Epic Magic Big Thread

[Cheiromancer]

It occurs to me that all my kernel analysis presumed that enhance = +4 levels. If it is really +2 levels, my numbers will be off. Of course, if Heighten is worth a little more it might balance. Heighten = +3 might work. Bumping a fireball type spell in the old system would increase the save DC by +4 and raise the dice cap by 10d6. So 4 levels = enhance (+8SP) + heighten (+8SP) = 16SP (plus some other benefit; delay, or whatever). In the new order it would be enhance (+4SP) + heighten (+12SP) = 16SP.

Not that kernel analysis is absolutely correct; it's basically a substitute for intuition, and when intuition differs with the calculations, guess which one has to go? In particular I'm having trouble reconciling certain spells of a suite with each other: plague vs contagion, for instance. Plague wants to come out as an epic spell, not a 7th level spell, at least the way I am calculating it.

Oops. Have to go to class!

[edit] Hmmm. You are right about Matt getting stiffed. Consider the proposal withdrawn.

I've been kinda assuming that a wizard's intelligence is going to be equal to his level at epic levels; so Matt will have an intelligence of 50 at 50th level (15 start + 5 inherent + 12 level up, +18 headband, assuming level cubed x 100 gp wealth and about 25% of base wealth in that headband.) His DCs are going to be 30 + spell level without Heightening. So 37 for a finger of death. Everything can save vs that; even the lowly black slaad saves on a 7 or better. He'd be better off using a no save spell like polar ray, or an esoteric spell like one of the power words or something. In particular, even if he uses all his AMCs on Heightens, he will still not be able to affect a KR 50 creature with a base save of 50; it succeeds on everything but a 1.

It might be we are scaling things wrong. Maybe we shouldn't be multiplying UK's numbers by 2/3. A lot of his CRs are equal to WotC's and we don't multiply *them* by 2/3. Then... let's see. If a mature red dragon is CR 34 (I'm going by v4 of the appendix), that means it is a moderate (15-25% resources) encounter for a 34th level party. Or a quite tough encounter for a CR 17 party (they have to rest afterwards, and somebody might die). Or a 16th level party with elite ability scores. Does that match your experience? A CR 14 marut would be a moderate encounter for a 13th level party with elite scores. That's pretty close to the truth, isn't it?

We might have to make a new list of sample monsters (with hit points and saves) to use as benchmarks. If we don't multiply by 2/3 then all the saves will be more reasonable; the system might not even be broken after all. At least it will be less broken.

That'll be my new proposal; using UK's CRs as written, and have them be the benchmark. Of course that means that a 50th level party will find a CR 100 boss monster to be a tough challenge (like a 17th level group faced by a mature red dragon or a lavawight or something), but possibly defeatable.

Do you have volume 1 of the IH:Bestiary?

[edit2] Ok, I have 681 creatures on a spreadsheet sorted according to UK's CR. What CRs should I give samples of with hp, HD and saves?

The creatures range from the tiny monstrous centiped (CR 1/16) to the mighty neutronium golem (CR 9721). Though there is really one a fairly continuous sampling of monsters up to CR 130 or so. After that there are lots of gaps.

 

[Sepulchrave II]

It occurs to me that all my kernel analysis presumed that enhance = +4 levels. If it is really +2 levels, my numbers will be off. Of course, if Heighten is worth a little more it might balance. Heighten = +3 might work. Bumping a fireball type spell in the old system would increase the save DC by +4 and raise the dice cap by 10d6. So 4 levels = enhance (+8SP) + heighten (+8SP) = 16SP (plus some other benefit; delay, or whatever). In the new order it would be enhance (+4SP) + heighten (+12SP) = 16SP.

Yesterday's thoughts on the bus. Feel free to shoot them down - they might prove useful, nonetheless:

Kernels. A different tack. Weak metamagic feats incur +1SP per metamagic level adjustment; medium metamagic feats incur +2SP per metamagic level adjustment; strong metamagic feats incur +3SP per metamagic level adjustment.

