]

Wild idea out of the blue - How about 1 rank of spell for every school, and the intdmg amount is simply based off the number of skillpoints invested in that school (and the MP cost is simply a % of the skill invested also)? - It removes the problem of downranking for efficiency, it removes the *huge* list of spells a hybrid mage has in their spell book. For those of us that often forget to update our hotkeys, it saves a small job. You simply make the mana cost, intdmg and base damage ratios the same as they are now (6intdmg on anti @ 500 skill, 5intdmg @ 400 skill, 1intdmg @ 0skill). Obviously for hybrid spells it would need to take the skill point value for the tree with the lower amount of the 2 (at 100 fire + 200 water your Antipode and Firebolt would effectively be what is now rank 20, your hydro would be what is currently rank 40).

yes. This is that elegant solution we've been looking for.

haha... MM you gamer, what an awsome plan... why didn't you post that in a guide for me? 

[Dlorak]

What if we punish people for NOT investing skill points by lowering their damage?

This formula:

Code:

skill-level + 20
-------------------
Level + 20

can be multiplied to the damage. It will be unused after level 500.
The result of the formula by the damage a spell does, we get a formula that just acts as a check on if you've invested the points into the fire skill up to your level and punishes if you haven't (by taking off about 5% damage.)

At this point we can take caps out entirely because punishing non-skill point investors (0%) is a much better alternative...


(everything below here is a waste of space.)

Wild idea out of the blue - How about 1 rank of spell for every school, and the intdmg amount is simply based off the number of skillpoints invested in that school (and the MP cost is simply a % of the skill invested also)? - It removes the problem of downranking for efficiency, it removes the *huge* list of spells a hybrid mage has in their spell book. For those of us that often forget to update our hotkeys, it saves a small job. You simply make the mana cost, intdmg and base damage ratios the same as they are now (6intdmg on anti @ 500 skill, 5intdmg @ 400 skill, 1intdmg @ 0skill). Obviously for hybrid spells it would need to take the skill point value for the tree with the lower amount of the 2 (at 100 fire + 200 water your Antipode and Firebolt would effectively be what is now rank 20, your hydro would be what is currently rank 40).

I'm calling this the "single-spell solution". We consolidate all "firebolt X"'s and all "hydro jets X"'s (etc etc) into a single rank and call it "firebolt" or "hydro jet". The benefits for "firebolt" scale with your skill points invested in fire.

I Love this idea.

*devil's advocate hat on*

Straight up the single-spell solution doesn't solve the MP:damage ratio problem. Players would simply invest no points until the %damage gains from spells become relevent.

The solution Helen mentioned might indeed work to releive the MP:damage ratio problem... Increase MP costs to lower the marginal change in MP cost.
Say, level 1 costs 10MP, then the jump to 11MP is an agreeable 10% increase in MP costs for the extra 1% damage. Or is it? read on...

The crux of the MP:damage ratio problem is that 1% increase in damage is puny in the early stages. This fault can be traced to the percentage-based spells (as opposed to fixed-damage based spells). These percentage-based spells are essential to end-game balance but truly have no place in the early game. Let me elaborate on this point...
When you are level 10, and have 50INT, 1% is 0.5 additional damage. Whoopee. may as well be zero.
When you have level 200 and have 1000INT, 1% is 10 damage and that adds up when you get +40% damage: +400 damage!
When you have level 500 and have 3000INT, 1% is 30 damage and that adds up FAST when you get +100% damage: +3000 damage! WHOMP!

So clearly, %based is right for the end-game, but wrong for the early game. In the early game the question is "Why would anyone want to pay more MP for +1 damage?" ... Answer is "they won't" and thus the mp-abuse problem.

A cap or soft-cap system solves the problem, it really does. But in the meantime it punishes high-INT builds.

SO, lets see if there is an alternative where (1) High-int ISNT punished (2) we keep the %-based spells in the end-game.


========= FIRST HUNCH (*EDIT: its a bad hunch, tho!... read critically*) =============
First, some credits: this is 100% triggered by Jonny's mention (in-game) that the single-spell solution leave the spellbook very "empty" and that we will just have to "fill it up." Excellent point. I think we'll fill it up sooner rather than later whether or not we go with this hunch   

Onto the hunch: a new spell for each tree, and slap a level requirement on the big-time %-based spells (cutting off the MP-abuse region).

Something like:

Code:

skill-level + AdjustNum
------------------------------------    + (0.001) * skill-level
Level + AdjustNum

Then that whole thing equates to something in the range of [0.95, 1] + 0.01 per 100 levels.
so 100% to 110% for the first 100 levels.

call that the INTdmg.

The adjustNum being 20 or so. This forces mages to invest into skill points to keep damage up to snuff, which combines nicely with the one-spell solution's quirk of increasing MP cost with skill-level. Nice.

In a nutshell this is a very slow-growing spell, always around 1intdmg. It is more of a penalty on players not to keep their skill-level up to snuff with their own level. Its like a mage-tax on the mage-build's SP.

