Poison effect

[galehar]

This is a topic tangential to both Formicid and rune lock. They both make poison more frequent and dangerous, so the issues with the poison effect are now more apparent. I've seen suggestions like remove rP- for formicid or make rP+ easier to find before the rune lock. I don't like either, so instead I propose to address the issues of the poison effect. What are they?

Tedium because of increased resting (which even need some special configuration to avoid long rest being constantly interrupted).
Variance of duration and damage. It's hard to predict how much damage you're going to take or how long you'll be poisoned.
Rarely tactically relevant. Because the damage is spread over a long time, the amount you take during the fight is rarely relevant.

So I suggest that poison duration is shorter but the average damage remains the same. Would be good to simplify it too, the current code is quite messy. Which is why this proposal doesn't contain actual formulas and numbers, I don't have time to source dive and make a complete proposal. Any help with that appreciated.

[nordetsa]

If the damage doesn't get modified I don't know what's the point of changing how poison works. Formicids will continue to be vulnerable to poison, as with any character that don't have rPois.

[tompliss]

One thing needs to be considered, in a "shorter but with the same mean" revamp is that character usually regen their HP while poisonned.

On your first point, I think rP- should stay for Formicids (as well as for most insect monsters), but its damage-boost should be decreased to the "standard" +50% rC- and rF- give.

[EDIT] @ nordetsa : Formicids usually die because of bad luck on poison. Being vulnerable give them twice the charges, so a mean of twice the ticks. Except charge removal is random, and you'll meet enough adders and scorpions that at least once before d:5, you'll be poisonned for twice the normal duration...

[galehar]

If the damage doesn't get modified I don't know what's the point of changing how poison works.

Less resting, less tedium, more tactically relevant (because more damage per turn).

Formicids will continue to be vulnerable to poison, as with any character that don't have rPois.

Well yeah, sure. The point is not to make poison less threatening.

[Galefury]

If it's made shorter, the total damage should probably be a lot lower than it is now. Because currently it is really high, but most of it is mitigated by HP regen. Also, making it last shorter weakens the distinction from other damage types, so I'm not sure poison is even worth keeping like that. We already have lots of things that do some quick extra damage.

Maybe I have played too much Desktop Dungeons, but poison could just block regeneration like sickness, but not go away by waiting. Curing should still cure it of course. It could go away by earning exp or by exploration, or maybe by using some other limited resource than curing. But such a change would require looking at everything in the game that poisons...

[Sar]

Well, right now poison is rarely deadly for non-Formicids outside of early game, I think. rPois is great for not being ouched by PArrows/Venom Bolts (not many things use those) but outside of those cases it's mostly an anti-annoyance measure (the place I like doing the least without it is actually Zot because draconians/golden dragons breath poison clouds non-stop). Otherwise you can usually use curing if you get really bad poison, or wait it out.
Oh, rPois is also good vs. meph clouds (so player ghosts and swamp drakes) and curare (Sonja, rare kobolds).

[galehar]

If it's made shorter, the total damage should probably be a lot lower than it is now.

Yeah, maybe. Would be good to have some hard data about poison damage.

Also, making it last shorter weakens the distinction from other damage types, so I'm not sure poison is even worth keeping like that. We already have lots of things that do some quick extra damage.

I think there is some room between "instant damage" and "spread over such a long time that it's mitigated by regeneration". Maybe start by dividing the duration by 2 or 3.

[tompliss]

Well, right now poison is rarely deadly for non-Formicids outside of early game, [..] rPois is also good vs. meph clouds (so player ghosts and swamp drakes) and curare (Sonja, rare kobolds).

That sums up the non-Formicid about Poison. And I think it's good that way : it's always nice to have rPois in the early dungeon because you won't worry about adders and scorpions, but not necessary, as rC is not necessary but still useful agaisnt Ice Beasts. And it's relevant in mid-game because of Curare and the Poison branch, and other side effects. And it's relevant in late-game because of side effects, as can be clarity or stasis. And I won't talk about extended as I don't know.