Weak metamagic feats are Heighten, Delay, Widen, Still and Silent

Medium metamagic feats are Enlarge, Quicken, Extend

Strong metamagic feats are Empower, Maximize

I think Heighten should be +1 in kernel analysis, all level-dependent variables should be set to maximum, and we should drop Enhance altogether and replace it with Empower (at +6). This would make one Heighten worth 1/6 of a full spell level, which seems appropriate - it's a weak feat, after all. Delay is a +3 level metamagic feat but it's way overpriced; I'd tag it at +3 for kernel analysis purposes; honestly, I think it should be a +1 level metamagic feat.

Delayed Blast Fireball
20 + 12 (twice Empowered) + 4 (Heightened) + 3 (Delayed) = 39

Disintegrate
20 + 12 (twice Empowered) + 8 (typeless) +8 (skewed) +1 (cannot be raised) -6 (targeted) +3 (Heightened) -2 (Range) = 44

(Disintegrate is a solid 7th-level spell, and would be better balanced as a touch attack. Skew should be +8 - average damage is 22.5d6, assuming 50% of targets make their save; average damage on an unskewed spell is 15d6, assuming the same - in the long run, skewing is as effective as adding two more Empowers to the kernel, but involves an element of risk.)

Horrid Wilting
20 + 12 (twice Empowered) +8 (typeless) +5 (Heightened) +8 (unrestricted # of targets) +3 (Widened) -2 (living creatures only)= 54

Horrid Wilting should really be 9th level.

Cone of Cold
20 + 6 (Empowered) + 2 (Heightened) + 1.5 (half-Widened) -6 (range) = 23.5

Cone of Cold is crap.


Thoughts?

 

[Cheiromancer]

I'm frustrated with kernels right now (I'm especially tired of anything related to poison, disease or hold monster) but I'm sure it will pass. There's a stylistic difference in newer books; it's not just that there are new factors (like swift casting times and spells that can be dismissed for a quick boost)- rather, they feel like they are put together differently. It could very well be that the implicit formula has changed over time; it may also be different from designer to designer.

Anyway, the proof of the pudding is in the eating, and if we can get good seeds and factors that make nice spells, that would be great.

One desideratum: clouds. At 20th level Incendiary cloud does 80d6 damage, but spread over 20 rounds. It's big selling point is that it ignores SR. I have a feeling that ignoring SR should be like ignoring saves in that it is not just a constant factor but a multiplier (maybe a doubling) of the base seed. If the mobility of the cloud were worth +4 (which seems reasonable; it adds greatly to the utility of the spell), then you'd have

Incendiary Cloud
(20 fireball + 4 mobile) x 2 = 48

I don't know how to phrase the wording of the "cloud multiplier", though. It would be a lot easier if it did 5d6 damage instead of 4d6. Then you could say that it does half damage per round, but has a duration that lasts 1 round/level. But 40% seems contrived. And the implicit heightening hasn't been priced, and the range is wrong (it should be long; but if you change the fireball kernel to 18 the numbers don't work out). It is suggestive, but the precise details are elusive.

Cloudkill is another example; it does a fair amount of Con damage each round, 1d4 (save for half), or a mean of 3.75 damage. Assuming this is 40% of something, and assuming 1 point of Con damage is worth 2SP, then this component is worth 9SP (9.375 actually). Its movement can't be directed by the caster, but it does have an esoteric effect that kills low HD opponents- I'm going to ignore it for now. It's formula might look something like

Cloudkill
(9 Con damage + 6 area effect) x 2 = 30

So, again, nothing about range or heightening saves, so it's kinda inconclusive. Dunno. Maybe "no SR" is x1.5, not x2. Maybe the heighten factor is added after multiplying. That might work; suppose that the base damage for incendiary cloud is 20d6, and the damage done per round is 1/5 of base. Then

Incendiary Cloud v1.1
(20 fireball + 12 double empower + 4 mobile -2 range shortened one step) x 1.5 + 5 heighten = 41 + 5 = 46.

Pretty close. And then cloudkill would be

Cloudkill v1.1
(9 Con damage + 6 area effect + 8 range) x 1.5 = 34

Even though the details are sketchy, these spells are highly valuable to metamagic specialists - an energy substituted, massively empowered incendiary cloud is a great way of dealing with swarms, mopping up isolated groups of mooks, damaging golems and so forth. And you can do this without worrying what the internal factors are.

 

[Sepulchrave II]

That'll be my new proposal; using UK's CRs as written, and have them be the benchmark.

OK, let's give it a shot.

