Implant hell crawl's charm school mechanism.

[sdynet]

I'm implanting part of the charm school system of hell crawl in my fork: https://github.com/sdynet/crawl/tree/perma-crawl
Download link: https://drive.google.com/file/d/1rNtTrfdQSWcyHetef32m4f2ItH1p6Pdn/view?usp=sharing

There are several changes, but most are similar to the original. I'm going to polish this system up and then PR.
Why I'm doing this? We know 'that some parts of the magic(charm, summoning, necromancy, etc)' require significantly more manipulation than others, which makes users easily tired. So there has been a lot of opinions to solve this problem, but I think it is hell crawl's system that is workable and has the best stability at this point. Here, for a moment, I'll quote it for those who don't know the charm system of hell crawl.

Sets up the framework for permanent buff spells. Cast them once and they stay on forever, reserving some maximum MP. Max mp reserved is listed in the spell info screen. The formula is spell level/success rate^2. Right now this is evaluated as a double, then summed up for all active buff spells and rounded down; more complicated than just using an int for each of them but there's a little less breakpoint nonsense this way. This is recalculated every turn so you cannot cheat the system with equipment swaps. You have an ability to end all your buffs if you have any active, to restore your max (but not current) mp. Your buffs will also expire if your max mp would become negative due to removing an mp ring, getting drained so hard that you need more max mp than you have available, etc.

Spells swapped over to this model so far are infusion, song of slaying, regen, spectral weapon, darkness, deflect missiles, ring of flames.

This was the explanation of the intentions and system I was working on.
The section I'm considering carefully about is consistency. Should only the spell containing charm be maintaining to permanently? If this answer is yes, charm must be added in order for the darkness to maintaining to permanently. The death' door cannot be permanently maintained, so charm must be removed( or change charm to hex).
One of my greed is to apply this system to some parts of summoning. For example, the battle of summoners will be surprisingly pleasant if 'summon small mammal' or 'summon mana viper' is perma-buffs style and the player automatically summons his colleague when he sees the enemy. But these spells do not include charm. In this case, the system is less consistent with the school.
Should the charm school' system be consistent, or should the convenience of spells be prioritized, even at the expense of some of the consistency? This is my concern.

[tealizard]

Applying a spectral weapon-style trigger-to-summon mechanic to most spells in the summonings school is one of the main unfinished tasks in hellcrawl, in my opinion. I would suggest either setting summon limits of one for everything or having a multiple cast/sustain mechanic. I don't think there was ever a clear theory of how the triggering should work, so that part will take some imagination. Trigger on sight might be problematic.

Untargeting most targeted spells is the other big one, but that's a major project. It would probably encompass expansion of melee-to-hex mechanics, fire-to-cast mechanics for some conjurations, and positional non-targeting along the lines of my comments in the controversial gdd threads from about six months ago.

Also, if you're thinking about this turning into a PR for dcss, I'd forget about that. They've dug in on their opposition to these mechanics over the years. No chance of acceptance in my opinion. I think your summoning project would be great for hellcrawl though.

[damerell]

Why I'm doing this? We know 'that some parts of the magic(charm, summoning, necromancy, etc)' require significantly more manipulation than others, which makes users easily tired. So there has been a lot of opinions to solve this problem, but I think it is hell crawl's system that is workable and has the best stability at this point.

We looked at this. Our issue with MP reservation systems, as found in Gooncrawl and Hellcrawl, is that some Charms have so short a duration that under some circumstances casting them constantly as they expire would drain your MP. Regeneration, for example, is much better if you can devote 3 max MP to have it on forever than if you have to pay 3 MP to recast it every time it expires.

[tealizard]

That's not the right analysis of how the costs work, but it's pointless to relitigate the issue now.

[Airwolf]

If you don't do a PR, it definitely won't get accepted.

If that's too many negatives for you, a PR is the best route to getting these changes made, even if it has a small chance of success.

[tealizard]

No, the best route to getting changes made is to create new forks and move them to a more central position with the playerbase or seek out a new playerbase. The particular change under discussion here is not a dice roll. There is a longstanding position on hellcrawl-style charm mechanics and it's not going to suddenly disappear. If you want "crawl" to have these mechanics, you have to change what "crawl" means.

[TBDA]

I'm implanting part of the charm school system of hell crawl in my fork: https://github.com/sdynet/crawl/tree/perma-crawl
Download link: https://drive.google.com/file/d/1rNtTrfdQSWcyHetef32m4f2ItH1p6Pdn/view?usp=sharing
The section I'm considering carefully about is consistency. Should only the spell containing charm be maintaining to permanently? If this answer is yes, charm must be added in order for the darkness to maintaining to permanently. The death' door cannot be permanently maintained, so charm must be removed( or change charm to hex).
One of my greed is to apply this system to some parts of summoning. For example, the battle of summoners will be surprisingly pleasant if 'summon small mammal' or 'summon mana viper' is perma-buffs style and the player automatically summons his colleague when he sees the enemy. But these spells do not include charm. In this case, the system is less consistent with the school.
Should the charm school' system be consistent, or should the convenience of spells be prioritized, even at the expense of some of the consistency? This is my concern.