[EDIT] @ galehar : I don't think the duration is the worst part of the poison effect (even if a little help in the "wait for the poison to end spamming 5" thing could be better), but about the variance. You have a 1 in 8 chance to loose one charge of poison. If you're unlucky, you're poisonned forever. (if the wiki is still corerct about it) :

  Code:

   Each turn, you have a poison_level in 5 chance of taking damage from poison; at 5+ levels of poison, this is automatic. The damage is as follows:
        If you have more than 10 levels of poisoning, if 1d(poison_level) is greater than 8, you will take 1d10+4 damage.
        If you have 5-10 levels of poisoning, there is a 50% chance you will take 2-3 points of damage.
        Otherwise, you will take 1 point of damage.
    You also have a 1 in 8 chance of losing one level of poison each turn (1 in 3 if you have been reduced to 1 hp).
    If your Cheibriados piety is at * or above (slowed/strengthened metabolism), there is a 50% chance you will take no damage – but this will bypass the check to lose a level of poison as well. This effectively halves the damage per turn of poison but doubles its duration (which also gives you more time to regenerate health).

[nordetsa]

Actually, what's the formula for the poison damage? LearnedDB doesn't seem to say anything about the player poison damage. Sometimes the poison wears off quickly, sometimes it seems to stay forever and force me to quaff a potion of curing.

[Sar]

But my point was that it's not actually very relevant to your survival (you can quaff !resistance if you encounter Arachne), but it's nice to have because it makes you press 5 less after fights. I'm not sure it's a good thing.

[nordetsa]

And I won't talk about extended as I don't know.

I think rPois is still relevant in the extended because of the occasional Green Death spawning in Pandemonium. Its poison arrow hurts without rPois, and it can get you to red poison fairly quick.

[tompliss]

I think it's relevant to your choices. You won't go in the Poison branch without worrying, if you haven't rPois. It's the same about going in the volcano without rF or in the Ice Cave without rC, but with a rune in the end

[Sar]

I will and I did. Unlike fire and cold damage, most poison damage doesn't kill you outright, and monsters have to get through your AC/EV first.

[pubby]

I'd like to see poison implemented as a fixed, predictable amount of damage spread out over time. When you get hit by a poison attack, the health bar would show how much damage you'll eventually take, in purple:

====================

This is calculated by taking your current regeneration rate into account, and so both the poison formula and the regeneration formula would need to be deterministic.

Resting while poisoned would only take one key press with this system.

[tompliss]

I will and I did.

without worrying

but thank you for reading...
And no, going there after clearing Elf is not "without worrying".

====================

<3

[Sar]

I'm not sure why you assume I didn't read that part. What is there to worry in Spider without rPois? Arachne? I don't have to fight her. That I'll have to quaff one of my 20 curing potions? Oh no! Whatever shall we dooo~!
That assumes I have good defences and HP, but if I don't I still probably worry more about wolf spiders (speed:15, dam:25, 15) mauling the shit out of me and not poison status.

[Mankeli]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average). If you want this to not be a buff, you can increase the damage for the poisoned status while it is active. Make this be the case for all species not just Formicids.

Flavor: you found an antidote when you killed the poisonous monster

[dck]

Yes, very yes.

[njvack]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average). If you want this to not be a buff, you can increase the damage for the poisoned status while it is active. Make this be the case for all species no just Formicids.

Flavor: you found an antidote when you killed the poisonous monster

If we're looking for a more serious change that would require all kinds of rebalancing (and come on, who isn't?) monster poison could be changed to a "HP siphon" effect -- some fraction of the poison damage heals the monster. Then it makes sense that killing the monster stops the drain.

This would, I think, be madly OP as a player effect.

[johlstei]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average). If you want this to not be a buff, you can increase the damage for the poisoned status while it is active. Make this be the case for all species no just Formicids.

Flavor: you found an antidote when you killed the poisonous monster

If we're looking for a more serious change that would require all kinds of rebalancing (and come on, who isn't?) monster poison could be changed to a "HP siphon" effect -- some fraction of the poison damage heals the monster. Then it makes sense that killing the monster stops the drain.

This would, I think, be madly OP as a player effect.

As a player effect it would also cause tedious behavior like stashing monsters and only hitting weak monsters until they are poisoned so you get maximum healing. It might be interesting as monster only.

[galehar]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average).

Reducing or removing the need to rest off poison is indeed one of the goals of the proposal. Although I think there are more elegant ways to go about it than that.

Flavor: you found an antidote when you killed the poisonous monster

So, every random snake and spider in the dungeon is carrying an antidote phial which you can instantly grab and drink once you've killed it even if it's 8 cells away? And you call that flavour?

[Mankeli]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average). If you want this to not be a buff, you can increase the damage for the poisoned status while it is active. Make this be the case for all species no just Formicids.