Do you have volume 1 of the IH:Bestiary?

Yep.

 

[Cheiromancer]

Ok. I have hp, HD, saves and SR for the dragons (man, there are a lot of dragons!), most of the creatures in the ELH, the SRD giants and principal demons and devils. I'm gonna add the hydras pretty quick. And a few other critters here and there.

It seems that the Gibborim's CR is underrated; compare to what UK thinks an ancient red dragon should be. I've asked about that over at his forum. I hope his CRs are generally trustworthy.

[edit] It seems that in the bestiary he multiplied the CRs by 2/3, to approximate the method that WotC uses to calculate the CR of "boss monsters". The real value for the gibborim is 73. Multiply by another 1.24 1.5 if you want to give it PC equipment.

That makes a lot more sense.

[edit2] I'm chatting with him right now on Messenger, and it seems some details are a bit murky- it's been over two years since he wrote v5. I'm gonna go with my interpretation of the numbers.

[edit3] He says he multiplied all the numbers by 2/3 to agree more closely with WotCs numbers. But that, i.e. an 18th level character is CR 12. It's like converting between two different currencies.

I think most of the discrepancies between the two are due to the way that WotC calculates high CR monsters; that they are boss monsters. I'm gonna try to stick with the v5 values, or 1.5 times the values of the Bestiary.

 

[Sepulchrave II]

Epic Warlock is up.

 

[Rikandur Azebol]

Unfortuantely for Epic Warlocks many of those feats are ... unbalanced to say it lightly. For example, Shadowmaster. For me it screams Salient Divine Ability, litte weaker but noone the less.

Another example is Epic Fiendish Resilence, wich quintuples benefits of the class ability. Where standard Epic feats aren't that generous ... but it is not that overpowered, after all 'lock needs his healing. Or Tomb-tainted soul feat.

P.S. Are You following U_K's rule about 1 epic feat equals 2 normal ones and Salient Divine Ability equals roughly 3 to 5 epic feats ?

 

[Cheiromancer]

We've more or less came to the conclusion that 1 epic feat equals 2 normal ones. Not entirely independently of UK (we've both read his stuff), but not just because he said so either. We haven't talked about divine abilities yet.

Here's a messload of monsters with CRs and, in many cases, hp, HD, saves and SR. I concentrated on the higher CR monsters to see what their saves and such would be.

This is CR = ECL. If a 13th level party encounters a CR 13 monster, they'll use about 20% of their resources. If it were WotC CR, a CR 13 monster would use about 50% of their resources. The CR that UK uses in the bestiary is 1.5 x CR = ECL. I converted it back for this file.

If we end up discarding this idea, it's easy enough to get excel to change the numbers back.

[edit] Oops. The small earth elemental somehow got the stats for an elder earth elemental.

Using the handy little chart function (XY scatter) it looks like HP/CR = 12 hp, and Fort = Will = 0.5 * CR. Ref = 0.4 * CR

With a lot of exceptions, that is. But those look like decent estimates for these statistics. And SR seems to stop being a factor at higher levels. Its basically looks like 20 + 0.4*CR.

So if an Nth level caster can manage save DCs of 10 + N/2, then equal CR opponents will save about half the time. Both of which are quite reasonable expectations, I think.

 

[Sepulchrave II]

So if an Nth level caster can manage save DCs of 10 + N/2, then equal CR opponents will save about half the time. Both of which are quite reasonable expectations, I think.

You're going to hate me.

If Matt (50th level) can prompt a Save with a DC of 59 with his 9th-level spells (19 +20 Int +18 Heightened +2 GSF) then we've got serious problems; he's 24 points clear of the 50/50 zone. With SR a wash, his spells are unstoppable.

Even without his headband and no stat-buffs running, Matt forces a save w/ a DC of 50. The very old gold dragon (KR50; new KR = 1.5 x old KR) still autofails (Will +26) - this can't be right.

This is gonna sound really weird, but I think AMC works better with WotC CRs than it does with Krusty's own CRs. MM lists a very old gold dragon as CR 22.

Stripling Matt
A stripling Matt (22nd level, MF) with GSF (Enchantment) can force a save against his dominate monster with a DC of 30. Intelligence of 28 (15 +5 level +6 headband +2 Inherent).