It doesn't need to be consistent with a school of magic, IMO. There are obvious spells that a player would want to have permanently maintained if they have the MP. Most of them are Charms and Transmutations, but some are Hexes, Necromancy and Summons. Its all relative to the spell, not the school.

For example, in HellCrawl, there are many non-Charm spells that use this approach already; Spectral Weapon and Servitor are good examples. It works well as they trigger when you cast a spell or attack, depending on the spell. The Animate and Death Channel spells are just permabuffs that activate based on your spellpower when you kill something. So the system is already there somewhat, it is just a matter of extending to other summons with easy to define triggers.

For summons with predetermined limits; Canine Familiar for example, you could have that companion always present and they would upgrade based on spellpower. It would be an easy way to keep summoners from being tedious to play. Most summons are far less clear about when you want them active; they either:

1) Don't have limit on the number of summons (Ice Beast, Small Mammals)
2) Need to be placed (Lightning Spire, Forest, Gateway)
3) Have very short durations (Hydras)
4) Are extremely variable in what they summon (Imp, Demons, Menagerie)
5) Have potential consequences to casting (Demons, Horrible Things)
6) Are very situation dependent (Butterflies)

What I am getting at is that most of the "easy" spells to assign a permabuff to in DCSS are already done in Hellcrawl. I can only think of a handful of summons that you could actually do the change to at this point; Canine Familar, Guardian Golem, Mana Viper, Dragon's Call. You might be able to get Shadow Creatures and Call Imp to work as well.

Also, permabuffs are definitely good (probably too good) and I am not sure how to keep them balanced. My biggest gripe about the reserved MP system is that you save a lot of time in combat; it takes no time at all to keep the full suite of Book of Battle spells, a Deflect Missile for good measure while also being in Dragon Form or Ice Form. My last win had 4-5 buffs for most of the game and 7 buffs on for my Orb run. In vanilla, each of those casts would cost you time in combat. You would have to pre-buff like crazy, which you might not have time for in many situations and you never lose time mid battle to keep them up. Its a big reason why I think permabuffs feel strong in general on Hellcrawl. You get all that power with almost no downsides except a limited MP pool, which isn't a problem for most aspiring Skalds and Transmuters since they were not planning on using other spells anyway.

Summons would be the same thing; each one that gets a permabuff would effectively be saved turns. Summoning is already pretty busted, so giving the player a bunch of free turns would only make that worse. I'm not sure how to fix this issue without aggravating players; you can't simply slow them down, since that would make the spells suck, but you can't just give all that time away freely either without expecting a shift in balance.

[tealizard]

Good comment.

If you look at dcss circa about .17, most of the classic duration/buff spells were still around. It was always possible to have 5 or so duration spells and/or summons active in any given fight without risk of being attacked or cast on by monsters while you cast your spells. You do it by breaking line of sight with corners, abusing stairs, and so forth. Since you can't use stairs in hellcrawl, it's true there are situations where the turns you might need to spend casting duration spells without permacharms have real combat ramifications. (And of course, the hellcrawl orbrun is far more intense than any situation that occurs naturally in dcss and there turns casting duration spells would certainly make a difference sometimes.) But these situations basically don't happen in reasonable dcss play and that's the baseline for balancing hellcrawl, not hypothetical versions of hellcrawl that don't have X mechanic.

All that said, there is a freestanding question of whether the ability to have so many duration spells active is overpowered. I would say the real problem is that max player mp gets so big relative to the cost of spells. I think there was some ambition to address this, but crawl mp just sucks so bad it needs a total redesign.

[damerell]

It doesn't need to be consistent with a school of magic, IMO. There are obvious spells that a player would want to have permanently maintained if they have the MP. Most of them are Charms and Transmutations, but some are Hexes, Necromancy and Summons. Its all relative to the spell, not the school.

I agree very strongly with this. Stoat Soup has done Portal Projectile, which doesn't have any Charms in it, and conversely we are unlikely ever to do Spectral Weapon, which to me is "just another summon" (and I don't want to do permasummons) with an added dose of "actually, I don't want it against every opponent because of the way sometimes it cops a massive AoE and does nothing except share me some damage".

1) Don't have limit on the number of summons (Ice Beast, Small Mammals)

These spells have limits. As far as I know every summon does except Dragon's Call (where the duration and time between dragons provides a practical upper limit); nearly all summons do, anyway.

What I am getting at is that most of the "easy" spells to assign a permabuff to in DCSS are already done in Hellcrawl. I can only think of a handful of summons that you could actually do the change to at this point; Canine Familar, Guardian Golem, Mana Viper, Dragon's Call. You might be able to get Shadow Creatures and Call Imp to work as well.

Call Canine Familiar might be an option because its long duration means it often lasts between fights anyway, but other than that I think permasummons would have the problem of saving a bunch of time at the start of the fight in spades, and summons are very strong anyway. (Also, either you have to get rid of the XP penalty or players will be incentivised to fiddle with exactly how many summons they have switched on (whether that is actually worth worrying about or not)).

My biggest gripe about the reserved MP system is that you save a lot of time in combat

That's mine too, but I can't see any way around it if permabuffs are actually to be nicer to use than ordinary spells. If they came up one by one on sighting a monster, you'd sometimes want to arrange the set of permabuffs (or the order, if that was possible) to get the right ones up first, and now it's as fiddly as non-permabuffs. We have mid-combat miscasts which force the spell off for a period - no immediate recasting - but it's still a huge improvement.

