Purpose of the Charms school

[Psieye]

TL;DR provided at end of post. There are 2 chapters to this post, feel free to skip the first.

------ Chapter 1: Recap ------

In principle, we like buffs in crawl. They are cool for changing otherwise-dire combat prospects. There are a few sources of buffs in crawl, ordered roughly in order of how 'spam-resistant' they are:
- Racials/Mutations
- Consumables
- Equipment
- Religion
- Spells

A buff needs to meet 2 criteria to 'float' in this community:
1) The buff needs to make meaningful impact.
2) The buff must not encourage tedious habits under optimum play.

Efforts have been made to reduce incentive for tedious equipment hotswapping. Consumables and god powers have limiters to prevent spamming. What of Spells? The strategic cost is EXP (and min/maxing Int) - both to get appropriate skills and sufficient spell slots. After that, they are functionally always available.

We are happy with Conjurations because doing direct damage is a very active step in ending combat. Summons are slightly less direct, but still feels substansive towards solving problems. Hexes are similarly indirect but bring powerful effects which don't always work (MR). A lot of Necromancy requires extra resources/conditions to achieve "powerful effects that can't be spammed". In general we are happy with how these schools function. Most need an enemy to be cast and those that make allies can't be abused with stairdancing (where most 'out-of-combat' preparations are done).

Transmutations (specifically, shapeshifting) apply buff packages that each apply a lot of upsides and downsides. None of those upsides bring "you gain ranged attacks" on their own. You can also only be one shape at a time. As such, even though you technically could spam them or cast them out-of-combat, there is minimal incentive to.

So we get to Charms. In the old days, they were effectively 'trade spell slots for small buffs requiring keystrokes to renew'. What habits did that nudge us towards?
- "Spam a variety of individual Charms, then walk - keep resting and casting them as need be." -> tedium
- "Spend turns in combat doing nothing but switch on a single word in your status bar that meaningfully affect your follow-up decisions." -> lack of impact

The strategic cost to Charms is mostly "pay and forget" (until you need ?Amnesia). Thereafter, gameplay only sees the tactical key-mashing that yields tedium. Devs have been removing the worst offenders and are trying to knead in meaningful downsides to the ones left.

Players who don't take things to the logical extremes of optimum play keep suggesting new Charms ideas with downsides hurriedly slapped on as an afterthought. Downsides like "drains your MP" or "a little contam" which translates to "play even more tediously to optimise". I've also seen "really short duration" but without any other restriction that just means "spend even more turns in combat pressing buttons that don't directly killdudes -> more tedium".

Perhaps the question for now buffs is not "what's the downsides" but "why can't this be a consumable or a god power?"

Who knows, once we 'solve' Charms we might find issues with the other schools to chew on. But for now, Charms is the problem at the forefront. Specifically, we know what we don't like about it but we're not sure what direction to take it instead.

------ Chapter 2: Charms Niche ------

What should we do about the Charms school? Let's look at successful buffs from Spells:
- Discourages/forbids out-of-combat casting
- Highly situational, because meaningful downsides or requirements

Do we have design space that doesn't step on the toes of other schools? These design spaces are nominally taken:
- Needs enemies: Hexes
- Needs corpses: Necromancy
- Big bundle of upside and downside: Transmutations
- Deals damage: Conjurations

With the removal of the Haste spell, it seems there are no single school Charms beyond the Skald's starting book. It exists to be a thread linking buffs in various other schools and to make the spell harder for single-school specialists. Perhaps that's not a problem: "put EXP in this if you want buffs, but they all require setup or have downside". If we mandate every new Charms to be a hybrid, this forces us to think of the pre-existing meaningful restrictions the other schools have established.

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

TL;DR Bad Charms ideas are tedious when optimised. Want to make buffs big and unspammable. Suggest new Charms ideas be consumables or god powers instead. Otherwise mandate they be hybrid school spells with thematic restrictions.

[VeryAngryFelid]

I've also seen "really short duration" but without any other restriction that just means "spend even more turns in combat pressing buttons that don't directly killdudes -> more tedium".

That's decision-making, not tedium. Do I want to walk to better terrain? Do I want to cast Blink? Do I want to use a consumable? It's repetitive "press tab to attack" which is tedious.

Perhaps the question for now buffs is not "what's the downsides" but "why can't this be a consumable or a god power?"

Most consumables have downside of lacking decisions about training skills (potions, scrolls) and I don't think having a separate god for buff will work provided we already have Okawaru (double damage from finesse, extra offense/defense from heroism) and Qazlal (RMsl, +SH, resists).

With the removal of the Haste spell, it seems there are no single school Charms beyond the Skald's starting book. It exists to be a thread linking buffs in various other schools and to make the spell harder for single-school specialists. Perhaps that's not a problem: "put EXP in this if you want buffs, but they all require setup or have downside". If we mandate every new Charms to be a hybrid, this forces us to think of the pre-existing meaningful restrictions the other schools have established.

Dual schools don't solve the issue as you can see many blaster casters who have Conjurations as their highest skill. Triple schools level 6-7 spells can work probably.

[Psieye]

I've also seen "really short duration" but without any other restriction that just means "spend even more turns in combat pressing buttons that don't directly killdudes -> more tedium".

That's decision-making, not tedium. Do I want to walk to better terrain? Do I want to cast Blink? Do I want to use a consumable? It's repetitive "press tab to attack" which is tedious.

Hmm, 'what counts as tedious?' is more ambiguous for this community than I had assumed.

I don't think having a separate god for buff will work provided we already have Okawaru (double damage from finesse, extra offense/defense from heroism) and Qazlal (RMsl, +SH, resists).

That's my point - "why should your new buff idea not just be added to Okawaru (optionally, replacing an existing god ability) or another pre-existing god?"

[VeryAngryFelid]

Hmm, 'what counts as tedious?' is more ambiguous for this community than I had assumed.

I am not sure we actually disagree here. If you always cast the buff, then it is tedious indeed. I assumed that the buff is not a no-brainer so you don't always cast it, sometimes you do and sometimes you don't.

[Psieye]

Hmm, 'what counts as tedious?' is more ambiguous for this community than I had assumed.

I am not sure we actually disagree here. If you always cast the buff, then it is tedious indeed. I assumed that the buff is not a no-brainer so you don't always cast it, sometimes you do and sometimes you don't.

Ah, yeah I was thinking of buff ideas which are no-brainers "except the duration is ultra short". Vine Stalkers (and other, less effective ways) get around "you'll run out of mp by spamming this" so you're just left with "always spam this whenever you can, it's always more efficient than tab spam".

