Curses

[dpeg]

Crawl's curses contested, see for example this recent posting.
evilmike and I have discussed the matter a bit, and some ideas can be found on this development wiki page. In a nutshell, we see four options:

1. No action. The situation is not ideal, but change is certainly not urgent.
2. Get rid of curses.
3. Curses wouldn't make items sticky anymore; rather, there would be bad effects for unwielding when uncursed. (evilmike)
4. Cursed would stay sticky, but have additional bad effects. (dpeg)

The last two proposals would strive to make use of scrolls of Remove Curse more interesting, without making it unfun.

The wiki page explains the suggestions in depth. Comments are welcome.

[KoboldLord]

Honestly, this sounds like it's at high risk of becoming another food reform. A feature with a mild effect on the game might not be improved by making it have a major effect instead.

Anyway, I like option 3, but with a major caveat. Unequip-activated curses should *always* have a fairly significant positive effect when used, and only cause trouble when switched. Distortion and vampiric brands are excellent examples of what such a cursed item might look like. For instance, a cursed ring of protection would be a normal +N to AC for as long as you left it equipped, but when you took it off you'd get a -N malus to AC that would time out one point at a time as you gained xp. A plain old -N ring of protection is simply rubbish whether it's sticky or not, but a cursed +N ring of protection that you don't want to swap out could justifiably be generated at an even higher positive value than normal rings of protection, since you have to compensate for the drawback.

Items that are a single penalty and absolutely nothing else probably aren't possible to make interesting. Might as well cull them.

[evilmike]

Honestly, this sounds like it's at high risk of becoming another food reform. A feature with a mild effect on the game might not be improved by making it have a major effect instead.

Yes, I agree. The article on the wiki already brings up the food reform for exactly this reason, and it's stated several times that we need to be mindful of pitfalls like this. This is why I went as far as to include "do nothing" on the list of possibilities.

I believe curses are a flawed system in this game, but I also think this is somewhat dangerous territory. Unlike a lot of things, I think if we make changes here, we run the risk of things going badly wrong. There are many problematic things in Crawl I see clear solutions for. Curses are not one of those (even the basic "remove it" option is more complicated than it looks, because no one knows what to do with Ashenzari in that case).

This is also why I feel that a less complex system is better than a more complex one.

[eeviac]

#4 feels like a different flavor of negative mutation. And, as I hate getting mutated more than just about anything in crawl (even if it's a trivial -1 str), I don't want it.

#3 might be good, but I can only see it working if they cause semi-permanent effects. Glow and stat loss just encourage 5'ing.

Honestly, I don't mind the status quo.

[dpeg]

As Crawl and its devteam grow bigger, I notice a tendency to become more and more cautious. Perhaps this is natural, but one lesson (I thought) we learned from Nethack is that expansion (of game and devteam) should not lead to timid and conservative development. If you say "well, here is an idea, but it might easily be better to do nothing", then you influence players and follow developers.

Yes, the food change backfired but each failure is compensated by many more success stories. Recall galehar's skill system, removal of branches/levels/species/etc., addition of Deep Dwarves; each of these met resistance, not just by players, but also among developers. That is absolutely okay, and in fact to be expected. My point is that the probable existence of negative opinions, and the real chance of failure, should *not* lead to more cautious design and development.

I'll speak about something I am more familiar: almost every god change I announced met opposition, sometimes lots of it. Examples are the existence of an Abyss god, the fact that not all Temple gods find a seat in the Temple, giving pacification to Elyvilon, Ashenzari's curse mini-game and the idea of a fruit god. It is plain to see that I had some awful ideas too, as was realised soon enough (the Zin debacle, Beogh's still not polished permanent followers, and on Cheibriados I am not sure). However, I am confident that Crawl's religion is in a much better state than pre-DCSS: in flavour, gameplay, interface and diversity (among gods, but also compared to other games), so altogether in fun. This would not have been possible with a fearful approach to design!

Going back to curses: Doing nothing is an option, but here we're talking about improvement. That status quo is playable means we can always come back to it.
Removal of curses hurts Ashenzari -- some might be not concerned or happy about this; I think it would be sad. Both evilmike's and my proposal try to turn Remove Curse scrolls into a more valuable commodity. As explained by evilmike on the wiki page, this has to be accompanied by modifications that actually make curses *interesting*, and 3 and 4 try to achieve that in different ways.

I believe that the game gets more interesting if it offers choices. This has worked well in a number of cases I was involved with (Lugonu as a new option to leave the Abyss; timed portal vaults as an incentive to explore quickly; bazaars to perhaps spend gold on things you otherwise wouldn't buy, there are more). In hindsight, the food approach may have failed because eating is an extremely frequent activity, hence not suited for choices of this type. Knowing that, it's easy to avoid that trap with curses -- proposals 3 and 4 assume that curses are rarer than now (and curse chance is a know to twiddle).