You get all that power with almost no downsides except a limited MP pool, which isn't a problem for most aspiring Skalds and Transmuters since they were not planning on using other spells anyway.

Well, you can easily run the MP pool dry with Portal Projectile - I've got a HESk whose jewellery happens to include MP+9 and MPRegen, and they've run the pool dry in spite of that (with MP costs comparable to repeated recasting, not MP reservation; PProj is one of the spells like Regeneration where MP reservation can easily work to your advantage compared to recasting in vanilla).

Another issue with importing Hellcrawl's system is it does nothing about hunger costs, because it doesn't have to. Until vanilla admits it's actually getting rid of food, something would have to be done about that.

[sdynet]

Yeah. Certainly, perma-buffs is very powerful. I am aware of that as well. They have to cost more at a level that doesn't hurt convenience. I think...
Add cost:
1. More mp reserve(the most convenient way)
2. Reserve hp by spell level(This, on the contrary, is too extreme.)
3. Noise generation
4. 10% reduction in travel speed while perma-buffs is maintained. This effect is not superimposed.
5. It consumes mp(like an dragon's call, portal projectile)
There is also a choice to accommodate all five options. Basically, apply one and four to all perma-buffs, two or five to perma-summoning and three to charm. As our game is moving away from hunger, I have excluded anything related to it.

[sanka]

For me, it seems hard to balance "perma" buffs with drawbacks or significant upkeep costs. If my understanding is correct, the point of permabuffs compared to the current system is that you do NOT need to recast them all the time. You loose this benefit exactly when the costs are relevant enough so you do not want to upkeep them all the time: you would switch on/off them based on the situation, which will be similar to current casting it before combat.

Either there is a decision from combat to combat whether to use them, but then there is player interaction. Or there are no player interaction, tactical decisions, but then could only be strategical costs, not temporary tactical ones.

In other words, if you want permanent buffs, you need strategical, long term costs. If you want tactical decisions, then some appropriate tactical cost (drawback, etc.) could be appropriate (although long term cost could be fine here as well).

[tealizard]

@sdynet: One thing you have to watch out for is the possibility that you defeat the purpose of permabuffs by imposing a cost that incentivizes toggling. For example, with spectral weapon, you might worry about AOE damage -- the damage sharing silliness was removed from hellcrawl to prevent this consideration from coming into play, compensating by increasing the spell level to 5, instead of the ridiculously low 3 where every character should get it. Things like movement speed reductions or overall increases in noise for actions unrelated to the spell effect are nonstarters. Things that might work are increases in spell level/difficulty for permacharms, cumulative, symmetric increases in spell difficulty when maintaining multiple permacharms (e.g. each permacharm gets +nK to base difficulty where n is the number of permacharms active and K is some constant), or some other kind of additional, preferably easily mentally calculated additional cost to maintain many permacharms.

Again, I warn you that any time you spend trying to get permabuffs into dcss will be wasted. If you're interested in this project it should be as the start of an independent fork or a contribution to hellcrawl. As I said, permasummons would be a cool feature if done well. Entering negotiations about the balance of hellcrawl-style permacharms in a game that has stairdancing is ridiculous on the face of it and doubly so given that there is no chance permacharms will be accepted anyway.

@sanka: Yes, that is exactly right. At minimum, any drawback has to be tied to the use of the effect, e.g. if the effect triggers on melee attacks, it may be acceptable to cause additional melee noise, though I'm not completely sold on that even.

edit: About the supposed power of hellcrawl permacharms -- it's true that permacharms are more powerful than very lazy, naive use of duration effects, but they are generally weaker than careful use of duration effects in dcss because they take more mp over time. The whole problem is that careful use of duration effects is unnatural and destroys any flow that might exist in crawl combat. It's mandatory luring before every fight at best.

Most of the observation about the power of hellcrawl permacharms comes from the fact that they are derived from the more powerful spells from dcss (historically), in terms of return on investment (e.g. rmsl) or absolute power (e.g. death channel, darkness). And of course this would be even more true with permasummons.

[TBDA]

I agree very strongly with this. Stoat Soup has done Portal Projectile, which doesn't have any Charms in it, and conversely we are unlikely ever to do Spectral Weapon, which to me is "just another summon" (and I don't want to do permasummons) with an added dose of "actually, I don't want it against every opponent because of the way sometimes it cops a massive AoE and does nothing except share me some damage".

Agreed. You can only have permabuffs for spells with well defined triggers. At least with Spectral, you have clear Yes/No situations so you can manage the spell.

These spells have limits. As far as I know every summon does except Dragon's Call (where the duration and time between dragons provides a practical upper limit); nearly all summons do, anyway. Call Canine Familiar might be an option because its long duration means it often lasts between fights anyway, but other than that I think permasummons would have the problem of saving a bunch of time at the start of the fight in spades, and summons are very strong anyway. (Also, either you have to get rid of the XP penalty or players will be incentivised to fiddle with exactly how many summons they have switched on (whether that is actually worth worrying about or not)).