[dpeg]

So we get to Charms. In the old days, they were effectively 'trade spell slots for small buffs requiring keystrokes to renew'. What habits did that nudge us towards?
- "Spam a variety of individual Charms, then walk - keep resting and casting them as need be." -> tedium
- "Spend turns in combat doing nothing but switch on a single word in your status bar." -> lack of impact

Sure thing, but lack of impact is not a given. For example, berserk rage is just a word on the status bar, but it means a lot. Closer to existing spells, Confusing Touch (even though it is a Hex). "Lack of impact" does not come from "stands on status bar", but rather from "too small an effect to affect your follow-up decisions", in my opinion. E.g. Repel Missiles is important enough to fire up, but once you have it, you don't need to think about it (until it runs out).

The strategic cost to Charms is mostly "pay and forget" (until you need ?Amnesia). Thereafter, gameplay only sees the tactical key-mashing that yields tedium. Devs have been removing the worst offenders and are trying to knead in meaningful downsides to the ones left.

Players who don't take things to the logical extremes of optimum play keep suggesting new Charms ideas with downsides hurriedly slapped on as an afterthought. Downsides like "drains your MP" or "a little contam" which translates to "play even more tediously to optimise". I've also seen "really short duration" but without any other restriction that just means "spend even more turns in combat pressing buttons that don't directly killdudes -> more tedium".

This is good, but let's try to be more systematic about it.

• Strategic cost: experience / skilling choice: This matters, but I don't think this will ever be the correct cost. For example, no matter what XL a charm has, there will be a time when investing in Charms for that spell is too expensive, and a time after which the spell is easy pickings. This does not make for interesting decisions.
• Tactical cost: maxMP reduction: This comes up all the time. The main issue is that there are characters who don't need MP for anything else. In fact, such characters often have great use for charms. In other words, this is another cost which does not really work.
• Tactical cost: contamination: This can work, as everyone can see from Haste and Invisibility. However, it is a heavy-handed solution, and I don't think I'd like this cure applied to all charms.
• Tactical cost: MP drain: Now I think this is a really good idea: no matter whether you use MP for anything else, you total and current MP suddenly matter. Moreover, you cannot slap on very many spells with this mechanic. I could imagine a Charms school where (almost) all spells get this.
• Tactical cost: turn: This is about really short durations. I disagree with your assessment. Let's look at the most extreme version: the effect lasts a single turn. That means you have to use it right away. You spend exactly one turn to power up your next action. If all charms had this, you couldn't pile them up. (I think this fix can only apply to a short list of existing spells, but I don't see the "more tedium" yet.)

Perhaps the question for now buffs is not "what's the downsides" but "why can't this be a consumable or a god power?"

That's not the correct question, in my opinion, because all kinds of things could be consumables or god powers. As I see it, in global design it works like this: gods are like tweaks to the rule set, these could be mild (i.e. close to the core game, e.g. Makhleb's powers) or bizarre) -- they're an extra thing, so have a lot of freedom. Next, consumables are by definition limited, so they can get away with all kinds effects -- controlled blink, and haste are extremely strong, but it is hard to break the game with consumables (and if there are too many, then you can tune the generation rates).
Spells, by contrast, are much harder to design or balance: these effects are intended to be used infinitely often; they are supposedly kept in check by a number of costs (MP, contimination, hunger, spell slots, experience). Old Divinations and Charms have shown that these checks aren't enough.

That said, I think that Charms as a spell school can work. Here are the desiderata I'd like to put up:

1. Spellpower matters for effect (not just duration). This means that skill should matter beyond casting success.
2. Meaningful choice of which charm to memorise, and to cast. (Generally, you now want "all of them". This is bad.)
3. Only useful in battle. Not during travel/exploration. Might be acceptable to cast before you go into battle, but see 2.

[Psieye]

"Lack of impact" does not come from "stands on status bar", but rather from "too small an effect to affect your follow-up decisions"

Yes, this is a good correction.

• Tactical cost: maxMP reduction: This comes up all the time. The main issue is that there are characters who don't need MP for anything else. In fact, such characters often have great use for charms. In other words, this is another cost which does not really work.
• Tactical cost: MP drain: Now I think this is a really good idea: no matter whether you use MP for anything else, you total and current MP suddenly matter. Moreover, you cannot slap on very many spells with this mechanic. I could imagine a Charms school where (almost) all spells get this.

Idea for Tactical cost: "MP = 0" - casting the buff sets your MP to 0 and keeps it there until it expires. Preferably convert all that 'lost MP' into the effect of the buff.

Tactical cost: turn: This is about really short durations. I disagree with your assessment. Let's look at the most extreme version: the effect lasts a single turn. That means you have to use it right away. You spend exactly one turn to power up your next action. If all charms had this, you couldn't pile them up. (I think this fix can only apply to a short list of existing spells, but I don't see the "more tedium" yet.)

Hmm, it seems tedium from 'number of keystrokes' is less a concern than I had perceived. Let's take the extreme case even further: "for 1 turn, gain x10 damage, +100 AC, +100 EV". This is definitely tedious because it's an utter no-brainer for characters who don't have MP concerns. But it's also 'tedious' because you've added extra keystrokes before every action. I suppose if those extra keystrokes are after meaningful decisions, then the additional user input is not something to avoid.

Spells, by contrast, are much harder to design or balance: these effects are intended to be used infinitely often; they are supposedly kept in check by a number of costs (MP, contimination, hunger, spell slots, experience). Old Divinations and Charms have shown that these checks aren't enough.

That said, I think that Charms as a spell school can work. Here are the desiderata I'd like to put up:

1. Spellpower matters for effect (not just duration). This means that skill should matter beyond casting success.
2. Meaningful choice of which charm to memorise, and to cast. (Generally, you now want "all of them". This is bad.)
3. Only useful in battle. Not during travel/exploration. Might be acceptable to cast before you go into battle, but see 2.

I still think 3 is best met by demanding they instantly expire when there are no enemies in sight. 1 is relatively easy to accommodate.

2.2, i.e. "you shouldn't want all these buffs up all at once" without resorting to shapeshifting's hard limit of 1... Hmm, we don't like Contam as a cost because it encourages you to rest a lot. The current Contam cost model is more for preventing repeat usage of Haste/Invis. What if the Contam immediately vanishes when the buff expires? So say all Charms add some (non-random, constant) Contam when cast and subtract that amount on expiry. Having multiple buffs up now sends you to badmut land. Well, it's not as satisfactory as putting in carefully coordinated downsides that are mitigable in isolation but are very punishing when stacked - which ties into 2.1, i.e. "you shouldn't want to memorise all these buffs at once".


I need to sleep here, but I want to review all the Charms buffs (specifically, what they do more than what they're called) currently in-game. Could I trouble someone to summarise them here?

[johlstei]

Lets just make them consumables they'd mostly be fun as consumables. Super rare potions of death's door sound very fun to me.

