Curses feature branch

[bio_hazard]

Some of the names are fairly weak. I prefer "Winter's chill" from Sil, to "chilled to the bone", and "Of Burning" for the other. Perhaps the former should grant RES_COLD[2], so temp cold resist gives temporary immunity? Or give off an aura like in some variants?

In any case, this is a whole lot of junk. I'd rather see fewer basic ego types, and more that combine different features. Further, IMPAIR_HP is already available on a ring. I don't see the point of putting it on some boring ego item. And I'm not sure of the point of the ones with obvious penalties like "boots of slowness", which are IDed immediately on use, or even on pickup, then discarded. Without a stick curse, things like this are just zero-risk junk. Even the ones with fire vulnerability are pretty dull. If you want to make something scary, make it a subtle thing that may only be discovered too late, like RES_NETHER[-1].

The death drake breathes nether. -more- You learn something new about your armor. -more- You die.

I kind of agree with this, although I haven't played the RuneID version enough to know how the mid-game would play out in this way. There's something to be said for having a surplus of both positive flags and curses, and for each game to take an unknown sample of these (important ones common to every game, and others showing up less frequently).

A few ideas- I don't know how to fully implement these but it seemed in particular lights and launchers weren't well represented

name:easy mark
type:amulet
type:ring
effect:SUMMON:THIEF
dice:1
time:1d100
msg:A thief steps out of the shadows
descccasionally summons a thief

#don’t know how to implement this
#maybe easier to have it as a timed effect
nameoverty
type:soft armor
type:hard armor
type:shield
effect: #lose gold on hit
msg:the hit jostles your purse
desc:each point of damage causes loss of 1 gold

name:crooked
type:bow
# doubles chance of ammo breakage
# halves range
desc: your launcher twists in your hands

name:famine
type:light
#hunger clock x100
#light radius + 1 more than typical
desc:the light is brighter, and uses your energy as fuel

name:fireflash
type:light
effect:TIMED_INC:FIRE
dice:1+1d4
time:1d1000
msg:your light scorches you
desc:burns you and the contents of your pack

name:darkshroud
type:light
effect:TIMED_INC:#light radius = 0
dice:10+1d10
time:1d50
msg:your light dims to an ember
desc:light periodically flickers to nothing

name:tarnished
type:ring
type:amulet
type:light
#hit with disenchantment every 1d1000
msg:a wave of evil magic passes over you
desc:chance to disenchant other equipped items

name:fractured
type:bow
type:digger
type:hafted
typeolearm
type:sword
type:boots
type:gloves
type:helm
type:crown
type:shield
type:cloak
type:soft armor
type:hard armor
type:dragon armor
#item has a chance to break
msg: crumbles!

[Pete Mack]

Yeah, those are good. Sil has a nasty one, where potentially out-of-depth undead appear more frequently. Not summoned in any obvious way, just often present.

[d_m]

Yeah, those are good. Sil has a nasty one, where potentially out-of-depth undead appear more frequently. Not summoned in any obvious way, just often present.

Really like things like this, especially on otherwise-great items.

E.g. Mormegil should probably just be something like this. Super powerful sword, super powerful dragons, good luck and god speed!

[Derakon]

Now I'm thinking about Nuclear Throne, which is an action roguelike with Crowns which bestow mixed-blessing effects on you. Off the top of my head, here's some adaptations for Angband:

* Blood: when equipped the spawn rate for new enemies on the level is vastly higher (initial level population remains the same)
* Curses: high chance of spawning out-of-depth equipment items, but they always are cursed
* Haste: items drop in larger stacks, but vanish after a few turns if not picked up
* Life: high chance that dropped items are replaced by (weak) healing potions
* Risk: deal extra damage when below 50% health
* Hatred: +10% experience from kills, but take (mlvl) damage per kill.
* Luck: increased chance of critical hits; enemy attacks always deal max damage

[Nick]

All right, a few points to make here:

1. Looks like everyone here is better at designing curses than me, which is wonderful. New object/monster properties can be created if necessary to implement any of these, so don't be bound by what appears currently possible. If I can't see how to implement something, I'll let you know.
2. I based a lot of mine on the old cursed egos, and I'm thinking that was OK as a starting point, but most of them should probably be just ditched.
3. More generally, the whole curse system can be regarded as a starting point. If any aspect of it isn't working or could be done better, we can just change it.
4. Do we think there is an optimal number of curses? Is more always better?

Also, glad to see curses being given the PowerWyrm treatment. Help certainly hasn't caught up, nor has object power (which along with monster power needs some work). I've fixed the bad free in development, and have to think a bit deeper about the issue you highlighted with the enveloping curse, and the item tester.

[Pete Mack]

The trouble with the old ego items as a base is they just are boring without sticky curse. You equip one, immediately find it is bad then take it off. For non-sticky curses to work, they either need to be an obvious malus on an otherwise great item (like Doomcaller, Calris, Deathwreaker) or they need to be in some way subtle. Otherwise they are just junk.

[Nick]

The trouble with the old ego items as a base is they just are boring without sticky curse. You equip one, immediately find it is bad then take it off. For non-sticky curses to work, they either need to be an obvious malus on an otherwise great item (like Doomcaller, Calris, Deathwreaker) or they need to be in some way subtle. Otherwise they are just junk.