What I meant by the cap is that there is a global summoning cap. Some spells only give you one ally regardless of the number of times it is cast. Others allow you cast multiple times until you hit the cap. Others ignore the cap entirely. The code wouldn't be able to tell which spell you are trying to hit the cap with first. With permabuffs, you would either need to limit each spell to one instance (which I think would work), or recast the spell with a higher associated MP cost (which would be wacky to code).

Regarding the time savings of permabuffs. That's mine too, but I can't see any way around it if permabuffs are actually to be nicer to use than ordinary spells. If they came up one by one on sighting a monster, you'd sometimes want to arrange the set of permabuffs (or the order, if that was possible) to get the right ones up first, and now it's as fiddly as non-permabuffs. We have mid-combat miscasts which force the spell off for a period - no immediate recasting - but it's still a huge improvement.

I would love to see miscast chances integrated to Hellcrawl.

The biggest offender with the reserved system is that you can ignore the spell failure of a permabuff spell. As long as you can hit the MP cost, you get the spell active. Thus, you can can get just enough EXP to cast the spell and enjoy its benefits. Playing an Unarmed Octopode? Just get that Dragon Form down to 20% and you'll be off to the races. Have a couple ice enhancer rings and a decent Charm skill? Ozo's is viable despite being at 20%. Those kinds of numbers would be risky in vanilla, especially on something squishy like a Octopode. It would reign in a lot of the shenanigans associated with permabuffs by adding risk back into the equation. You would actually be forced spend enough EXP to get the spell failure down low enough for the risk/reward to be reasonable, rather than strictly looking at the reserved MP cost.

Well, you can easily run the MP pool dry with Portal Projectile - I've got a HESk whose jewellery happens to include MP+9 and MPRegen, and they've run the pool dry in spite of that (with MP costs comparable to repeated recasting, not MP reservation; PProj is one of the spells like Regeneration where MP reservation can easily work to your advantage compared to recasting in vanilla).

I actually like that some spells have both a reserved and active cost. I am a big fan of using MP as the limiting resource for casting. In fact, I think that the formula's for Reserved MP cost are too forgiving; I shouldn't be able to have a buff up "at cost" until I have that spell failure down to 1-2%. I don't even think that "at cost" is high enough, especially for spells like Regeneration. Maintaining the spell should probably be 150% of the vanilla cost, even when you are very good at casting the spell. Those changes would limit the number of buffs you have active and actually force the player to make meaningful decisions on both the number of buffs active, as well as the EXP you are willing to spend to get the costs down.

Right now, a blaster can rock an Ozo's, Repel Missile and a Regeneration with pretty low EXP investments and it barely effects their ability to blast (9 MP reserved). Higher reserved costs would force you to look at the opportunity costs; is the Ozo's worth it in exchange for an extra Fireball or two before hitting 0 MP? Is upgrading to Deflect Missile worth it if I can't dedicate a significant amount of EXP to bring the MP cost down? You don't really have to make those calls right now. Bumping up the Reserved MP cost would make running buffs a meaningful decision for players that need that MP to function.

Another issue with importing Hellcrawl's system is it does nothing about hunger costs, because it doesn't have to. Until vanilla admits it's actually getting rid of food, something would have to be done about that.

On the subject of food;

1) I don't think it is a good limiting resource. It pretty much only affects casters, Rage users and a few species. All of those situations are better addressed with other mechanics. There is a reason why weapon-dudes are so popular in vanilla; people don't want to deal with food.

2) Casters already have a limiting resource in MP, so balance should be based around that. If having no food makes casters super strong, start bumping up MP costs. We already know that using a formula that changes MP costs based on your skills is doable (see the Reserved MP mechanics). Don't jerry-rig a weird hunger system to bog them down when you have a perfectly good system already there. Two half baked systems is not better than one well implemented system.

3) Rage is great in vanilla, but forcing Hell onto all players and amping up the difficulty of Zot more or less fixed the issues around that ability in Hellcrawl. You can't just Rage your way through the easy runes and walk out without addressing Rage's glaring flaws in those environments. You get great early game answers to problems, but they fizzle as the game progresses. If Rage is still a problem, look at the ability itself. If it is too strong for its drawbacks, either amp up the drawbacks or scale back its power. Don't slap a hunger mechanic on it and call it a day. Actually look at the ability itself.

4) Species; food mechanics add a lot of flavour (haha) but it doesn't have to be there. Trolls can be managed by slowly giving them their good mutations over time, rather than all at once. Spriggans are only busted because of the way vanilla does stairs; you could give them Swiftness as an ability rather than a full time mechanic to address that. Undead can be tempered with aptitude/HP changes or forcing more holy enemies and Dispel Undead users on them. My point is that the strong species that are balanced by their food mechanics can be addressed with other mechanics; it doesn't have to be food.

Things that might work are increases in spell level/difficulty for permacharms, cumulative, symmetric increases in spell difficulty when maintaining multiple permacharms (e.g. each permacharm gets +nK to base difficulty where n is the number of permacharms active and K is some constant), or some other kind of additional, preferably easily mentally calculated additional cost to maintain many permacharms.

I really like this idea. Stacking buffs is the real problem. If there was a cost to doing so, the player would have to make the choice on which buffs are worth running.