[dpeg]

johlstein: This is GDD... do you care to flesh out your argument? I don't see what you mean: particularly for Death's Door, the parameters used are spellpower and Kikubaaqudgha's piety. How to emulate with a potion?

[Shard1697]

I still think 3 is best met by demanding they instantly expire when there are no enemies in sight. 1 is relatively easy to accommodate.

Then it might be useful to cache a weak enemy like a goblin on upper floor, letting it follow you down to unexplored floor so you can manually explore with him following you harmlessly to keep your buffs up. Extreme, but not impossible for people to do.

I used to be more attached to the school, but at this point I think it would be very reasonable to remove charms as a school and split off what isn't deleted into other schools, with a bit of reworking.
IMO:

• Remove infusion, shroud of golubria, ring of flames, excruciating wounds
• Repel Missiles is now pure air.
• Swiftness is now air/tmut
• Ozocubu's is now ice/tmut, buff cannot be renewed, when it runs out you become Slow(is this worth having when ice form exists?)
• Regeneration is now pure necromancy, and requires enemies slain to work-like Powered By Death but influenced by spellpower
• Spectral weapon is now Hexes/Summoning
• Death's Door is now pure necromancy, level 9

Also possibly the idea I saw earlier(sorry, forgot by who!) which proposed Repel Missiles as a buff purely for approaching(or fleeing from) enemies which has high chance to ignore projectiles but ends when you attack/cast a different spell. Although this has issues with caching enemies as said above.
Could also become a spell that prevents you from attacking in any way(with really high chance to repel missiles), but in that case that's treading on transmutation's "big upside, big downside" deal, and could instead be like "Air form" or something.

[VeryAngryFelid]

I like Shard1697's idea. Now that Haste spell is gone there is little point in having species with good aptitude in charms IMHO.

[phloomp]

Tactical cost: turn: This is about really short durations. I disagree with your assessment. Let's look at the most extreme version: the effect lasts a single turn. That means you have to use it right away. You spend exactly one turn to power up your next action. If all charms had this, you couldn't pile them up. (I think this fix can only apply to a short list of existing spells, but I don't see the "more tedium" yet.

Just mentioning, since it's similar but hasn't come up: charms could also take a really long time to cast. Even haste wouldn't have been a no-brainer if it took as long to cast as 5 fireballs.

[Shard1697]

That would just mean you had to be sure to cast it before getting into trouble, so you'd reset fights a lot in order to have time. (And you could often eat a bit of damage from 5 turns to haste, run away, rest off contam, haste on other side of stairs, head back to fight)

[Psieye]

I still think 3 is best met by demanding they instantly expire when there are no enemies in sight. 1 is relatively easy to accommodate.

Then it might be useful to cache a weak enemy like a goblin on upper floor, letting it follow you down to unexplored floor so you can manually explore with him following you harmlessly to keep your buffs up. Extreme, but not impossible for people to do.

Ah point. Hmm, that means Charms which are strictly undesirable to cast unless fighting a non-trivial enemy.

New idea:
- Buff cannot be renewed
- Buff keeps track of EXP earnt over its duration
-- Alternative to "EXP earnt": number of attacks landed on non-plant targets
- If fail to earn a 'quota' before it expires, get Serious Downsides (like, losing maxHP or badmuts)

Ozocubu's is now ice/tmut, buff cannot be renewed, when it runs out you become Slow(is this worth having when ice form exists?)

Hmm, maybe go bigger and combine with Leda's Liquefaction:
- +AC while standing in the liquid floor of "slow movement"
- Non-amphibious enemies can 'fumble melee' in liquid floor, but not you
- Cannot fly, -Tele
- Cannot use stairs
- Cannot renew: you have to sit through that tail end when only the tile you're on is liquified

Repel Missiles as a buff purely for approaching(or fleeing from) enemies which has high chance to ignore projectiles but ends when you attack/cast a different spell. Although this has issues with caching enemies as said above.
Could also become a spell that prevents you from attacking in any way(with really high chance to repel missiles), but in that case that's treading on transmutation's "big upside, big downside" deal, and could instead be like "Air form" or something.

Hmm, a 'travel form', i.e. a spell that you can ONLY cast while out-of-combat. Off a few minutes of thinking, some dead-ends I ran into:
- "Out-of-combat = you can't see any enemies" -> duck around a corner, cast Air form to escape on the 1 turn enemies lose sight of you
- "Out-of-combat = no monsters are tracking/chasing you" -> information leak

Then I thought of how body armour hotswapping is discouraged. This gets me to:
- Takes 10 turns (100 aut) to finish casting 'travel form'
-- Can abort casting if enemy sighted

- While casting:
-- AC = 0
-- EV = 0
-- All damage taken x3

- While in 'travel form':
-- Stealth = 0
-- Cannot use stairs
-- Cannot interact with items on floor
-- Cannot use doors
-- DMsl

- Infinite duration until you do anything other than move

[dpeg]

New idea:
- Buff cannot be renewed
- Buff keeps track of EXP earnt over its duration
-- Alternative to "EXP earnt": number of attacks landed on non-plant targets
- If fail to earn a 'quota' before it expires, get Serious Downsides (like, losing maxHP or badmuts)

If you have to introduce a bunch of new rules for a spell school, then it is probably too much effort.

However, I don't think that's necessary: we can simply work off the things that make spells go, and you already said so: most/many Charms could deplete MP, either continuously (like Infusion), or at once. That would make MP relevant, and you wouldn't cast the spell out of the blue (this assumes that resting to get the MP back takes longer than whatever the spell effect was).

I'd like to do a spell-by-spell review of Charms (and perhaps some related spells), but I cannot do that right now. Would also be useful to collect some of the novel ideas that have been floating around.

[Hurkyl]

What does gameplay with good charms look like?

---

I think that's the most important question to get a handle on; it doesn't make sense trying to design charms if you don't know what result you want!

(I have my own vision, but I'm pretty sure it's diametrically opposed to what dpeg wants to see in gameplay so I won't clutter up the thread)

[dpeg]

Hurkyl: This is the place to put down your ideas on what Charms should look like, so please go ahead.

I mentioned what criteria Charm spells should meet, in my opinion. Of currently existing spells, I am happy with the effect and choices provided by: Infusion, Song of Slaying, Spectral Weapon. I have qualms (of different kinds) about Repel/Deflect Missiles, Swiftness, Ozocubu's Armour, Regeneration, Shroud of Golubria, Ring of Flames, Battlesphere (this is after the changes to some of these spells). Buff-type spells that work well, in my opinion, are: Confusing Touch, Passwall, Darkness, Invisibility.