Modestly-Advanced Matt
Matt at 27th-level (MF, AMC x4) now has an Intelligence of 32 (15 +6 level +6 headband +5 Inherent). His DC is 36.

Now, I'm not suggesting that the CR of a very old gold dragon should be 27 based on the saving throw for Modestly-Advanced Matt's dominate monster, but you get my point.

Edit: without belabouring said point, an old prismatic dragon (WotC CR 48) has a base Will save of +50. Anecdotal, I know; illustrative nonetheless. In practice, the dragon might have a mind blank up, or a cloak of epic resistance +8, and might laugh at 50th-level Matt's puny attempts to dominate him. There again, he might not.

Maybe we shouldn't count ability bonuses from items or spells at all when figuring the power curve of the metamagic specialist (or anyone else): hopefully, there will be a natural balancing achieved against the target's cloak of resistance or protective spells, or inherent bonuses, or whatever.


Addendum: I am reminded of the fragility of the notion of game balance at these levels; that CRs - or KRs - really don't mean too much beyond a certain point. Sure, there are CRs that I think are wonky in the ELH: I boldly assert that I could guesstimate their challenge in play with reasonable accuracy against a group of 4 characters better than any formula.

 

[Cheiromancer]

Suppose we look at that list of monsters and say that the question is what to scale the CRs by. I.e. Should we keep the numbers as they are, or should we multiply by 2/3, or 1/2, or what?

Your remarks indicate your belief that a very old gold dragon can't possibly be CR 50, because 50th level Matt is guaranteed to dominate the dragon in a single round. According to this reasoning, Matt isn't CR 50 either, since his saves and SR aren't anywhere high enough either; if Matt encountered his double first thing in the morning (or after a mordenkainen's disjunction) he'd be able to dominate his double a lot more easily than he would a dragon.

Basically, Matt is a guy with a gun. A guy with a gun can kill another guy at the cost of a single bullet. Even if the other guy is the same level as he is. His main problem is not in dominating the gold dragon; it's in not being dominated (or energy drained, polymorphed, petrified, blasphemied or imprisoned, or any number of other things) when some other 50th level character decides to come after him.

But that's OK. Let's say that CR is the ability to withstand bullets, and the dragon comes up short. Maybe if Matt retires, a very old gold dragon would make a great PC to replace him; all that shows is that ECL is different from CR. It matters what side of the trigger you are standing on.

Now, before we decide on criteria for determining what the scaling factor should be (whether to multiply that list of CRs by 2/3 or 1/2 or leave it as it is), I want to make sure we agree on a certain philosophical stance. Should have done it about 7 pages ago, but better late than never:

The basic assumption of the WotC system is that adventuring occurs in a very narrow band of CRs. If something's is your level +8, your party is going to be mincemeat. But if your level is 8 above something's CR, you shouldn't even get any experience from stomping all over them. Such a system assumes that power is exponential; one creature at CR X+4 = 4 creatures at CR X. And the WotC system also assumes that a character's power curve is similar.

Upper_Krust says this isn't so. He says that 1 creature of level 2X is equal to 4 creatures of level X. He believes that the power curve is quadratic. Whatever scaling factor you use (i.e. whether you multiply by 2/3 or 1.5 or whatever) it is still quadratic.

Which system do you think is true? Or, given the ability to craft the rules, which system would you like to make true?

I'm hoping that you are with UK on this. Then we can use that table even if it means using a different scaling factor (2/3 of the listed values, perhaps. Or 1/2, if you find that epic characters can fight above their weight class). I think you are, but I know I'll feel silly that we've worked on this for two months and we still haven't gotten one of the basic presuppositions settled.

Anyway. If we use the dominate monster criterion to say that dragons and similar creatures are 2/3 or one half the CRs listed, then we are implicitly expecting that casters will use most of their AMCs for Heighten Spell, at least when there is a saving throw involved. That means no AMCs for Empower Spell or other feats. Which means that our efforts in working on the additional damage factors for [blast] and [destroy] is totally misguided. We should have been using polar ray and melf's acid arrow as the base spells from the beginning. Or recognize that the damage curve for [blast] is base 10d6 +1d6/2SP. And that [destroy] does 5d6, but is useful for bringing down force effects.

All this would be OK; but it would mean that we were putting our work in the wrong place. Epic spells are going to look quite different if we assume high saves.