[crate]

If you want to make curses matter I would suggest you also need to rework mummy death curses. It is already possible, but tedious, to avoid these altogether (even in Tomb, usually) but not worth the effort because right now curses don't mean much unless they curse your distortion weapon. I do not want to have to carry around a lantern of shadows to kill mummies that I run into.

Some ideas that I had:
1) Remove death curses from ordinary mummies entirely
2) Make ordinary mummies apply their curse immediately upon entering LOS
3) Make mummies appear at a much earlier depth than they currently do and increase their speed to 10, so they are actually dangerous (this is still going to be annoying though, because even after you can kill the mummy you do not want to do so).

Guardian mummies kind of have these problems also.

I have no real comments on reforming curses. They mostly do not exist right now outside of ashenzari and I do not really see a pressing reason to change this, though if ashenzari did not exist I would say that removing curses entirely is better than the current situation.

[pubby]

Problems I see with curses:
- Early game it can mean unfair death to get cursed gear
- Middle to late game curses don't matter as you have plenty of remove curse
- Extended your whole inventory can get cursed which is very annoying to uncurse and burns all your scrolls. Kind of a newbie trap as the second time you'll drop almost everything that can be cursed before doing tomb/zigs.

These are the changes I would make:
- Prevent non-ego floor trash from being cursed
- Change remove curse to work on the whole inventory
- The main method of removing curses should be wearing them until the curse expires. It is challenging and interesting to wear suboptimal equipment in short spurts as it mixes up the gameplay. Remove curse would have to be rare for this to work.

[eeviac]

I should mention that Ash is one of my favorite gods and, whatever happens here, leave my bro alone please.

[varsovie]

What I find broken in crawl, aside from curse and other minor things, is how some items are plainly bad (ring of hunger, potion of poison) with no (or really stretchy) good way to use them. And the fact that after you have identified one of those (let say ring of hunger) you have identified every ring of hunger, and a big part of the items get (auto)ignored past early/mid game.
A good item for me is ring of STATS (str,dex or int). They can gives a advantage to you, but when you see a new one you don't know if it's better than the one you currently have, do you take the risque to burn an ID scroll, or wield ID?

My suggestion is that the curses should be RELEVANT to the item cursed. Remove ring of hunger, but make ring of sustenance give hunger when cursed. Make a ring of +rF gives when cursed -rF or +rC or even both. Yes we will virtually remove a lot of intrinsically bad items, but since they are already always cursed, that don't change there effects, but just gives more options. Eg ; if scroll of remove curse become way rarer or only single targeted, but we add a scroll of destroy curse item, it can add an interesting mechanic and choices. Also the player that get a new cursed junk will not only get frustrated and only throw it away asap, but may actually try to get a reward from it.. The drawback is that this system is more complicated than the actual one, and may gives a feel of nethack (eg: the way their cursed scrolls and pots work).

In my opinion, the removal of a curse should be an active, or at least involving thing. Even if there is an XP "timer", I would like to choose what item I uncurse first, how much XP I want to invest in uncursing (can be added in the skill menu). The reason is simple, if curse are given randomly as for now, if they are removed randomly it's not a choice anymore, but only a random nuisance, like traps.

As for Ash, fix the whole game first, then the little pieces that get broken in the process. Let's keep the biggest patch the simplest possible.

[snow]

I think the first step to fixing curses is to make remove curse only work on one item. No matter how I see it having them remove all curses is the major reason why the curse system is broken at the moment. Anyway with Ash... you can call sticky curses something else and give Ash abilities to bind and unbind items.

[galehar]

If we remove curses, here is how we could change Ash: Wear an item for long enough (xp gain no time), and Ash will curse it. You get a piety costing invocation to uncurse items. Ash is lacking in piety costing abilities anyway.

[Eji1700]

I've thought about this for awhile. My various thoughts-

1. Redo cursed rings. Theres VERY few bad cursed rings, and the most important one is hunger or maybe TP. Other than that anything you put on that is cursed early isn't really enough of a threat to get off right this second, and they certainly aren't common enough to really screw you over. My suggestion would be to first make invokeable rings not auto id(lev, invis, etc) until invoked, and to then add a few negative invokable rings. Some cursed versions of said rings could even auto invoke like TP. I'm thinking noise, immolation, possibly rot, mutation, pain, hell maybe even a super rare torment? This makes putting on rings without remove curse scrolls actually a risky process as opposed to now where it's basically a non risk. I'd actually like to see better invokable rings in general to help smooth out artificer progression as well as making the ring/id/curse game better ,but that's another topic.