As to the gameplay: ideally, you wouldn't use the same spells for each fight. Positively phrased, I would be happy if only one Charms spell were active at a time, and it's effect would be meaningful to follow-up decisions. Song of Slaying is a good example (despite its shortcomings): you don't want it on every fight, and if you use, it has an impact on what to do next.

[Psieye]

What does gameplay with good charms look like?

I think that's the most important question to get a handle on; it doesn't make sense trying to design charms if you don't know what result you want!

(I have my own vision, but I'm pretty sure it's diametrically opposed to what dpeg wants to see in gameplay so I won't clutter up the thread)

It's certainly very important. A new vision would be welcome - at minimum it would serve as a reference from a different vantage point.


@dpeg: you've labelled Passwall as 'buff-type'. Do Blink and Passage of Golubria also count as 'buff-type'?

[VeryAngryFelid]

Buff-type spells that work well, in my opinion, are: Confusing Touch, Passwall, Darkness, Invisibility.

May I question Darkness as well-working spell? I tried to use it in my last game as Enchanter along with manual exploration and found that I regenerate 5 MP before the spell expires. So I can rest a few turns until I regenerate that 1 MP back and cast Darkness again. This was really tiresome so I just stopped using the spell. It's probably worse than old Swiftness IMHO.

[Hurkyl]

In my opinion, the concept of charms is primarily interesting in the form a strategic choice. If you find a book on D:3, the decision to be a charm user comes at the cost of wearing lighter armor and spending XP on the Charm skill (and relevant secondary spell skills) rather than on things like improving your physical defenses, or making your weapon faster and more accurate, or getting the training to use your biggest blasting spell.

And you hope to make up for all those deficits with the buffs and other effects you get from the charm spells.

I'm, of course, not opposed to making charms tactically interesting -- but, IMO, that is of secondary concern.

The XP and armor costs are the primary ways to put strategic costs on spells, so I mainly think in terms of having effects in accordance with that investment.

(I realize that this has a consequence that low-level charms can be picked up cheaply, but I don't find that any more troublesome than the corresponding fact about weapons or armors or summons or other things)

---

The main gameplay effect I see is strategic things like a player deliberating if they want to stick with their melee build or if they'd rather ditch the chain mail for leather and let their weapon and physical defenses stagnate as they train up charms and assorted spell skills to get assorted buffs and effects.

If done well, it could even lay the groundwork for more interesting strategic choices -- e.g. a melee character dipping into elemental charms could then be in a position to decide whether to put focus back on their melee or to further exploit their existing spell skills by picking up some elemental conjurations. Conversely, access to the appropriate charms could push an elementalist into melee a lot sooner.

---

On the tactical side, I have no problem with charms to include semi-permanent effects that one usually has going, but can peter out in a prolonged fight. In fact, this is something I would actually expect from a charm user.

Additionally, I would expect a dedicated charm user to have a wide variety of effects useful in specialized situations. Maybe even to the point where spell slots are actually a relevant limiting factor.

[dpeg]

@dpeg: you've labelled Passwall as 'buff-type'. Do Blink and Passage of Golubria also count as 'buff-type'?

Blink: no: PoG: yes, perhaps.

An implicit assumption I'm making is that "buffs" don't have an immediate effect -- it's a power you can put on a status light. From this point of view, Lignification, Might, Invisibility, Confusing Touch, Passwall are all buffs. Passage of Golubria is questionable, I wouldn't know myself how to classify it. In my system, Crawl's Teleporation isn't a buff either: while the effect is delayed, it is triggered once and then automatically occurs.

But I am not sure we need a very strict definition of "buff".

VeryAngryFelid: sure, but that could be a detail, i.e. a parameter we can tweak.

[VeryAngryFelid]

VeryAngryFelid: sure, but that could be a detail, i.e. a parameter we can tweak.

I am not sure what parameter you mean. Will I still need to recast the spell manually every time it expires? Probably the spell is primary candidate for testing "toggle to turn it on at the cost of max MP". Decrease max MP by 10-15 and add an ability to turn Darkness off. Spending 1 turn to recast Darkness in fight is not worth the tediousness of repetitive manual recast of the spell IMHO.

[Implojin]

With respect to Charms counterbalance, and adding a form of mechanical dissuasion to casting spell buffs as heavily and as often as possible:


I think that contamination could serve the design purpose of providing an effective counterbalance to the Charms school if it were reworked slightly. Contamination (as a spell penalty) largely doesn't work right now simply because it can be rested off. When contamination does work, it's in combat situations that have gone far downhill, when the player sees a need to recast buffs multiple times before danger has passed. (And also doesn't have any !cancellation, but that's a separate discussion.)

The existing design of contamination doesn't leave much room to dial a balance knob in this area: Either a spell causes enough contamination that it can begin to mutate you (and counteract invis) within one casting, or it doesn't. Glow that isn't severe enough to trigger mutations is largely irrelevant. [Excepting the Invisibility/Glow interaction, which is probably the best gameplay behavior of the current contam design.]

Things like variance in the frequency of mutation gain, or variance in the duration of the resulting mutations, or variance in severity of the resulting mutations, are not taken into account with the way that _magic_contamination_effects() is currently written. I propose that we change this; if we are to use contamination as an effective spellcasting counterbalance, it needs more granularity in its penalty effects.


1) Instead of adding a fixed amount of contamination at each initial spellcast that wears down over time, change contamination to grant a larger amount of contamination at spellcast that does not wear down while the spell remains active. Grant contaminating spells an ability toggle, to turn off both the spell effect and contamination at will.

2) Add the above contamination-while-active effect to all Charms deemed problematic.

3) Next, change contamination-granted mutations from MUTCLASS_NORMAL into MUTCLASS_TEMPORARY. Permanent mutations don't work so well as a spellcasting penalty -- Their results are inevitably either cancelled immediately with !cmut, ignored by the player as irrelevant minor category muts, or forcibly retained as a long-term penalty of unknown duration to be ameliorated only when the floorgen decides to generate !cmut. As a result of this potential for long-term crippling character effects, handing out permanent muts has to be done carefully from the designer's point of view -- You can see this ripple effect baked into the current code here: https://github.com/crawl/crawl/blob/mas ... ts.cc#L790

Temporary mutations as from wretched stars, etc., would serve the purpose of a spellcasting glow penalty much more effectively: They can be handed out in bulk without fear of crippling player characters under the weight of stacks of unresolvable strategic penalties. All the player needs do is kill some monsters to begin removing those penalties. Moreover, as the duration of temporary mutations cannot be reduced by quaffing !cmut, they would actually do a better job of discouraging stacked charmcasting than permanent muts, because the player would be forced to face the consequence of allowing temporary muts to pile up instead of deciding "Okay, I have some !cmut, I can ignore this."