The first of these was the aim - items generated with a curse have their generation level increased. I think this mechanic needs some work, though.

[d_m]

My subjective experience is that a lot of curses show up as runes and are sometimes quite hard to ID, which ends up being more annoying than the actual curse in some cases.

I'm not sure if separating runes and curses makes sense, but it might make the ID-by-use story a little easier.

[Derakon]

My subjective experience is that a lot of curses show up as runes and are sometimes quite hard to ID, which ends up being more annoying than the actual curse in some cases.

I'm not sure if separating runes and curses makes sense, but it might make the ID-by-use story a little easier.

I'd be inclined to say that curses should be runes, but they should also be attached to powerful-enough abilities that you're tempted to keep the item anyway. So then if you ID a curse rune, you either a) know why this item is otherwise so powerful, or b) are left wondering what the cool power the item has to compensate for the curse.

[Nick]

My subjective experience is that a lot of curses show up as runes and are sometimes quite hard to ID, which ends up being more annoying than the actual curse in some cases.

I'm not sure if separating runes and curses makes sense, but it might make the ID-by-use story a little easier.

Yeah, that might make sense - certainly there are currently some curses (like impair mana) which pretty much coincide, so you would learn both at once.

How do you think it should work?

[luneya]

Curses as runes makes sense, but there should be a bit more to help the player figure out what's what. Perhaps include a curse-identification spell which is easier to cast than regular rune id (if we're keeping some magical form of the latter) but only works on curse runes. This should be available to both types of casters (and probably also in scroll form), perhaps at a lower level for divine magic than for arcane.

I haven't actually playtested the development builds, but is the player told how hard a particular curse is to break, or is this information hidden? If the player isn't told, they should be, given the negative consequences of failed curse-breaking. Perhaps this is another useful function for the curse-id spell. We could have different instances of the same curse at different levels of difficulty, have the nature of the curse be revealed by rune, but the breaking strength require magical id of the individual item to find. This doesn't lead to the having-to-id-everything annoyance that the rune system is meant to stop, as you only need to use an id scroll/spell if you want to break the curse.

[Ingwe Ingweron]

....I haven't actually playtested the development builds, but .....

So, you just have a host of assumptions about how the new curses work without any actual experience of them? That's confusing.

I'll try to clarify, but you might want to at least give it a shot before making all these assumptions.

- Pick up a cursed item into the pack - get the message, "Oops, it feels deathly cold."

- At this point, you know the item is cursed, but if you don't yet know the "rune" for that curse, you must use an ?Identify or spell of Identify to learn the rune. Or wield it and have the associated event occur. Some curses would be obvious immediately, others would might be more difficult to learn. But you can remove the object unless it has the "sticky" curse.

- Use ?Remove Curse, ?*Remove Curse*, or spell of Remove Curse or Dispel Curse. You will get a message of which curse to remove (if multiple curses are on an object, select which one), and the "power" of the curse is displayed numerically.

- A failed Remove Curse can cause an item to become more "fragile" and successive removal attempts have an increased chance to destroy the item.

I think that's it. Nick, have I missed anything?

[Nick]

I think that's it. Nick, have I missed anything?

Nope, sounds about right.

[Pete Mack]

Cloak of the Necromancer [1,+7-15] Random high resist (-1) (Based on Cloak of aman.)

[PowerWyrm]

- Pick up a cursed item into the pack - get the message, "Oops, it feels deathly cold."

Didn't see anything in the code that does it (seems you get the message when equipping a cursed item)... but of course I didn't check everything yet. My problem atm is being able to generate cursed items with the game... I didn't manage to generate one yet.

The big problem with this system is having attached a lot of things to curses: combat mods, stats, flags... This means you will have to replace all instances of obj->something in the code with a function that iterates over obj + obj->curses, and do the same with obj->known->something. This will rapidly become an unmanageable mess...

Moreover, it seems that the code reveals a curse as soon as one of its component is learned. Doesn't feel right. At this point, I see two ways of dealing with the problem: the simple way and the complete way.

First and easy way: learn curses completely on wield, and only on wield. All properties attached to the curse are noticed on wield, and the rune is learned. No individual rune is learned. For example: an item with "anchor" curse is generated (RES_NEXUS + NO_TELEPORT), you learn "anchor" and notice nexus resistance + no teleport on wield, but don't learn "rnexus" and "no teleport" runes, even if you get breathed on by a nexus hound or teleported around by a nexus Q. Consider that the flags attached to the curse are "abstract" flags and have no meaning outside of the curse, although they provide the corresponding effects.

Second and complete way: don't make curses as runes, just consider them as any other object. Basically they are just a bag of properties -- or a bag of runes. In this case, object knowledge is made by browsing the object itself and the associated curses. For example: an item with "anchor" curse is generated (RES_NEXUS + NO_TELEPORT), you notice the "anchor" curse on wield, you learn "rnexus" and "no teleport" runes like any other object, when you get breathed on by a nexus hound or teleported around by a nexus Q for example. Consider that the flags attached to the curse are "concrete" flags and that the curse is abstract.

In either way, there will be a substantial amount of reworking to do for curses to work fine...