Flavor: you found an antidote when you killed the poisonous monster

If we're looking for a more serious change that would require all kinds of rebalancing (and come on, who isn't?) monster poison could be changed to a "HP siphon" effect -- some fraction of the poison damage heals the monster. Then it makes sense that killing the monster stops the drain.

This would, I think, be madly OP as a player effect.

I'm looking for the least complex and easiest solution to a (mostly) gameplay problem. If the poison damage per turn would be left untouched in my proposal, it would be a buff to players. With the rune lock being in place forcing you to (possibly) tackle a poison branch earlier than before, I don't think that even leaving the poison damage per turn as-is would be bad . I don't understand why the solution should be one that would require major rebalancing if we can avoid it. I also fail to see how changing monster poison attacks to vampiric draining would help and don't see why it would make any sense. Also if I have to choose between "making sense" and reducind pointless fiving, I'd go for the latter.

[Mankeli]

So, every random snake and spider in the dungeon is carrying an antidote phial which you can instantly grab and drink once you've killed it even if it's 8 cells away? And you call that flavour?

I call that improving gameplay. The flavor part is for those who care, I don't. Besides, currently there is too much flavor: In real life, you can be poisoned for months, or at least have effects of being poisoned (molds for example) and Crawl emulates this very nicely with the after-getting-poisoned-five-a-thlon you have to go through when poisoned.

[Mankeli]

Flavor: you found an antidote when you killed the poisonous monster

So, every random snake and spider in the dungeon is carrying an antidote phial which you can instantly grab and drink once you've killed it even if it's 8 cells away? And you call that flavour?

I call that improving gameplay. The flavor part is for those who care, I don't. Besides, currently there is too much flavor: In real life, you can be poisoned for months, or at least have effects of being poisoned (molds for example) and Crawl emulates this very nicely with the after-getting-poisoned-five-a-thlon you have to go through when poisoned.

EDIT. About producing antidotes: many antidotes are actually produced from injecting poison into an animal like horse and then collecting the antivenom substances the horse has produced in its body. So the poison of the poisonous animal actually is an important part of the process of making the antivenom. I don't think we need a spesific "Laboratory" branch in Crawl though, where you should go after every poisoning but yeah, it would add flavour if it's important

[minstrel]

What about having poison counters disappear quickly with low tension? That takes care of the click 5 syndrome, while leaving the in-combat tactical effects alone to be adjusted, or not, as desired.

[duvessa]

Make the "poisoned" status disappear as soon as you have killed the monster that has poisoned you. There. No more "fiving" after adders/sea snakes/nagas (or if you escaped before killing the monster, at least much less fiving on average).

Reducing or removing the need to rest off poison is indeed one of the goals of the proposal. Although I think there are more elegant ways to go about it than that.

Flavor: you found an antidote when you killed the poisonous monster

So, every random snake and spider in the dungeon is carrying an antidote phial which you can instantly grab and drink once you've killed it even if it's 8 cells away? And you call that flavour?

Crawl also misidentifies numerous venomous animals as poisonous, and nobody seems to complain about that. I see no further problem with also misidentifying them as producing antivenom, and "8 cells away" is a pretty ridiculous nitpick when you are also e.g. notified when something dies on the other side of the level; these things are abstracted away all the time (you aren't shown actually taking your gloves off when you put on a ring, or actually unwielding your great mace to butcher a corpse).
That said this proposal doesn't address the annoyance of waiting out poison when you don't kill the source, so I prefer making poison faster, though you'd need to divide its duration by way more than 2.
Or maybe make it go away if you don't suffer any non-poison damage for 5 consecutive turns. That would have no effect on genuinely dangerous fights but would make resting much less of a pain.
It'd also be nice if poison weren't on so many monsters just for the hell of it. Monster nagas, jumping spiders, wolf spiders, spiny worms, ghost moths, etc. do not need poison attacks.

[dck]

Well you heal from getting hit with a battleaxe to the point of near death by laying around for a bit so I don't think having your character automagically get the antidote from whatever poisoned him is bad.

[some12fat2move]

I think the best way to flavor it would be that your character, as an adventurer, knows treatments to poison but can't apply them when enemies are right there trying to kill you.

[galehar]

The flavor part is for those who care, I don't.