4) Increase the chance per turn of triggering magic_contamination_effects(), based upon get_contamination_level() instead of player_severe_contamination(). By increasing granularity here to check against the several levels of severity that the game *already tracks* instead of checking against only SEVERE_CONTAM_LEVEL, we can fire off alternate resulting effects based upon current severity. [Read: based upon currently active buff spell count.]

5) As one of the above alternate effects, we could increase the experience required to remove granted temporary malmuts based upon current contamination severity.

6) We could also add another mutflag (mutflag::DANGEROUS ?), to enable further granularity in handing out muts that have more potential to be deadly than the (rather broad grouping of) stuff currently grouped under mutflag::BAD.

A shallow glow could result in a chance at some short-duration tier 1 mutflag::BAD results. A middling glow might result in a higher chance at medium duration tier 2 mutflag::BAD. A very deep glow might result in a high chance at a longer duration tier 3 mutflag::DANGEROUS, coupled with a lower chance for an additional tier 2 mutflag::BAD. Mix and match as necessary for balance later; this is the balance knob these changes are trying to install.


The culmination of the above changes is to create a balance shim: By decoupling the rather difficult job of balancing spells based solely around their effects, from the concomitant gameplay choice of how frequently those spells ought to be used, we would open up the resulting gameplay to allow for more depth in possible strength of spell effects without automatically making them things that every player wants to cast as often as possible. (Side note: This solution could also be extended to things like Summoning and Translocations later.) (Double side note: This might also allow for player-castable Haste to be brought back, if the right balance can be found.)

With an approach like this, the player would want to activate their Charms as late as possible preceding a dangerous situation. They would want to leave their Charms active for as short a duration as possible* once the battle has begun - and they would be encouraged to think about which Charm would provide the most benefit in the current situation, instead of simply casting every Charm they know. They could, of course, still choose to activate everything in dire emergencies, but they would pay for it later with multiple, more durable, severe temporary malmuts to fight through after the battle is won.

*(Possibly shutting them off before the combat is fully resolved! This would be a very good thing, especially if we add a Tornado-style recasting timer to increase the choice depth of "Do I shut this off now, or do I need to wait?".)


There are ripple effects to consider here, but they seem approachable:
- We probably don't want to change the existing contamination+invisibility interaction, as that plays reasonably well in current Crawl.
- !cancellation would have to be changed to function under this system.
- Interaction with undead races: To use temp muts as a balance tool in this way, it would be simplest to change them to also apply to undead.
- Interaction with Zin: Thematically, forbidding tempmut spells under Zin would seem appropriate, but perhaps there's an alternate solution?
- +++?

edit: ripple effects

[Psieye]

In my opinion, the concept of charms is primarily interesting in the form a strategic choice. If you find a book on D:3, the decision to be a charm user comes at the cost of wearing lighter armor and spending XP on the Charm skill (and relevant secondary spell skills) rather than on things like improving your physical defenses, or making your weapon faster and more accurate, or getting the training to use your biggest blasting spell.

And you hope to make up for all those deficits with the buffs and other effects you get from the charm spells.

The XP and armor costs are the primary ways to put strategic costs on spells, so I mainly think in terms of having effects in accordance with that investment.

The main gameplay effect I see is strategic things like a player deliberating if they want to stick with their melee build or if they'd rather ditch the chain mail for leather and let their weapon and physical defenses stagnate as they train up charms and assorted spell skills to get assorted buffs and effects.

If done well, it could even lay the groundwork for more interesting strategic choices -- e.g. a melee character dipping into elemental charms could then be in a position to decide whether to put focus back on their melee or to further exploit their existing spell skills by picking up some elemental conjurations. Conversely, access to the appropriate charms could push an elementalist into melee a lot sooner.

Ah, I see why you were reluctant. Yes, recent dev work has run counter to this position. Let's take a closer look at this vision. It boils down to "you find something on the floor, it makes you seriously deliberate on changing your EXP plan so you become a hybrid". For this to be considered interesting, it must pull you in the midgame. "I started out as a melee dude but learnt some Charms" is not interesting when you do it in lategame or extended: it's a no-brainer by that point.

Your vision for Charms would also have to compensate for the Str vs Int imbalance of the starting conditions. The lack of Int on a melee start is keenly felt compared to the lack of Str on a book start. Combined with the abundance of a gradual scale of "light -> heavy armour", it's not hard for a book start to go hybrid if they desire it and compensate for the lack of Str. It's already common sense that everyone should train some Fighting which further assists the "book -> hybrid" transition. The prompt a melee start needs to go hybrid early on is a spell that doesn't care about spellpower (and ideally spell success) because of low Int. Those spells are found in Translocations (blink) and Necromancy (animate dead) already. These schools also offer very powerful tools in the endgame so the EXP does not feel 'wasted'.

I think that's why devs want to move away from your vision: an interesting dilemma in the midgame falls away to a tedious no-brainer (both for strategy and for tactics) in the lategame. Perhaps "the midgame prompt that pushes you to hybridise earlier" is better served with new equipment ideas, like medium armour of Wizardry (but only for one school).

A further detail in your vision: "by having this Charms bridge in the middle, a character is incentivised to tech both melee and conjurations". I think that end is best served with more blunt designs - conjurations (i.e. not 'buffs') which explicitly demand good melee to get the most out of. Though 'spells which really want you to be good at melee' already exist: Transmutations. It's just that most shapeshifting is geared for UC, so perhaps a new Lv 3 transmutation that allows weapon wielding (with whatever other upsides and downsides) would be the prompt you seek.

[Psieye]

The culmination of the above changes is to create a balance shim: By decoupling the rather difficult job of balancing spells based solely around their effects, from the concomitant gameplay choice of how frequently those spells ought to be used, we would open up the resulting gameplay to allow for more depth in possible strength of spell effects without automatically making them things that every player wants to cast as often as possible.

With an approach like this, the player would want to activate their Charms as late as possible preceding a dangerous situation. They would want to leave their Charms active for as short a duration as possible* once the battle has begun - and they would be encouraged to think about which Charm would provide the most benefit in the current situation, instead of simply casting every Charm they know. They could, of course, still choose to activate everything in dire emergencies, but they would pay for it later with multiple, more durable, severe temporary malmuts to fight through after the battle is won.

Hmm yes, I do believe we have a breakthrough idea here. The other schools have their own brand of limiters, Charms overall having "you get nuked by temp malmutes" as its brand is very interesting. Especially the part on making even tiny amounts of contam have a chance to induce temp malmutes.

Possibly unnecessary extra detail:
- Each (serious) Charm gets an "if you get temp malmutated by contam, you get this specific temp malmut on top of that" if you've memorised it.
- Or switch "if you've memorised it" to "if you've got this buff up" if the former is too harsh.