So the cap-less spell will be better for 90% of mages for level 1-250. Low-int mages that have perfected having int RIGHT at the cap and no more will no longer be rewarded for that talent and can pump into INT like everyone else.

Why does this help? because the original trees can be made to have FIRE=250 requirement to cast, cutting off all the low-level abuse, and providing a target unlocking moment where the "firebolt" will become available.

Is this better than caps? maybe. Maybe not.
EDIT: on the re-read it seems intuitively clear this implementation ends up punishing all mages instead of just high-int mages. It cuts off their excellent level 100-250 spell boosts that keep them evenly paced with fighters. It is just plain bad thinking!


=========== SOFT CAPS ============

this could be the solution. But as mentioned, it punishes high-INT characters...


=========== Punish non-investment ========
This formula:

Code:

skill-level + 20
-------------------
Level + 20

When we multiply the result of the formula by the damage a spell does, we get a formula that just acts as a check on if you've invested the points into the fire skill up to your level and punishes if you haven't (by taking off about 5% damage.)

At this point we can take caps out entirely because punishing non-skill point investors (0%) is a much better alternative...




====

Thoughts?

[Lyrael]

The solution Helen mentioned might indeed work to releive the MP:damage ratio problem... Increase MP costs to lower the marginal change in MP cost.
Say, level 1 costs 10MP, then the jump to 11MP is an agreeable 10% increase in MP costs for the extra 1% damage. Or is it? read on...

I'm afraid he misquoted me there Dlor, I wasn't suggesting a flat increase in cost, I was suggesting that the bigger the gap between your level and your skill level (or, pre-single-spell soluton, the skill level of that particular rank), the higher the mp cost was, hence stopping people downranking but not punishing anyone for using the appropriate ranks of spells.

[Dlorak]

oh LOL... even better... that solves it too! In retrospect I way over-thought this!

[Lyrael]

I blame Jonny. He should apologise for making you think so much

[JonnyMonroe]

I blame Jonny. He should apologise for making you think so much

somebody's got to do it : /

EDIT
I need to read your posts more closely. Pretty much everything I commented is already covered. Ah well, I'll bold what I think is still relevant and you can ignore the rest.

On skill - level punishment;

I felt that was implicit in the fact that as you level up the things you fight get stronger, so you need to keep your attack up to scratch if you want to stay on top of your game. If you have your spell diminish as you level then you're getting close to homogenizing with melee and you might as well then impliment a spell power that works the same as weapon power. At that point there really isn't much left to mark a difference betwen spells and melee. In a world where intdmg purely scales with the skill in a tree is it really necessary to punish someone for not investing in that tree beyond them not having an optimal setup? Sure they might get better precision or thick skin or something but thats rarely going to compare to the results you'd get from a stronger intdmg conversion.

On mana:
have you considered having mana dynamically cost a flat % of the damage done? (before armour mitigation) - the result should be that all levels of spell skill have similar efficiency (identical for spells that bypass armour, a slightly more blurred line for spells that don't), the side-effect of ths is that it means a caster actually has to care about which weapon they use since that's a major source of mana (right now I could happily equip a club or some throwing knives if they had better caster stats). If the mana problem is solved then is there hope for the lifting of damage caps? you talk often about not punishing high-int builds but damage caps completely invalidate them before level 255.

------------------------------------------------------------------------

I really need to level my character so I can get a better sampling on all of this.

*eagerly awaiting counterarguments*

[Dlorak]

have you considered having mana dynamically cost a flat % of the damage done? (before armour mitigation)

No, I hadn't... Two painful consequences come to mind:
higher int would mean higher MP costs.
higher levels, when intDmg really really takes off, MP costs would be skyrocketing, and MP stores will be close to flat-lining in comparison.

On skill - level punishment;

I felt that was implicit in the fact that as you level up the things you fight get stronger, so you need to keep your attack up to scratch if you want to stay on top of your game. If you have your spell diminish as you level then you're getting close to homogenizing with melee and you might as well then implement a spell power that works the same as weapon power. At that point there really isn't much left to mark a difference betwen spells and melee. In a world where intdmg purely scales with the skill in a tree is it really necessary to punish someone for not investing in that tree beyond them not having an optimal setup? Sure they might get better precision or thick skin or something but that's rarely going to compare to the results you'd get from a stronger intdmg conversion.

On this point I disagree it creates equivalence. We're not using weapon power, or in any other way being weapon dependent. We're never scaling down or up for full-investment. Mages would be completely as they are, with the expection that we punish anything less than full commitment to a spell tree (ie: further punish bad choices).

I also want to point out that the diminishing returns effects 0% of the mage population. Especially after the single-spell solution gets implemented: more damage for each point invested, what could be better? As your rightly point out, other skill expenditures are "rarely going to compare to the result's you'd get from a stronger intdmg conversion." So again, we punish no one! just the attempts to abuse MP:dmg ratio and the foolish who don't invest fully into their core attack (no one is in this second category!)

[Celestia]

I submitted a list of some spells some time ago, back before 5.0 came out.