Yeah, but we're not talking about Mankeli's crawl. If you don't care about how crawl is designed your contributions are useless.
Understand that flavour is not just about aesthetics. If something is logical and make sense, it's much easier to explain it to the player. With your proposal, people will ask why they are suddenly cured of poison after killing a snake.
When gameplay demands it, flavour can be torn appart. It is less important. But I'm pretty sure we can find better and more elegant way to address the tedium of poison resting.

[dck]

While we're on the subject of poison, I believe the huge red poison you can randomly get from zot traps and which can kill you through 300 hps, troll regen, two rings of regen and TH without ever dropping from extremely sick shouldn't exist.

[Baldu2]

How about summons decrease their time remaining at twice the rate if their summoner is dead and poison status is guaranteed to lose 1 level per turn when no tension? At medium tension, poison could be identical, at high tension it could inflict higher damage ("your excitement helps the poison course faster through your body").

[Galefury]

I think the best way to flavor it would be that your character, as an adventurer, knows treatments to poison but can't apply them when enemies are right there trying to kill you.

This sounds pretty okay. Poison could just go away fast if there are no monsters in sight (yes blablabla stepping around corners and healing yourself while running at full speed, it already works for HP).

Edit: oh, also, please don't use tension, it's pretty bad.

[galehar]

Poison could just go away fast if there are no monsters in sight
Edit: oh, also, please don't use tension, it's pretty bad.

Well, "monsters in LOS" and tension is almost the same thing. At least, they are strongly linked. It would be terrible to link any of the 2 to poison removal. Imagine this:

newbie: I'm highly poisoned, I have 5 HP left and no healing potions! What do I do?
crawler: read a scroll of fog
newbie: Whhaaat?

[some12fat2move]

Poison could just go away fast if there are no monsters in sight
Edit: oh, also, please don't use tension, it's pretty bad.

Well, "monsters in LOS" and tension is almost the same thing. At least, they are strongly linked. It would be terrible to link any of the 2 to poison removal. Imagine this:

newbie: I'm highly poisoned, I have 5 HP left and no healing potions! What do I do?
crawler: read a scroll of fog
newbie: Whhaaat?

Good point. Maybe poison damage and duration could be determined by the distance from you to the closest enemy on the floor - so the farther away you are from your foes, your character doesn't have to be as cautious about danger and is better able to tend to illness or perform actions in a way that doesn't aggravate poison as much. Mathematically one way to do this would be to divide poison DoT by the number of tiles between you and the closest enemy (damage/distance). So when adjacent (1 tile from you to the enemy, so distance is 1) poison does its full damage (which should probably be increased on a monster-by-monster basis, to make it intense) and you are encouraged to get away from there. Similarly, the rate at which poison status goes away would be multiplied by the distance between you and the closest enemy. Distance could be capped at, say, 10. Under this system the hypothetical dialogue would be more like:

newbie: I'm highly poisoned, I have 5 HP left and no healing potions! What do I do?
crawler: get out of there and stay away from enemies...use blink and run, hope teleport works, something like that

[duvessa]

This is exactly why I suggested this:

Or maybe make it go away if you don't suffer any non-poison damage for 5 consecutive turns.

[Galefury]

So you suggest having action and attack speed affect it?

Also AC and EV. Of course those already kind of affect poison damage. So do action and attack speed for that matter. If someone implements something like this whoever implements it is going to come up with their own thing, so talking about the specifics is probably not very useful.

I think "delayed extra damage that you know is going to happen but can prevent by leaving combat or quaffing curing" is still really really similar to all other kinds of extra damage, but probably more interesting than "extra damage with a very slight delay that you can prevent by quaffing curing". So the general idea of making poison go away quickly when out of combat sounds pretty good to me. Better than just making poison faster, although that would probably also be an improvement.

[TwilightPhoenix]

Off-hand suggestion: if a player is able to rest for, say, 20 turns or so, poison can either cure itself instantly or drop down to the next level (red to light red to yellow to none), whichever is better for gameplay. 20ish turns is long enough that a player can't get healed by ducking behind a corner or reading a scroll of fog, but short enough that they shouldn't need to hit rest more than once.

Of course, the "you feel sick" messages shouldn't break the 20-turn "resting chain" because that'd just be obnoxious.

[roctavian]

Can we continue discussing deterministic poisoning? It seems like an excellent solution. What disadvantages would there be to implementing it (besides the Broguification of Crawl)?

(I don't actually think that making Crawl more like Brogue is a bad thing).