Oh and Zin is another ripple effect but I'm sure we can work that out. Irradiate may need adaptation too.

[VeryAngryFelid]

Random temp mutations really scare me (berserk, teleport). Maybe provide specific "mutations" for each charm spell similar to Ozo armour and Swiftness?

[Psieye]

Random temp mutations really scare me (berserk, teleport). Maybe provide specific "mutations" for each charm spell similar to Ozo armour and Swiftness?

By sorting the temp malmutes into re-designed tiers, the message would be "if you're conservative with buffs, you won't ever get berserkitis or teleportitis".

[VeryAngryFelid]

By sorting the temp malmutes into re-designed tiers, the message would be "if you're conservative with buffs, you won't ever get berserkitis or teleportitis".

I think this is a good idea which can be applied to all mutation sources, potion of mutation should not give robust or evolution probably.

[dpeg]

VeryAngryFelid: sure, but that could be a detail, i.e. a parameter we can tweak.

I am not sure what parameter you mean. Will I still need to recast the spell manually every time it expires? Probably the spell is primary candidate for testing "toggle to turn it on at the cost of max MP". Decrease max MP by 10-15 and add an ability to turn Darkness off. Spending 1 turn to recast Darkness in fight is not worth the tediousness of repetitive manual recast of the spell IMHO.

This is one way to do it. The way I like better, because it is more local and less strategic: casting Darkness reduces MP while it's on, and the effect goes out when MP runs out. In this system, there is still a parameter to tweak (the rate of MP depletion).

[VeryAngryFelid]

This is one way to do it. The way I like better, because it is more local and less strategic: casting Darkness reduces MP while it's on, and the effect goes out when MP runs out. In this system, there is still a parameter to tweak (the rate of MP depletion).

Personally I dislike MP depletion as it makes the spell much more useful for melee characters than for true hybrids/casters, it reminds me Minotaurs in CPA with nothing but Haste. I don't see a problem in casting Confuse, Invisibility, Silence or even Airstrike/Bolt of Cold under Darkness.

[dpeg]

Implojin: Yes, that is an interesting direction, especially because it also works off an existing mechanic (contamination from casting, most notably for Haste and Invisibility). I think there are various ways to keep Charm spells in check as well as interesting...

There is one potential drawback to temporary mutations (I never found them crippling in the Abyss, in my experience you can erase them well enough): players can leave some xp-rich parts of the dungeon to clear temporary mutations. (Of course, the real culprit is having such safely gathered xp in the first place.)

It would be very cool if some schools had discernible, relevant and flavour-ful mechanics: for example, Necromancy spells apply self-damage, Charms cause temporary mutations.

[dpeg]

This is one way to do it. The way I like better, because it is more local and less strategic: casting Darkness reduces MP while it's on, and the effect goes out when MP runs out. In this system, there is still a parameter to tweak (the rate of MP depletion).

Personally I dislike MP depletion as it makes the spell much more useful for melee characters than for true hybrids/casters, it reminds me Minotaurs in CPA with nothing but Haste. I don't see a problem in casting Confuse, Invisibility, Silence or even Airstrike/Bolt of Cold under Darkness.

This is true, but it's still much better than the other, more often proposed change: casting a buff spell turns it on forever, but reduces max MP.

Because with the latter mechanic (max MP reduction), your Minotaur, who doesn't use other spells, casts buff X once and has it forever, whereas with MP depletion, the spell actually runs out. Interface-wise, max MP reduction is better; balance-wise, it is much worse, in my opinion: with MP depletion, you can end up in a situation where you don't have the buff up, *and* MP is low.

[Psieye]

The use of MP as a buff cost loses meaning because of Charms-only lategame melee dudes who have fat MP bars for one purpose only: buffs. Since we don't want that tedium, we cannot rely on MP as a limiter. We could do weird things like "while Darkness is up, non-magic damage you deal is nerfed" and "once this buff expires, set MP = 0". But since inventing lots of new rules is undesirable, we need something simple yet effective. "Extra MP costs" is not effective whether that's from MP depletion or maxMP reduction.

It would be very cool if some schools had discernible, relevant and flavour-ful mechanics: for example, Necromancy spells apply self-damage, Charms cause temporary mutations.

I believe the other schools already had that flavourful justification of limiters. It was just Charms needing something.

The hyperoptimal play of "I stack so many temp malmuts on me because I've reserved cheap EXP elsewhere" is a symptom of "stacking all these buffs lets me dive way out of my usual depth". I think it's just a reminder that we should not go overboard with Charms if we do adopt the Malmut Nuke as the limiter.

[Implojin]

There is one potential drawback to temporary mutations (I never found them crippling in the Abyss, in my experience you can erase them well enough): players can leave some xp-rich parts of the dungeon to clear temporary mutations. (Of course, the real culprit is having such safely gathered xp in the first place.)

I would consider this both a benefit and a drawback: If the player leaves XP-rich portions of the dungeon uncleared, it directly means that their character is weaker than it could be.

The drawback of not wanting to encourage backtracking could potentially be addressed by changing the tempmut clear condition to only decrement when killing "challenging" monsters, if it doesn't already function that way.

[VeryAngryFelid]

"Extra MP costs" is not effective whether that's from MP depletion or maxMP reduction.

I agree with this. I wonder what position devs have about cooldowns. Turn-based cooldowns lead to annoying retreat and resting but XP-based cooldowns work fine for lamp of fire and Ru's draining is similar to XP-based cooldown too. Probably we might invent some new cooldowns like "Explore N new tiles", "Use N scrolls/potions" or "Enter new floor".

[Implojin]

"Extra MP costs" is not effective whether that's from MP depletion or maxMP reduction.

I agree with this. I wonder what position devs have about cooldowns. Turn-based cooldowns lead to annoying retreat and resting but XP-based cooldowns work fine for lamp of fire and Ru's draining is similar to XP-based cooldown too. Probably we might invent some new cooldowns like "Explore N new tiles", "Use N scrolls/potions" or "Enter new floor".

Additional exploration-based cooldowns would work very well as Charms limiters, in my opinion.

The reason I settled on temporary malmuts above (really just an XP-gain timer) is because they are already implemented and easy to translate into the existing Glow system. Alternate exploration cooldowns would work just as well, though!

[VeryAngryFelid]

I think exploration-based cooldown for Darkness might be very flavorful and balanced, you would basically cast Darkness before entering a new floor, opening a door or after finding a group of dangerous monsters. It would be a bad idea to have Darkness active during normal exploration because that would mean you might be on cooldown when you actually need the spell.

[Psieye]

Alternate exploration cooldowns would work just as well, though!

There's also the considerable advantage that players are already used to this with recent "piety by explore" developments.