2. Fewer remove curses. I think it's ok that you can die from a cursed item. Either through starvation or having an item where you don't want it. It SHOULD be an actual factor when you're dealing with floor loot that trying on every glowing/randart item you find could wind up really hampering you for some time. Some of my most fun games have been where i grabbed a cursed item very early and didn't find a remove curse/enchant item until something like 8 floors later. Having to deal with a suboptimal approach can actually force you to use food chunks or play extra cautious. By mid game i've usually got double digit remove curse scrolls and there's NOTHING to do with them. Further most cursed objects are so non threatening you can leave scrolls at the stash so they don't even get hit by things like sticky flame.

3. If you don't want fewer remove curses then I'd suggest more early game enemies who can curse your stuff. Mummies are the only ones I can think of at this point, but what about a "curse item" spell for quasits or the like?

4. More dangerous cursed objects. Maybe after a certain floor depth to prevent early game issues, but I really think there should be more dangerous cursed stuff in general. Again the only things i find myself worrying about are hunger and tele, and even then tele is usually more annoying than threatening(and honestly kinda fun since it can actually force consumable use). If you could find more cursed stuff like the singing sword you'd probably care more about just trying to auto ID every sword you see by using it.

[eeviac]

If we remove curses, here is how we could change Ash: Wear an item for long enough (xp gain no time), and Ash will curse it. You get a piety costing invocation to uncurse items. Ash is lacking in piety costing abilities anyway.

sounds good to me.

[Eji1700]

Ironically I just found the scythe of curses. I love this thing and don't know why the "can curse your other stuff" mechanic isn't found on more item.s

[dd]

Personally I'd like a situation where there'd be different types of curses, kind of like how there are different brands, and each item could have one curse, like each item can have one brand/ego. One of the curse types could be sticky curse, then there could be others that would give bad effects, but would not be auto-id:d because that would just make them pointless.

[snow]

Personally I'd like a situation where there'd be different types of curses, kind of like how there are different brands, and each item could have one curse, like each item can have one brand/ego. One of the curse types could be sticky curse, then there could be others that would give bad effects, but would not be auto-id:d because that would just make them pointless.

I think curses have a LOT of potential and having them like brands might be interesting. However most of the "bad" brands that appear on artifacts already (glow, -cast, etc.) pretty much make those artifacts unusable so I'm not sure if this is the route to go. Would players honestly use cursed items or would they just become floor trash?

If these "branded" curses are removable, and remove curse is exceptionally rare, I could imagine it being fun and interesting to chose which item to remove curses from. For example say you find a demon trident and some gold dragon armor and both have curses on them (one curse summons random monsters like hell effects, and the other is doubly cursed to randomly teleport you around and stops regeneration) and you have 2 remove curse scrolls. Use one on the first or both on the second?

I mean there's a long way to go (the current sticky curse system is a joke) but I do see lots of potential.

[dpeg]

I think curses have a LOT of potential and having them like brands might be interesting. ... I mean there's a long way to go (the current sticky curse system is a joke) but I do see lots of potential.

Do you realise that two developers just tried to fill some of that potential? Keeping to throw around general comments will never produce a change. At some point you have to get into numbers and lists.

[snow]

I was voicing my opinion that curses should be improved instead of removed and brand-like curses are probably the way to go if they're removable in some manner. As for remove curse itself it could be scrapped altogether (depending on how curses are implemented). After all you can remove curses with enchant weapon and armor scrolls.

some curse ideas:

teleportitis, -teleport, -cteleport, -cast, glow, hunger, random things haunting you, slow movement speed, inaccuracy, -regeneration

Or... they could just randomly give you a mummy curse (like torment, stat drain, etc.) every so many turns.

If you go the mummy curse route, the curse can stick to you instead of the item and remove curse would make you stop getting hit with the curse effect every so often.

I don't know of any numbers or tables sorry.

[varsovie]

I've thought about this for awhile. My various thoughts-

1. Redo cursed rings. Theres VERY few bad cursed rings, and the most important one is hunger or maybe TP. Other than that anything you put on that is cursed early isn't really enough of a threat to get off right this second, and they certainly aren't common enough to really screw you over. My suggestion would be to first make invokeable rings not auto id(lev, invis, etc) until invoked

The reason why they auto-id, is because all you really need to ID them is to press a-->x until you get the effect. It's just more keystrokes for the same effect.


PS. What do you think of a foo that temporally curse (sticky) all your equipment when it enter in LOS, or for a short duration as a hex spell? Would it be significant enough to be a really dangerous monster on certain char when it is in a group of foo (like giant eye) or in a zig?

[Eji1700]

I've thought about this for awhile. My various thoughts-

1. Redo cursed rings. Theres VERY few bad cursed rings, and the most important one is hunger or maybe TP. Other than that anything you put on that is cursed early isn't really enough of a threat to get off right this second, and they certainly aren't common enough to really screw you over. My suggestion would be to first make invokeable rings not auto id(lev, invis, etc) until invoked

The reason why they auto-id, is because all you really need to ID them is to press a-->x until you get the effect. It's just more keystrokes for the same effect.