[Kalma]

Does poison status damage need to be able to kill outright? Maybe it could be proportional to fraction of max health - sort of a slow torment mechanic. Then you could make the damage (at high health) bigger so it becomes more relevant in battle.

[shock]

Tedium because of increased resting (which even need some special configuration to avoid long rest being constantly interrupted).

Galehar's suggestion to shorten the duration would solve this, although it may reduce differentiation of the poison damage type. Consider even a drastic reduction, where all damage is spread over 3-5 turns. This may be too similar to sticky flame. So far, other suggestions resolve poison status by something other than time have flavor issues.

Variance of duration and damage. It's hard to predict how much damage you're going to take or how long you'll be poisoned.

Pubby's suggestion to make poison damage + regeneration deterministic and display potential damage in health bar is an elegant solution to this problem.

Rarely tactically relevant. Because the damage is spread over a long time, the amount you take during the fight is rarely relevant.

Galehar's suggestion to shorten duration while maintaining total damage would make it tactically relevant in the sense that you're taking more damage per turn. Tactical decisions like "drink curing now vs. deal with enemy first" would become more meaningful.

Another approach would be to give poison temporary and tactically relevant non-damage effects. Effects like stat damage, MP loss, slow, confuse, pass out+exhaustion, and paralyze could all fit from a flavor perspective. Poison could even function like contamination/glow where poison level is tied to severity of probabilistic effects.

[Steel Neuron]

Just addressing the flavor issues:

Why not, if you manage to rest uninterrupted for (arbitrary but low number of turns) you get a "you suck the poison out" message and remove it?

[XuaXua]

The potion-making spell that Crawl used to have would have been great to use here to cast and create an anti-venom from a corpse.

[Mankeli]

Understand that flavour is not just about aesthetics. If something is logical and make sense, it's much easier to explain it to the player. With your proposal, people will ask why they are suddenly cured of poison after killing a snake.

Well, I did try to explain why I feel actually getting the antidote does make at least some sense flavourwise and I also feel that this flavor-case is blown out of proportion in this particular instance. So in the part of my post that you deliberately failed to quote, I did try to provide flavor for the people that do care about that kind of things but you either 1) didn't and still don't understand the rationale behind producing antidotes and hence fail to realize why I considered my suggestion be to sufficient or 2) simply think that was not good enough in which case I recommend reading the two paragraphs below.

I see no further problem with also misidentifying them as producing antivenom, and "8 cells away" is a pretty ridiculous nitpick when you are also e.g. notified when something dies on the other side of the level; these things are abstracted away all the time (you aren't shown actually taking your gloves off when you put on a ring, or actually unwielding your great mace to butcher a corpse).

Yes and yes. Nowdays you can also train skills you are not actually using (like training a weapon skill when killing monsters with spells or with a weapon belonging in other weapon class). I doesn't make any sense flavorwise that you can learn how to use a sword simply by keeping it in your backpack or holding it stationary in your hand without actually hitting anything with it but it makes much sense gameplaywise.

That said this proposal doesn't address the annoyance of waiting out poison when you don't kill the source, so I prefer making poison faster, though you'd need to divide its duration by way more than 2.
Or maybe make it go away if you don't suffer any non-poison damage for 5 consecutive turns. That would have no effect on genuinely dangerous fights but would make resting much less of a pain.

Yeah, this very true and it's a real problem for my suggestion (and I acknowledged it myself too). I have no idea how common the situations are where you don't kill the poisoner compared to the cases you do. I'm completely fine with (much) faster poison too if people feel its better. Or you could combine the two ides maybe so that it would always be faster than now but you could end it instantaniously when you kill the poisoner. Or maybe the devs come up with another idea better than any of these.

[Tiktacy]

I don't know about everyone else, but it feels like after early game, poison is pretty much non-existent. Maybe we should make It require multiple points of resistance like fire and ice? As is, improving something that is rarely seen beyond the early games sounds strange(but still makes sense none-the-less).

[galehar]

Can we continue discussing deterministic poisoning? It seems like an excellent solution. What disadvantages would there be to implementing it (besides the Broguification of Crawl)?

Yeah, I think we should start with that. Then, it becomes much easier to understand and balance.

So in the part of my post that you deliberately failed to quote, I did try to provide flavor for the people that do care about that kind of things but you either 1) didn't and still don't understand the rationale behind producing antidotes and hence fail to realize why I considered my suggestion be to sufficient or 2) simply think that was not good enough in which case I recommend reading the two paragraphs below.