[tealizard]

The biggest offender with the reserved system is that you can ignore the spell failure of a permabuff spell. As long as you can hit the MP cost, you get the spell active. Thus, you can can get just enough EXP to cast the spell and enjoy its benefits. Playing an Unarmed Octopode? Just get that Dragon Form down to 20% and you'll be off to the races. Have a couple ice enhancer rings and a decent Charm skill? Ozo's is viable despite being at 20%. Those kinds of numbers would be risky in vanilla, especially on something squishy like a Octopode. It would reign in a lot of the shenanigans associated with permabuffs by adding risk back into the equation. You would actually be forced spend enough EXP to get the spell failure down low enough for the risk/reward to be reasonable, rather than strictly looking at the reserved MP cost.

Miscast effects for duration spells have absolutely no impact on normal play. You generally only need one successful cast per encounter and casting happens prior to direct engagement, so even theoretically dangerous effects like sleep or whatever are unlikely to have any impact. I've had people claim that miscast effects make a difference so I actually tried recasting hellcrawl transmutations repeatedly to try to get significant miscasts at around 20-25%. After about 30 recasts, the worst I saw was like 3 damage. Duration spells in dcss are well known for their insensitivity to fail rates. If anything the hellcrawl system makes fail rate more important, since high fail rate often means you can't cast the spell at all, whereas in dcss you simply reset a fight that begins with many failed duration casts.

Also, you are exaggerating the issue of stacking permacharms. It is basically false that stacking permacharms makes the player stronger in hellcrawl versus the dcss player at a comparable stage of the game.

[damerell]

What I meant by the cap is that there is a global summoning cap. Some spells only give you one ally regardless of the number of times it is cast. Others allow you cast multiple times until you hit the cap. Others ignore the cap entirely. The code wouldn't be able to tell which spell you are trying to hit the cap with first.

Unless I am very much mistaken, no, there isn't. There's a per-cap spell for most summoning spells. There's no global summon cap. I've just started a wizard-mode game, set Summonings and Spellcasting to 27, memorised Summoning spells until I ran out of spell slots, and cast them all; the butterflies from my first cast are still around.

Just get that Dragon Form down to 20% and you'll be off to the races.

I can't see us doing permaforms either; you can only have one form at once and you know when it expires.

Well, you can easily run the MP pool dry with Portal Projectile - I've got a HESk whose jewellery happens to include MP+9 and MPRegen, and they've run the pool dry in spite of that (with MP costs comparable to repeated recasting, not MP reservation; PProj is one of the spells like Regeneration where MP reservation can easily work to your advantage compared to recasting in vanilla).

I actually like that some spells have both a reserved and active cost. I am a big fan of using MP as the limiting resource for casting. In fact, I think that the formula's for Reserved MP cost are too forgiving; I shouldn't be able to have a buff up "at cost" until I have that spell failure down to 1-2%. I don't even think that "at cost" is high enough, especially for spells like Regeneration. Maintaining the spell should probably be 150% of the vanilla cost, even when you are very good at casting the spell. Those changes would limit the number of buffs you have active and actually force the player to make meaningful decisions on both the number of buffs active, as well as the EXP you are willing to spend to get the costs down.

I wouldn't go that far, but I'm definitely softening towards MP reservation as a mechanic to stop you hitting fights with full MP, even given that right now in Stoat Soup you have greatly lowered MP regeneration once you start getting permabuff benefits so that in a prolonged combat you see no more MP than you would in vanilla.

[sdynet]

I modified the mp reserve formula.
1 - spell fail rate = success.
Existing: spell level / success * success
New: spell level / success^10 (I used this formula because I couldn't think of a better formula.)
The effect is certain. When the probability of failure is 5%, the new method requires approximately 150% of the cost compared to the traditional method. And, When the probability of failure is 10%, the new method requires approximately 275% of the cost compared to the traditional method. For example, regen requires 6 mp in the old method when the probability of failure is 30%, but the new method requires 106 mp. Users should reduce the probability of failure to below 10% in order to use it in reality. Look at the table below.

MP reserve cost.png (35.41 KiB) Viewed 17846 times

In addition to this, a measure will be added to raise the spell level of perma-buffs. Especially regen. I'll raise it to one level. That alone will require the user to make a lot of effort to use the regen. To prevent the toggling of the buff, spectral weapon does not share the damage and the spell level will rise 5. And though it's still a long plan, summoner intends to make the MP consume one every time he summoning(perma-spell) a followers(like dragon's call).
I think that this has increased the cost users have to bear to a level that is tactical and strategic. (maybe)

edit: The download link has been updated. Subsequent patches have been applied.

patch note
* modified the mp reserve formula
* song of slaying is now a level 3 spell.
* regen is now a level 4 spell.
* spectral weapon is now a level 5 spell. The damage sharing function of this spell has been remove.

[petercordia]

I like this idea. It feels less cheesy than Hellcrawl's implementation. I hope some of the other branches will copy this.

[Hellmonk]

Keep in mind that your formula makes permabuff spells uncastable at low levels for most combos. This is actually fine imo, but you need to remove or rework the skald background and maybe a few others. Idk if this actually ends up working better than a player mp reduction, probably depends on whether or not stairs exist and how comparatively good conjurations end up being in the context of your fork's other mechanics (since mp is supposed to function as an in-combat limit for those spells).