PS. What do you think of a foo that temporally curse (sticky) all your equipment when it enter in LOS, or for a short duration as a hex spell? Would it be significant enough to be a really dangerous monster on certain char when it is in a group of foo (like giant eye) or in a zig?

I'm aware of that. That's why i'm suggesting they don't auto id and you ADD effects. All the current evokable effects are positive and not dangerous EXCEPT teleport so having them not auto id is pointless. If you added evokable effects that could be dangerous(immolation, mutate, curse other items, noise, monkey kung fu, whatever) you could make the ID game deeper and more risky, plus open up some more design space on rings which are pretty...meh....as is right now.

[mikee]

Maybe I am in a minority on this, but I do not want curses to be 'more relevant'. For me, curses in roguelikes have always ranged from a nuisance to making me want to quit a character, and I could never imagine saying to somebody, "This roguelike has a pretty cool curse system."* To me, 'interesting' or 'challenging' is best a result of actual danger and not negative strategic effects, so I would prefer option 2 or option 1.

* Actually, there is one case that I thought was OK. DoomRL has a permanently cursed item that is fairly good, but it takes damage over time and can become worse than useless if it is not repaired properly. I don't know if he remembers, but MarvinPA used it once and then died.

[cerebovssquire]

"I could never imagine saying to somebody, 'This roguelike has a pretty cool curse system.'"

Cursed items are done really well [...].

That said I think curse-like effects are only interesting when they attach a "cost" to a very strong item (see DoomRL's Cybernetic armour or distortion brands in Crawl).

[mikee]

I

[twelwe]

Just to elaborate on that quote, I'm new to Brogue, and as far as I can tell, the curses are very limited, applying only to rings, and have the exact effect the ring would have given its enchantment level, but in reverse. For example, a ring of transference gives you 10% life steal per enchantment level - a +5 RoT would give you 50% life on hit while a cursed -3 RoT would hurt YOU for 30% of the damage you deal, turning every single offensive action you make into a very careful choice. There are only seven types of rings in Brogue, so there are only seven types of curses, meaning the system is simple and effective - which is why I would classify it as done "really well."

[Sar]

Just to elaborate on that quote, I'm new to Brogue, and as far as I can tell, the curses are very limited, applying only to rings, and have the exact effect the ring would have given its enchantment level, but in reverse.

Armor and weapons in Brogue can be cursed too, which usually (always?) means they carry some nasty runic - how about a weapon that has a chance to heal your foe or an armor that has a chance to become heavier when you're hit? And if you uncurse a cursed ring of foo -x it won't become a ring of foo +x (IIRC, it's been a while), so, unless you really want it and have enough scrolls of enchantment, it still isn't very useful.

[ontoclasm]

I am not actually suggesting this, since it might be super dumb, but it's food for thought.

In the Golden Sun series curses act similar to Crawl's, but are permanent: "removing the curse" lets you take the item off, but it remains cursed and will stick to you again if you reequip it. Furthermore, curses occur exclusively on items which are quite good, but which have a minor, but noticeable, drawback (in Golden Sun's case, wearing cursed items makes you lose turns I think?).

They don't really do a good job of exploiting the potential of the system -- it's Golden Sun, what do you expect -- but it does make finding a cursed item a somewhat meaningful decision: if you want to use this (good) item, you have to commit to dealing with the downside for a while, and it will cost you to go back on that commitment.

[VargrVeum]

Anyway, I like option 3, but with a major caveat. Unequip-activated curses should *always* have a fairly significant positive effect when used, and only cause trouble when switched. Distortion and vampiric brands are excellent examples of what such a cursed item might look like. For instance, a cursed ring of protection would be a normal +N to AC for as long as you left it equipped, but when you took it off you'd get a -N malus to AC that would time out one point at a time as you gained xp. A plain old -N ring of protection is simply rubbish whether it's sticky or not, but a cursed +N ring of protection that you don't want to swap out could justifiably be generated at an even higher positive value than normal rings of protection, since you have to compensate for the drawback.

I think curses can work like mutation but on equipment - but i think keeping this distinction is a must, the idea of a curse that alters your character (even only briefly) once you've dropped the offending item seems a little messy to me. I'd be more interested in seeing solution 4 investigated - interesting curse effects on sticky items (with a scattering of good and bad stats perhaps - although to my mind a cursed item should be negative, a few nice examples of "cursed" artifacts with beneficial properties spring to mind (the one ring and mourneblade for example).

Perhaps special curses can be applied almost like brands are now - we've already got it with the ring of hunger and teleportitis...