[Psieye]

I'll go hunting for other links later, I'll just post this one for now: viewtopic.php?p=295428#p295428

Arbitrarily recastable duration spells cost max mp, they are cancellable, they do not time out, but they can be broken by combat actions like rmsl and shroud. If they're broken, you cannot no longer cancel or recast them and their max mp cost is not refunded until they automatically recast after a cooldown.

As for appropriate combat breaking actions, dispelling breath attacks already do this. Other sources could be figured out. The tactical issue is not casting, which is not a real choice for existing charms, but the possibility of getting your spells broken. The decision is whether or not to disengage, which is more interesting than whether or not to cast/recast a charm.

In other words, the limiter is "I've just lost my buffs due to avoidable In-Combat Conditions, I can't renew them in this fight, do I disengage or finish it?"

Could perhaps be extended to: "due to avoidable In-Combat Conditions, my buff has changed into a nerf, I can't fix that in this fight, do I disengage or finish it?"

Provided the risk of Condition triggering was significantly mitigable via some conscious effort, this wouldn't feel like "I must tediously reset fight until my buff lasts long enough".


On a related note, Death Channel was grouped in the same 'arbitrary duration spell' as bad Charms. Given Death Channel has strong incentive to "wipe the whole floor in one go before your army expires", I think it's fine as is: it's a bad idea to rest after a fight generated spectral pets.

[VeryAngryFelid]

On a related note, Death Channel was grouped in the same 'arbitrary duration spell' as bad Charms. Given Death Channel has strong incentive to "wipe the whole floor in one go before your army expires", I think it's fine as is: it's a bad idea to rest after a fight generated spectral pets.

I am not sure I like the "don't heal to maximize benefit" idea, it increases gap between weak and powerful characters, Death Channel is especially problematic because of snowball effect.

[dpeg]

The use of MP as a buff cost loses meaning because of Charms-only lategame melee dudes who have fat MP bars for one purpose only: buffs. Since we don't want that tedium, we cannot rely on MP as a limiter. We could do weird things like "while Darkness is up, non-magic damage you deal is nerfed" and "once this buff expires, set MP = 0". But since inventing lots of new rules is undesirable, we need something simple yet effective. "Extra MP costs" is not effective whether that's from MP depletion or maxMP reduction.

I'm not convinced yet: sure, the character your describe has MP to spare, let's say 30, all ready to use for buff spells. Casting them in a MP-depletion system means that the more spells you trigger, the faster they run out. And once they've run out, you cannot recast. So it's a difference between one spell with longer duration, or several with short ones.

I think I'm fine with temporary mutations as a special cooldown for Charms. I don't like evoker-style XP-timers as much: that's a mechanic already in use, and "you cannot cast this spell now" does not feel spell-like to me. With glow (no matter if current type or special temp-mutation glow), you can cast as long as you have MP.

[dpeg]

There is one potential drawback to temporary mutations (I never found them crippling in the Abyss, in my experience you can erase them well enough): players can leave some xp-rich parts of the dungeon to clear temporary mutations. (Of course, the real culprit is having such safely gathered xp in the first place.)

I would consider this both a benefit and a drawback: If the player leaves XP-rich portions of the dungeon uncleared, it directly means that their character is weaker than it could be.

The drawback of not wanting to encourage backtracking could potentially be addressed by changing the tempmut clear condition to only decrement when killing "challenging" monsters, if it doesn't already function that way.

Your first point is really good. Also, we're slowly but continuously trying to make those areas fewer and smaller.

The bit about "challenging" is really, really hard. Best to avoid it. Tension works so-so for Xom and Demonspawn guardians, let's leave it at that.

[Psieye]

The use of MP as a buff cost loses meaning because of Charms-only lategame melee dudes who have fat MP bars for one purpose only: buffs. Since we don't want that tedium, we cannot rely on MP as a limiter.

"Extra MP costs" is not effective whether that's from MP depletion or maxMP reduction.

I'm not convinced yet: sure, the character your describe has MP to spare, let's say 30, all ready to use for buff spells. Casting them in a MP-depletion system means that the more spells you trigger, the faster they run out. And once they've run out, you cannot recast. So it's a difference between one spell with longer duration, or several with short ones.

Hyperoptimal ("lolsanitywhatsthat") play with MP-depletion-per-turn:
- Cast buff at full MP, out-of-combat
- Take 2 steps into new territory with unknown chance of enemy encounter
- Spam '5' until MP reset
- Repeat until combat initiates with maximum turns to react and near-max MP bar and buff duration

Hmm, to fix this we could do:
- MP depleted per turn while enemy in sight
- Rate of depletion partially (like, 30%) mitigable with super-high spellpower
- Infinite duration while "MP > 20% maxMP"

This would have weird side effects like buffing VS melee dudes, incentive for melee dudes to switch to Sif in lategame and some guy fueled by the insane idea of "MP on kills for charm-stacking!"

[dpeg]

Psieye: That's a good point, although a new one. You've just argued that MP-depletion is not enough to prevent out-of-combat casting. (I was arguing about MP-depleation as a crank *during* combat.)

I accept that, and I guess that slapping building up temporary mutations is a good way to go about this: if you honestly play, then you'll collect xp which reduces temp-glow; this prevents the behaviour you just described.

[Psieye]

MP-depletion-during-combat and Malmut-nuke-for-stacking-buffs would need one extra detail to work together:
- If insufficient MP, the buffs don't do anything. But you still have the "any amount risks temp malmut" glow until the natural expiration of buff duration.

Because "I only want this buff-stack for 2 turns, I don't want it for longer" situations will arise. Being able to spam spells (e.g. the very same Charm spell over and over again) to fast-cancel the downsides would be bad.

[Implojin]

The bit about "challenging" is really, really hard. Best to avoid it. Tension works so-so for Xom and Demonspawn guardians, let's leave it at that.

Sure, you're quite right that it's a hard subject and a separate discussion.

I would like to clarify, though, that I didn't necessarily mean to use Xom tension: There are other proxies for monster difficulty that could be checked; some among them being "absdepth at which this monster begins to spawn", "raw experience prior to player XL stepdown that would be granted at monster kill", etc.

For the purpose of testing monster challenge for decrementing tempmut duration, we wouldn't need it to be a perfect measurement: As one example, just a rough estimate of "Does this monster come from a shallower depth than would be challenging at the player's XL?" might suffice. (e.g. Don't decrement durations if killed monster's minimum spawn absdepth < ((XL*racial_XP_mod)-3))

In the abstract, it's easy for a player to tell if a monster is challenging: Rats basically aren't anywhere after D:1, Yaks basically aren't anywhere after Lair entry depth, etc.; obviously this becomes more challenging to measure accurately if you try to include monster abilities into the challenge rating. The trick here is not to bother caring about monster abilities! We can assume that the devs editing mon-pick-data.h are trying to place depth-appropriate challenges, so the hard work of estimating relative threat has already been done for us. [Minimum monster spawn absdepth] checked against [player XL with a racial modifier to account for XP gain aptitude] thus seems to me to be a "close enough" standin for the question of "Should killing this monster decrement timers?"