[tealizard]

I'm kind of scratching my head here. I thought the issue was having many buffs active, which this new formula does nothing to address. As hellmonk says, this makes these spells uncastable for a sizable chunk of the early game. I mean, of course the formula isn't a big deal -- if you come up with something good on summons that will be useful regardless of what specific formulas you use in your own fork.

Another old suggestion I liked for addressing the multiple buff thing was setting a minimum fail rate that increases with the number of active buffs. At something like 5%*(n - 1) with n being the number of active permacharms, you would find it's difficult to keep more than 4 or 5 active, which I think is a pretty conservative approximation to dcss behavior. It's also relatively easy to understand. This would be with the original success^2 formula and it would apply to all active permacharms, of course. That might encourage charm swapping, which would be bad, but anything you try to do to limit the number of active permacharms will probably run into that.

[petercordia]

The issue as I perceive it is simply that Hellcrawl perma-charms cost too little mp. In Hellcrawl you can have dragonform up with a 70% failure rate; in normal crawl I'd only do that to prepare for a battle with a ghost. It's fine to have a lot of buffs, so long as it comes with a large enough cost.
As for the idea you wouldn't be able to cast charms in the early game, I don't think that'd be the case. You can get the low-level charms down to 5% fail rate fairly easily. You wouldn't get spectral weapon nearly as quickly because it's now a level 5 spell, which is a bit weird though. You'd probably get spectral weapon much later than an earth elementalist got LRD.

[tealizard]

I doubt you have ever cast dragon form at 70% fail in hellcrawl. By my calculations, that would cost about 77 mp.

[damerell]

I like this idea. It feels less cheesy than Hellcrawl's implementation. I hope some of the other branches will copy this.

It seems draconian to me - for many permabuffs it is simply saying that you can't use them at moderate spell failure. In vanilla, at 20% failure, you could chance your arm on Song of Slaying - here it wants to reserve 27 MP for the privilege! Surely the point of a permabuff system must be that it is _less_ obnoxious for the player than vanilla recasting.

I fear we're going to have to move to a MP reservation system but since we make spell failure chance important in other respects, I can't see it going beyond reserving the flat cost of the spell.

[svendre]

I don't like the idea of non-persistent buffs because of the quality of game play issues. I like deflect missiles for example, because you don't need to recast it over and over and you could play tediously if it did expire. I do remember in the past where the sequence was like: cast, cast, cast, channel, channel, channel, <now you can proceed into the unknown>, loop. That sucked.

I don't like the idea of a big pile of spells all being like deflect missiles because it's too overpowered. The mana pool will just appear to many as an extended equipment slot.

My idea for a solution is to implement persistent spells like deflect missile, but *each time* they have an interaction in the game, charge the mana cost. So let's say you get shot with an iron shot, and deflect missiles blocks it... charge 6 mana, but do not turn deflect missiles off. Another example, let's say you have mutated claws from the level 1 spell: each time you hit an enemy, charge 1 mana. If you enable 10 spells permanently on, your mana could take massive hits depending on the situation.

If your mana runs out, I see two scenarios:
1) the spell effect won't trigger, nothing else happens, or
2) the spell gets disabled and needs to be cast to turn it back on again

I haven't given it a lot of thought, but I think option #1 would be a better choice. Also, there needs to be a mechanism for turning those kinds of spells off.

This system would solve both problems of tedium of recasting spells over and over, and give them a real cost to keep the game balanced and interesting.

[damerell]

My idea for a solution is to implement persistent spells like deflect missile, but *each time* they have an interaction in the game, charge the mana cost. So let's say you get shot with an iron shot, and deflect missiles blocks it... charge 6 mana, but do not turn deflect missiles off. Another example, let's say you have mutated claws from the level 1 spell: each time you hit an enemy, charge 1 mana. If you enable 10 spells permanently on, your mana could take massive hits depending on the situation.

Except for Infusion and PProj, this is a huge increase in MP costs compared to vanilla. For example, it is impractical for a blaster caster to use DMsl when any centaur pack can suddenly suck all your MP away. I think it's important that any permabuff implementation doesn't leave the player wishing they could go back to vanilla Charms.

[sdynet]

OK, I will do it step by step.

Keep in mind that your formula makes permabuff spells uncastable at low levels for most combos.

At the same time as solving this problem, I still didn't want to allow the use of perma-buff spells with a failure probability of more than 30%. So I modified the formula one more time.
1 - spell fail rate = success.
Existing: spell level / success * success
New2: spell level / success^5

MP reserve cost2.png (33.19 KiB) Viewed 17691 times

And I played Gh and Mi. To see if they can use the spell in the early game. If the magic and the least intimate two can proceed with skald normally, then I can say I've solved this problem. The result was not as bad as the following.

Gh(1).png (132.15 KiB) Viewed 17691 times

Gh(2).png (179.29 KiB) Viewed 17691 times

Mi(1).png (128.97 KiB) Viewed 17691 times

Mi(2).png (214.94 KiB) Viewed 17691 times

Mi(3).png (114.16 KiB) Viewed 17691 times

Another old suggestion I liked for addressing the multiple buff thing was setting a minimum fail rate that increases with the number of active buffs. At something like 5%*(n - 1) with n being the number of active permacharms, you would find it's difficult to keep more than 4 or 5 active, which I think is a pretty conservative approximation to dcss behavior. It's also relatively easy to understand. This would be with the original success^2 formula and it would apply to all active permacharms, of course. That might encourage charm swapping, which would be bad, but anything you try to do to limit the number of active permacharms will probably run into that.