[palin]

In the Golden Sun series curses act similar to Crawl's, but are permanent: "removing the curse" lets you take the item off, but it remains cursed and will stick to you again if you reequip it. Furthermore, curses occur exclusively on items which are quite good, but which have a minor, but noticeable, drawback (in Golden Sun's case, wearing cursed items makes you lose turns I think?).

Some artifact in Crawl has recurse too. This is an annoying property for artifacts which are actually useful in some other way... I don't think this mechanic should improve the current situation much.

[Lasty]

Let me put in another vote for a curse overhaul like the one I suggested at the end of the other curse thread. The basic premise would be 1) curses aren't sticky and instead do something else, 2) there are no scrolls of remove curse (or maybe a very rare 1-item target scroll), 3) curses appear on otherwise desirable items, and 4) curses can be removed by accomplishing some task, sort of like minor "achievements" in an achievement system

Some example possible curse effects:
* reduce all/some skill aptitudes by x, and/or reduce experience gain by 1-10%.
* Spawn higher depth monsters and/or spawn monsters more frequently.
* Intelligent monsters know your location at all times while the item is in use.
* Monsters may turn into hostile zombies after death (no xp for these zombies).
* Stat rot.
* Randomly corrupt/destroy consumables (infrequently)
* Apply Abyss-like teleport interference

Some example possible curse lift conditions:
* Kill x monster w/ this weapon or item equipped (for weapons and armor -- perhaps the weapon could force you to wield it while that monster is in view?)
* sacrifice a specific item or items (like Treasure Vaults, but with somewhat easier to acquire items)
* survive x number of x type of attack (acid, fire, giant rock . . .)
* earn x experience while wearing/wielding this item and no other item
* teleport x number of times
* become petrified
* become confused
* become paralyzed
* spend x turns or gain x experience while at a certain level of nutrition (best options are probably around Engorged and Near Starvation/Starvation)
* take x damage
* gain rot
* gain yellow (red?) glow
* perhaps steal some ideas from Xom's list of amusements.

So, an example item could be something like an ego weapon which gives a small experience penalty when wielded until you take 100 damage with it wielded, or a +5 plate armor which causes monsters to know your location until you teleport 10 times.

I think the benefit of a curse system like this is that it gives players an opportunity meet challenges that may involve altered play style (the goal of most curse systems) and get commensurate rewards, but which also allows people to opt out if the rewards are too small or the penalties too steep or the player is just not brave enough.

As for what to do with Ashenzari, I like the bind/unbind abilities already described.

[sardonica]

Let me put in another vote for a curse overhaul like the one I suggested at the end of the other curse thread. The basic premise would be 1) curses aren't sticky and instead do something else, 2) there are no scrolls of remove curse (or maybe a very rare 1-item target scroll), 3) curses appear on otherwise desirable items, and 4) curses can be removed by accomplishing some task, sort of like minor "achievements" in an achievement system

Some example possible curse effects:
* reduce all/some skill aptitudes by x, and/or reduce experience gain by 1-10%.
* Spawn higher depth monsters and/or spawn monsters more frequently.
* Intelligent monsters know your location at all times while the item is in use.
* Monsters may turn into hostile zombies after death (no xp for these zombies).
* Stat rot.
* Randomly corrupt/destroy consumables (infrequently)
* Apply Abyss-like teleport interference

Some example possible curse lift conditions:
* Kill x monster w/ this weapon or item equipped (for weapons and armor -- perhaps the weapon could force you to wield it while that monster is in view?)
* sacrifice a specific item or items (like Treasure Vaults, but with somewhat easier to acquire items)
* survive x number of x type of attack (acid, fire, giant rock . . .)
* earn x experience while wearing/wielding this item and no other item
* teleport x number of times
* become petrified
* become confused
* become paralyzed
* spend x turns or gain x experience while at a certain level of nutrition (best options are probably around Engorged and Near Starvation/Starvation)
* take x damage
* gain rot
* gain yellow (red?) glow
* perhaps steal some ideas from Xom's list of amusements.

So, an example item could be something like an ego weapon which gives a small experience penalty when wielded until you take 100 damage with it wielded, or a +5 plate armor which causes monsters to know your location until you teleport 10 times.

I think the benefit of a curse system like this is that it gives players an opportunity meet challenges that may involve altered play style (the goal of most curse systems) and get commensurate rewards, but which also allows people to opt out if the rewards are too small or the penalties too steep or the player is just not brave enough.

As for what to do with Ashenzari, I like the bind/unbind abilities already described.

This all sounds awful. I would never ever wield anything unidentified ever again, even if it meant transporting items to orc and letting orcs pick it up and try it first.

[njvack]

I am not actually suggesting this, since it might be super dumb, but it's food for thought.

In the Golden Sun series curses act similar to Crawl's, but are permanent: "removing the curse" lets you take the item off, but it remains cursed and will stick to you again if you reequip it. Furthermore, curses occur exclusively on items which are quite good, but which have a minor, but noticeable, drawback (in Golden Sun's case, wearing cursed items makes you lose turns I think?).