The above example obviously isn't a perfect metric: Character *power* at relative XL is highly dependent on racial aptitudes, god choice, and found equipment. Some weak chaff monsters might spawn deeper than they should (This is probably a problem that should be addressed separately still). Uniques don't show up in mon-pick-data.h, but we could easily mark them as "always challenging" and it wouldn't be too inaccurate, or we could grab their minimum vault depth locations. This measurement doesn't need to be perfect, it just needs to be in the ballpark.

(Edit again: Rather than "minimum spawn absdepth", a more appropriate phrasing might be "spawn absdepth of the highest average weighting", to account for the distrib_type differences between FLAT, SEMI, PEAK, RISE, and FALL.)

But anyway, back to the discussion at hand...

edits: clarity

[Psieye]

Here I will aggregate stuff from learndb:

Charms: infusion, repel missiles, shroud of golubria, swiftness, ozocubu's armour, regeneration, spectral weapon, excruciating wounds, deflect missiles, ring of flames, death's door

- Infusion: over duration, trade 1 MP -> +2~3 dmg on melee attacks that hit
- Song of slaying: +slaying per non-magic kill over duration, popcorn stop contributing after a threshold, lots of noise
- Repel Missiles: Enemy ranged attacks suffer -ACC and have (spellpower dependent) chance to expire this buff
- Shroud of Golubria: sometimes challenges enemy melee attacks. After AC reduction, weak damage is likely to be negated, high damage instead cancels buff with near certainty.
- Swiftness: +move speed duration, followed by -move speed duration
- Ozocubu's armour: +AC for -move speed
- Regeneration: faster healing for faster hunger
- Spectral weapon: pet that melees when you do. Its stats scale with spellpower
- Excruciating wounds: apply temp pain-brand, makes noise on cast
- Deflect missiles: like Repel Missiles but better
- Ring of flames: flame cloud immunity and generation, rF++, rC--, Fire enhancer
- Death's Door: invincibility, no heals, set HP to low, cooldown of 1d3 after it expires

Other non-Charms buffs:
- Confusing touch: confuse-brand for UC attacks, resistable with HD - the confuse-inducing attack doesn't deal damage
- Darkness: -2 LOS over duration
- Invisibility: +Contam and invisibility over duration

I mentioned what criteria Charm spells should meet, in my opinion. Of currently existing spells, I am happy with the effect and choices provided by: Infusion, Song of Slaying, Spectral Weapon. I have qualms (of different kinds) about Repel/Deflect Missiles, Swiftness, Ozocubu's Armour, Regeneration, Shroud of Golubria, Ring of Flames, Battlesphere (this is after the changes to some of these spells). Buff-type spells that work well, in my opinion, are: Confusing Touch, Passwall, Darkness, Invisibility.

So, earlygame buffs which only give upside when you actively participate in physical combat are well liked. Darkness has been discussed above as "should probably have MP-depletion-in-combat". Invisibility is the poster child of "interacts with Contam", might need changes with the new Charms-Contam idea mentioned above.


I'll think more later, posting this recap for now.

[Psieye]

A recap of downside ideas that don't seem to fall victim to hyperoptimal tedium or other issues:
- Ultra-short duration
- (Rapid) MP-depletion on seeing enemy, infinite duration until low MP
- Malmut Nuke: Rewrite Contam so it's now "per turn chance to temp tiered malmut" at any level of glow. Give lots of Charms immediate Contam that dissipates very soon after the buff is over.
- Incentivise exploration between casts. Either a hard "can't cast this yet" limit or by taking the above Charm-malmut and declaring it wears off by seeing new tiles.
- Requires kills for benefit
- Buff lasts forever until some consciously avoidable In-Combat Condition, possibly turning buff into a nerf when triggered
- Continuous loud noise

Also, my Contagious Curse of Awesome idea which boils down to:
- Throughout duration, buff does serious harm to you per-turn
- You shave off duration by attacking/killing stuff
- Failure to significantly shave off duration will leave you in a horrible state

[Psieye]

I'm generating numerous (stupid) ideas but they aren't as elegant as a unified "all these good charms have Malmut Nuke attached".

Chargin' Mah Barrier:
- Infinite duration buff that absorbs 33% damage of the next penetrating beam attack that hits you
- On the very next turn, you must perform a ranged weapon attack (or throwing)
-- If you do, your attack benefits from +dmg
-- If you don't, immediately suffer backlash damage (leaving you worse off than no buff) and buff expires

Energise Weapon:
- Like Excruciating Wounds but noise is continuous and duration is super short
-- Instead of loud noise, could just Sentinel's Mark you
- Your temporarily branded weapon causes killed enemies to explode in cone of energy, i.e. AoE damage spreading away from you
- Enemies that don't die on first hit will take extra damage if your temp branded weapon hits them again before the duration expires

Immobilised Armour:
- +AC
- Apply Barbs (Manticore) to you throughout (short) duration

Cursed Swiftness:
- While up, cannot use stairs
- On expiry, summons hostile fast demon which follows you on stairs use

Allergic to Walls:
- +EV
- Standing next to walls damages you (ignoring EV) like in Slime
- The wall reaction also slows you

Degenerate Potions:
- Like Shroud of Golubria, but higher damage threshold and does not dispel on being hit too hard
- All potions act like Degeneration potion throughout duration
-- Mummies can't cast this

[Siegurt]

A recap of downside ideas that don't seem to fall victim to hyperoptimal tedium or other issues:
- Ultra-short duration
- (Rapid) MP-depletion on seeing enemy, infinite duration until low MP
- Malmut Nuke: Rewrite Contam so it's now "per turn chance to temp tiered malmut" at any level of glow. Give lots of Charms immediate Contam that dissipates very soon after the buff is over.
- Incentivise exploration between casts. Either a hard "can't cast this yet" limit or by taking the above Charm-malmut and declaring it wears off by seeing new tiles.
- Requires kills for benefit
- Buff lasts forever until some consciously avoidable In-Combat Condition, possibly turning buff into a nerf when triggered
- Continuous loud noise

Also, my Contagious Curse of Awesome idea which boils down to:
- Throughout duration, buff does serious harm to you per-turn
- You shave off duration by attacking/killing stuff
- Failure to significantly shave off duration will leave you in a horrible state

Also "spell only has any effect of cast in combat" (effect on kill is one way to achieve this, but not the best imho)