This was an attractive opinion for me. But I'm worried that imposing two types of penalty on this system will complicate the game.

[tealizard]

If you want to address this thing about number of active permacharms, you need new mechanics. That means more complex rules. If your analysis is that the issue is not important enough to justify added complexity, I completely agree with that. Otherwise, you need to bite the bullet and add new mechanics. Fiddling with the exponent cannot accomplish your stated goals and will only move away from dcss behavior. There is a reason that hellcrawl uses the exponent 2 -- it is the integer exponent that best approximates dcss duration spell mechanics without making duration spells strictly easier to cast.

[petercordia]

The stated aim is to keep the power level of charm spells OK, whilst reducing keypresses and tedium. In the current state of charms I can never be bothered to use them optimally, whereas with permacharms I would use them almost optimally. Hence to keep the effective power level the same (for me), the new permacharms must be less powerful than the old temporary charms.
From tealizard's words, he might be using charms more optimally than myself. I guess that changes your perspective.
Maybe Hellcrawl charms keep the powerlevel the same and this is part of the reason why Hellcrawl charms feel overpowered. Otherwise Hellcrawl permacharms are actually more potent than normal charms. (I certainly think regeneration is more powerful in Hellcrawl because of the DOOM counter.) Either way, permacharms need to be nerfed compared with Hellcrawl, and fiddling with the exponent seems like a very satisfactory method for implementing such a nerf. If Infusion affected Spectral Weapon, it would only make it much more complicated to calculate whether it is better to cast Infusion, but I don't think it'd make the game better. One of the things I like about Hellcrawl-style permacharms is that it is obvious for a skald whether casting Infusion is good - If it doesn't stop you casting a higher-level charm it is always good.
There is no reason to stick closely to dcss behaviour except to keep the game balanced. (Assuming DCSS is currently balanced.) DCSS behaviour is, to use damerell's words, obnoxious.

[damerell]

The stated aim is to keep the power level of charm spells OK, whilst reducing keypresses and tedium. In the current state of charms I can never be bothered to use them optimally, whereas with permacharms I would use them almost optimally. Hence to keep the effective power level the same (for me), the new permacharms must be less powerful than the old temporary charms.

I think my aim is different; to make the power level the same for everyone as it is now for Hypothetically Optimal Man who can be bothered to optimise their use. Once that's done, we can look at whether they are actually overpowered in that scenario (although I hope they won't be, since I expect the vanilla developers to always more-or-less have HOM in mind), but the way to do that seems to me to be to get the implementation out and then play. (Indeed, I just finished a 5-rune HESk, unusual because I developed our implementation in parallel with playing it...)

One thing that's been bothering me about MP reservation schemes is the "Troll Skald problem"; that a level 1 TrSk can do nothing with Infusion at all. However, on reflection a vanilla troll skald can't either - the odds of them getting enough duration out of it to regenerate their 1MP before it expires are very small. (I'm not saying Hellmonk isn't in general right that some of these proposals make it impossible for a low-level Skald to use their spells and that that's bad, it's just I've been fretting over this particular case and I shouldn't have been.)

[tealizard]

@petercordia: Obviously it is impossible to balance a game around the subjective opinions and playstyle of players -- this will lead incoherent standards. That's the point of optimal play analysis, to interrogate what the best thing the player can do is and what the consequences of these possibilities are for the game.

So about your "feel" that hellcrawl permacharms are overpowered, I suspect that most of the difference in this perception is indeed the difference between what you do in dcss and what is actually possible in dcss. This isn't a comparison between you and me though. It's a comparison between what you actually do and what you could do. I think it's absolutely true that the QoL benefit of permacharms also comes with a leg-up for newer players in that that it quickly takes them to spell usage consistent with a knowledgeable but also extremely careful, tedium-tolerant player. The realization that you can and should cast these spells a lot more than might come naturally and in counterintuitive situations is a part of the game we should be willing to lose.

[svendre]

My idea for a solution is to implement persistent spells like deflect missile, but *each time* they have an interaction in the game, charge the mana cost. So let's say you get shot with an iron shot, and deflect missiles blocks it... charge 6 mana, but do not turn deflect missiles off. Another example, let's say you have mutated claws from the level 1 spell: each time you hit an enemy, charge 1 mana. If you enable 10 spells permanently on, your mana could take massive hits depending on the situation.

Except for Infusion and PProj, this is a huge increase in MP costs compared to vanilla. For example, it is impractical for a blaster caster to use DMsl when any centaur pack can suddenly suck all your MP away. I think it's important that any permabuff implementation doesn't leave the player wishing they could go back to vanilla Charms.

Perhaps the numbers would need fiddled with such that less mana is consumed per activation, but I think the two main objectives can still be achieved.

[bel]

OK, I will do it step by step.

Keep in mind that your formula makes permabuff spells uncastable at low levels for most combos.

At the same time as solving this problem, I still didn't want to allow the use of perma-buff spells with a failure probability of more than 30%. So I modified the formula one more time.
1 - spell fail rate = success.
Existing: spell level / success * success
New2: spell level / success^5

And I played Gh and Mi. To see if they can use the spell in the early game. If the magic and the least intimate two can proceed with skald normally, then I can say I've solved this problem. The result was not as bad as the following.