They don't really do a good job of exploiting the potential of the system -- it's Golden Sun, what do you expect -- but it does make finding a cursed item a somewhat meaningful decision: if you want to use this (good) item, you have to commit to dealing with the downside for a while, and it will cost you to go back on that commitment.

My take is: this is not super dumb, and is a worthy suggestion. So, in Crawl world, this might mean:

* get rid of the cursed status
* change {curse} to always stick the item to you
* allow ?RC (and Ash?) to remove a {curse} item
* either make {curse} automatically do Bad Things to you, or
* (better, I think): always pair {curse} with a negative property that would make you want to swap the item frequently.
* apply {curse} and its drawback counterpart fairly liberally to randart weapons & jewelry

So, Cekugob might get {curse} in this world. Or something that had great +Int but rF-. Or if there were a {Statrot} ego or something.

Then, instead of mixed randart jewelry/weapons being all like "I'll remove the curse and then swap it as per normal," it's more of a "Can I mitigate the drawback? Will the drawback come up rarely enough that I'll plan to ?RC in an emergency?" situation.

[Lasty]

This all sounds awful. I would never ever wield anything unidentified ever again, even if it meant transporting items to orc and letting orcs pick it up and try it first.

I'm assuming here that the fact that an item is cursed would auto-ID, even if the specific curse isn't known. Should have mentioned that.

[Bloax]

If anything, I'm in for the "Cursed items ought to be good, but unpractical to switch to and from."

Because you know, making bad items even worse is great stuff, right? Right??

[dre]

Here's a play on #4: parasitic curses

Cursed items provide (mostly) beneficial effects over time, but their hold on the player also increases as does the bad stuff that happens when the item is removed. Theory: the cursed items are parasitic and desire to remain attached to the player to feed on his soul/enjoy the thrills of dungeon diving/whatever. As such they'll provide a symbiotic relationship that grows over time, but so does the cursed item's hold over the host, such that the cost and difficulty of removing the item increases.

It could work something like this:
4 levels of curses, after a significant amount of time (several thousand turns, experience?) the curse is strengthened.
1st is benign curse as currently employed, just sticky. Curses are removed entirely with scroll or can be unwielded for minor glow or single point stat loss AND paralysis or summon minor hostile (so not during combat).
2nd level, curses begin to have minor positive effects (below), but unwielding now causes worse effects . Remove curse reverts to 1st level.
3rd level, more of the same, remove curse pushes down to 2nd level
4th level, at this point parasite has deeply attached itself to the host, and unwielding will have major effects. Something along the lines of high % chance of banishment, heavy stat loss, yellow+ glow, and paralysis (which occurs before banishment so it's not an escape button). Remove curse does nothing or only works at some p << .5

Benefits would have to balanced so that staying below level 4 doesn't induce micro-management. Significant warnings should let the player know when levels are about to increase. Benefits items provide at level 2, 3, 4 could be something like:
weapon - slaying of +1, +2, +4
body armour - AC (or EV for robes/leather) of +1, +2, +4
other armour - AC & EV of +0, +0, +1
rings - MP of +1, +1, +3
amulet - nothing, nothing, random resist likely to be useful

So players would have to choose between item flexibility and the benefits of allowing the curse to strengthen.

[kaikaun]

Here's my proposal:

Cursed items simply apply magical contamination on unequip. The amount of contamination is variable: the more cursed the item, the more contamination it applies on unequip. Cursing an already cursed item increases the contamination it applies. Different sources of curses apply different amounts of contamination (less for scrolls and early game mummies, more for Zot traps and other late game nasties). The amount of contamination on a cursed item is visible in inventory (as a level of glow like that on a magically contaminated player). The level of contamination on items decays slowly with time (XP gained), but not below a minimum value (cursed items stay cursed if just allowed to decay). Curses are renamed to "contamination", and the concept of magical contamination of the player and his items are aligned.

Scrolls of remove curse become scrolls of decontamination. They decrease the level of magical contamination on all items and the player (potentially to zero, "uncursing" the item). They remain common, but the amount of contamination they remove is tweaked up or down to balance them. If possible, more magical contamination granting effects are added to the game over time, given there is now a tool to deal with it (e.g. powerful spells may automatically contaminate the player even on success). Magical contamination becomes something to manage strategically rather than just a punishment = more choices.

Stickiness becomes completely divorced from the concept of curses, and purely associated with Ashenzari. Ashenzari becomes god of "bondage" instead of "curses". Followers immediately get a "Bind item" invocation to bind any worn item, making it sticky and granting a small immediate amount of piety (and the various bonuses Ash gives for wearing sticky items). They also immediately (or maybe at 1*) get an "Unbind Item" invocation to make items unsticky again, costing a large scaling amount of piety (1/3 current piety or so).

The advantages of this are simplicity (only one curse effect represented by a single integer, small changes to Ash, reuse of contamination concept) while giving scope for different levels of curses, making magical contamination a strategic choice, and splitting Ash from item curses so they can be developed separately. The main disadvantage is blandness - magical contamination just seems blah.

[ElectricAlbatross]

Stickiness becomes completely divorced from the concept of curses, and purely associated with Ashenzari. Ashenzari becomes god of "bondage" instead of "curses". Followers immediately get a "Bind item" invocation to bind any worn item, making it sticky and granting a small immediate amount of piety (and the various bonuses Ash gives for wearing sticky items). They also immediately (or maybe at 1*) get an "Unbind Item" invocation to make items unsticky again, costing a large scaling amount of piety (1/3 current piety or so).

That just sounds much more fun than the current Ash, honestly.

[Bloax]

Draining 1/3rd of the piety is equal to taking off an amulet of faith, which is a quite hefty price.

If unbinding items with Ashenzari did that, you wouldn't bind anything until you found some stuff you were sure would serve you for a very long time.

1/8th Or so sounds more like it. Perhaps going up to 1/5th depending on how long you've had it bound.

[Galefury]

If unbinding items with Ashenzari did that, you wouldn't bind anything until you found some stuff you were sure would serve you for a very long time.

No.

[Bloax]

I hope you enjoy getting set back a long way if you find some sweet, sweet loot then.

Of course piety isn't all that important with Ashenzari, but getting your piety halved for switching two items is a bit too punishing if you ask me.

[dpeg]

Albatross: I doubt that it makes Ashenzari more fun, because it removes decisions. (Effectively, instead of using scrolls, a finite currency, you'd use piety, an repleneshable one.) It would be playable, of course, and a reasonable course of action if curses (as used by the god) are not in the game anymore.

Anyway, here is a recap of the various proposals (doing nothing or scrapping everything are always options, so are not listed):

1. Non-sticky. Harmful effect upon unequip.
   The original proposal uses glow. Advantage: Just a single effect (not a list), although there was some criticism about glow as a curse effect. Drawback of the system: does not scale.
2. Sticky. Curses decrease with time (not below sticky) and give bad effects on the item.
   Scales well (items found on D:1-10 and cursed by plain mummies would be only sticky; later on, curses would be stronger). Drawback: linking curse reduction to xp obfuscates the state of your cursed item.
3. Cursed items re-curse. Only good items are cursed.
4. Curses increase with time (until permanent) and give good effects on the item.

I have been reading through all the comments again. I got the impression that two bits got confused: cursed items you find on the ground, and curses you get on items you have chosen. The former have a tendency to be boring, the latter have good chances to be interesting (because it probably affects a good item). It seems hopeless to me to go for an interesting mechanic if only curses of the first kind are considered.

crate has a point that mummy death curses should not be scummable. This can be achieved in a number of ways.

[njvack]

4) Curses increase with time (until permanent) and give good effects on the item.

Am I understanding this right (sigh, I'm getting lost in the wiki and thread) that this would be roughly comparable to the existing Gourmand mechanic, except more permanent?

[Galefury]

Maybe a more general look at the role of curses in crawl is needed to properly evaluate the proposals. These are the things current curses affect (that I could think of):
1) Resource drain (?RC) for use-iding equipment.
2) Resource drain (?RC) for killing plain mummies without jumping through hoops.
3) Resource drain (?RC) for swapping artifacts with the recurse property and for using the scythe of curses.
4) Strategical and tactical effects from sticky equipment. These are only very noticeable early on when you don't have ?RC, and on deep mummy zig floors where literally everything you are carrying will get cursed, you often want to swap equipment, and turns are extremely valuable. Even in Tomb you often have the tactical flexibility to avoid getting many items cursed, delay swapping, or uncurse things in combat if necessary.
5) Sticky curses are integral to Ashenzari gameplay, where their relevance is elevated by compensating for the bad effect in other ways and multiple resources are needed to deal with curses because you sometimes want to curse things, sometimes to uncurse.
6) Enchantment scrolls can be used instead of ?RC. This is currently irrelevant outside the very early game because ?RC are so abundant.

When looking at it like this it becomes immediately obvious why current curses are boring and rarely relevant. They drain a resource that is used for nothing else and is usually abundant. The effect of curses is extremely strong, so if you're out of ?RC curses can be extremely crippling and a death sentence early on. If you have ?RC they are completely irrelevant unless you are in a situation where you want to swap equipment and every turn counts.