You can try something like:

MP reserved = spell level / (success)^(spell level).

This formula could be tweaked a bit, of course.

A level 9 spell at 30% fail would cost much more MP (relatively) than a lvl 2 spell at 30% fail. This method will also solve (or mitigate) the "Troll Skald" problem, because it's ok for low level spells to have high miscast %, and still be fine to cast (which is current DCSS behaviour).

(I cast spider form at 30% fail all the time in my Tm games, btw.)

More generally, I am not really convinced that Hellcrawl permacharms are somehow more powerful than DCSS charms. However, if you want to discourage casting high level permacharms at 30% fail, it is a legitimate objective.

[petercordia]

I'd suggest
MP reserved= spell level / (success) ^((1+spell level)/2)
That way statue form costs 16 MP at 75% success rate, and regeneration is still easy to cast.

[damerell]

More generally, I am not really convinced that Hellcrawl permacharms are somehow more powerful than DCSS charms.

I think it's hard to work around the fact that you save a turn (or five) at the start of every fight, and this is a large benefit. It's not clear what can be done about that - the game can hardly actually take the turns away from you [1], and making the spells significantly less good to compensate is a major penalty for, say, a Centaur Skald who can just back up and get all their buffs up in vanilla.

[1] Buffs come up one turn at a time when you see a monster? When you wait with a monster in view? Now you want a way to sort the order they come up in and to carefully fiddle with the ordering and the exact set you're using... and the idea was to eliminate that kind of thing.

[bel]

You're assuming that buffs are cast at the start of the fight. But that's not the only option (or even the most common option). Buffs could be cast before the fight starts, in anticipation of a fight (say, just before you open a door, or before you go downstairs.)

Also, keep in mind that permabuffs always reserve MP, while you may not actually cast a particular buff in a fight. So the power adjustment doesn't just go one way.

[petercordia]

More generally, I am not really convinced that Hellcrawl permacharms are somehow more powerful than DCSS charms. However, if you want to discourage casting high level permacharms at 30% fail, it is a legitimate objective.

(emphasis mine)

I agree with this.
We shouldn't focus on keeping charms the same - we should focus on good gameplay.

[tealizard]

Still scratching my head about this thread. This notion that there's a problem with casting a spell at 30% fail is very original. 30% fail is double basic mp cost in hellcrawl mechanics and usually quite safe to use for dcss duration spells. People wildly overrate the danger of miscasts, especially when the player is not directly engaged in combat. And don't even get me started on this "turn cost" nonsense.

If you want to make a rule that is good, you need to focus on simplicity. You guys are trading recipes for mechanics no one understands.

[damerell]

Also, keep in mind that permabuffs always reserve MP, while you may not actually cast a particular buff in a fight. So the power adjustment doesn't just go one way.

I'm not sure how true that is of the low-level ones, and of course to win the game you must first survive the low levels. I'd always want Infusion if I planned to hit any monsters and don't have other uses for the MP, Shroud of Golubria if anything is going to try and hit me, Regeneration if I might take any damage - and Stoat Soup's non-terrible Song of Slaying if there's any significant number of monsters to fight.

ETA: It's also less true in general, because if I think I'm unlikely to want it in many fights I can keep it off, and cast it as needed - I have the option of having a vanilla-like behaviour. So they don't really "always reserve MP".

I see what you're getting at, but my experience with our implementation - which is not, now, tremendously different to Hellcrawl's - is that I'd always rather the spells were permabuffs, and not just because of the convenience factor; if they aren't diminished in some other way, they're stronger overall.

[bel]

Also, keep in mind that permabuffs always reserve MP, while you may not actually cast a particular buff in a fight. So the power adjustment doesn't just go one way.

[...]
ETA: It's also less true in general, because if I think I'm unlikely to want it in many fights I can keep it off, and cast it as needed - I have the option of having a vanilla-like behaviour. So they don't really "always reserve MP".
[...]

Specifically on this point:

If you cancel a permacharm mid-fight, you don't immediately get the MP back. Only your max MP increases, not your current MP. If you're talking about cancelling pre-fight, then we're back to the case of not casting buffs in-combat.

As for the rest of the post, I do not see what the level of the permabuff has to do with anything. The issue is whether you get to have it on before a fight starts. I can just as well cast Spectral Weapon as I can Death Channel before opening a door.

[damerell]

If you cancel a permacharm mid-fight, you don't immediately get the MP back. Only your max MP increases, not your current MP. If you're talking about cancelling pre-fight, then we're back to the case of not casting buffs in-combat.

I meant neither of those things (and I know, of course, that the MP are not returned); I meant what I actually wrote - I can keep one normally off, and cast it only when I see a specific need for it, getting something very like the vanilla behaviour. Hence, it's not really the case that permabuffs "always reserve MP" - you're quite at liberty to have them not do so and if you do you're no worse off than you would be in vanilla.

As for the rest of the post, I do not see what the level of the permabuff has to do with anything.

Because, as I said, it so happens that the low-level ones are ones you might well want on all of the time, so you don't care that they always reserve MP because you are almost always getting a benefit from them; in vanilla, you'd be always casting them if you had the opportunity.

[bel]

deleted