The new proposals should be assessed in the same way to see if they might be better. Multiple resources (at least two) should be useful in dealing with curses, and each of them should be valuable and scarce. To be interesting curses need to have relevant tactical or strategical effects (sticky curses actually work quite well for this, but may be too harsh early on, especially if ?RC would become rarer). It is fine if it is possible to avoid curse effects completely by spending resources: if the resources are scarce and useful for other things it may be preferable to just deal with the curse.

[nicolae]

Multiple resources (at least two) should be useful in dealing with curses, and each of them should be valuable and scarce.

Don't scrolls of enchant whatever also remove curses?

[Galefury]

Yes, I mentioned that. But who cares? That's where the "each of them should be valuable and scarce" part comes in.

[nicolae]

Yes, I mentioned that.

Oh, damn, so you did.

[dpeg]

galehar: Many thanks for the analysis. That's the way one should think about it, in my opinion.

As I tried to indicate in my last posting, curses will not be interesting if we only consider items on the floor. Yes, item can be created with a curse (and ?identify is a way to get around that, but, *warning* currently monsters can be tediously employed for that as well), but that is more like a gamble, and the decisions restrict to "Do I have enough ?RC to use-id this blindly?" (also goes for randarts later in the game), "Should I rather ?id before" (which you may have to do with jewellery anyway) and that's about it.

So for this to be interesting, we need an external source of curses. As listed by galehar, there are two: plain mummy death curses and re-cursing randarts. Both of these are interesting, but not good enough. Regarding re-cursing randarts, I believe the property comes up too rarely to matter: you need it on a really good randart and even then you just stick with it; no need to regularly uncurse.
Plain mummy curses are irrelevant because quite early it just reduces to ping-pong: they curse, you uncurse.

I believe that certain monsters (including all kinds of mummies, but also liches) could have a curse spell which curses an item in use (this addresses "all jewellery cursed post Tomb", and prevents mummy dancing). In my suggestion, these curses would be higher level, so come with additional bad effects (think temporary bad mutation, but also "very noisy" on weapons). This is something that may or may not require immediate action, action after the battle, or perhaps can be waited out. That's the hope.

[Galefury]

Tired? I'm not galehar.

[Lasty]

Is anyone a fan of sticky curses on floor items? Because scrolls of identify are usually better used on artifacts/jewelry/potions/other scrolls, use-IDing is the main way to ID weapons and armor. Remove curse scrolls effectively a scroll of "allow yourself to use-ID until you hit a curse". If we keep sticky curses on floor items, I'd really like to see a better way to ID weapons/armor than use-IDing or burning (relatively) precious ID scrolls.

[dtsund]

Here's a proposal that I briefly alluded to on IRC, based on some of the stuff here:

Everything in the game is cursed when you find it. When you equip something new, you are forced to either wear it until the curse wears off through experience gain or until you uncurse it with a (rare) scroll of remove curse (which would only work on one item at once).

Most curses will be relatively short-lived, on the order of clearing out a dungeon floor, and will simply prevent you from removing the item. Once you've kept it on for long enough, it is permanently uncursed and you can swap it on and off freely. Some items will have nastier curses on them, which have negative effects on the player and take longer to break.

Purely bad equipment probably should not exist; 'bad' items could persist as nasty curses on good items. One possible idea, for jewelery, would be an effect-inverting curse, which turns (for example) =sustenance into =hunger until it wears off.

I would say, in this system, all things should auto-ID when worn, so the player can gauge how much trouble a given curse puts him or her into.

What this would mean: if you find a new piece of equipment and decide to use it, you are typically committing yourself to its use for the next floor or so (that or burn resources). If you equip an item and it turns out to have a nasty curse on it, you have the choice between burning one of your scrolls or suffering through it for a while; ideally, incautious players should typically have to deal with more nasty curses than he or she can instantly break with scrolls.

Ashenzari, for this system, could simply automatically apply a weak curse to every item worn, even if it had been previously uncursed.

[njvack]

Everything in the game is cursed when you find it.
...
Purely bad equipment probably should not exist; 'bad' items could persist as nasty curses on good items. One possible idea, for jewelery, would be an effect-inverting curse, which turns (for example) =sustenance into =hunger until it wears off.

My first impression is that I rather like this idea. Inversion seems particularly fun -- maybe your =cTele is, for a while, either *Tele or even Stasis.

Would you imagine that bad egos would still exist on randarts, or that once the item is fully uncursed, it would wind up purely good? (Some items, of course, will never be good enough to make it worth waiting out the curse.)

Would items sometimes unsticky before the other bad effects wore off? Would ?RC fully uncurse an item, or just make it unsticky? (I can see either or two scrolls, depending on rarity.) Would curses gradually wear off -- so a curse that was -AC might be -5AC at the start but only -1AC at the end?

If one of the followon effects of this would be "identify stuff before trying it on" that seems like a net positive -- there are lots of identify scrolls in the game, but not so many that identifying *everything* would be a no-brainer.