I didn't quote all your proposal but I read it. And I find it dumb, I thought I made that clear. It's not enough to explain it on the forum, you have to think how you're going to explain it in-game.
Minmay's argument is that other features are abstracted, so it's fine to ignore flavour. I disagree. As I've already tell, when it's the only way to improve gameplay, I'm fine with dismissing flavour. But it's not the case here. We've only just started discussing it so I'm not willing to go straight to a solution with convoluted flavour, especially since its gameplay consequences aren't that great.

As is, improving something that is rarely seen beyond the early games sounds strange(but still makes sense none-the-less).

Why shouldn't we improve the early game? Also, early game usually means pre-Lair.

[dpeg]

As always, the Brogue way is interesting. It's approach already been mentioned upthread: you accumulate deterministic poison damage, how much HP you're going to lose to poison is indicated as a bar, and as soon as your current HP vs poison damage is fatal, that's also displayed. However, there are situations against spiders (Brogue's notorious early poisoning monster) which feel desperate: curing is rare, and if you don't react quickly enough, you may be able to kill the spider but then you watch helplessly as the poison slowly kills you. Mind you, almost always this is still operator error.

Here is an idea how of what we could do in Crawl:

• Poison damage is deterministic: if you get hit by a poison attack, that means a certain amount of HP. Modifiers like poison level and rPois apply, of course.
• Incoming damage of poison type is applied normally but also tracked separately (can use the ordinary HP bar for this). If you would in principle be able to heal if off naturally (assume that no monsters are around, so it's just your current HP plus your regeneration vs the remaining poison HP reduction), then nothing further happens. As soon as you had no incoming damage for sufficiently long (say 30 turns), the poison accumulation is cleared.
• If not, i.e. too much remaining poison damage incoming, then you don't die, but excess damage is taken off your maximal HP (rot style). There might be some conversion factor necessary, e.g. 5 excess poison damage = 1 rotted maxHP.

Bottom line: poison will rarely kill you directly, but using !curing (or !heal wounds) during combat is often better than waiting out. Strong poisoners can be costly, so you may want to adapt a bit. Note that there will be no 5-mashing whatsoever. So poison damage is just a more dangerous damage type, like rotting or draining. You can play around a bit (ignore and estimate you'll be able to not-hit the threshold) and you can use resources in battle to clear it.

[roctavian]

However, there are situations against spiders (Brogue's notorious early poisoning monster) which feel desperate: curing is rare, and if you don't react quickly enough, you may be able to kill the spider but then you watch helplessly as the poison slowly kills you.

I actually think this is a great advantage of the Brogue mechanic -- it's heartbreaking when this happens to a promising character, in a way that simply getting killed can't quite match. On the other hand, it could lead to more Q->Yes deaths, which I don't think is desirable.

[*] (if there is) too much remaining poison damage incoming, then you don't die, but excess damage is taken off your maximal HP (rot style). There might be some conversion factor necessary, e.g. 5 excess poison damage = 1 rotted maxHP.

I think I got this, but for clarification: supposing I have 50 hitpoints, and I am poisoned 53 points. I lose my hitpoints down to 1/50, then I lose a max HP and have 1/49 (50)?

Supposing I have 15 hitpoints, I get poisoned for 30 points. I lose my hitpoints down to 1/15, then I drop over the next few turns to have only 1/12 (15)?

I would suggest that maxHP lost to poison in this way should recover more quickly than Rot, and should be completely restored with a single !curing after the fact, so that Rot remains more threatening.

[Lasty]

Of the ideas being thrown around, I like dpeg's best, though there's still a bit too much regeneration versus rate of poison damage in it for my taste. So, let me throw out one other idea:

Poison should be a threat in combat, and it should be a threat while running or evading, but it should quickly evaporate when the player has time to breathe. So, every turn you spend resting knocks off a unit of poison. You can do this no matter what else is going on. This allows you to quickly terminate poison status when you've safely resolved an encounter, but makes poison still a risk when you're forced to take actions to fight or escape. It also allows the possibility of spending time during a fight to rest off some poison before continuing to fight back, though presumably that's a risky maneuver.

If this would recover from poison too fast due to the exact quantity of poison units as things currently work, then multiply the unit scale. It would probably also make sense to increase the amount of damage dealt by poison over time, since this would greatly speed up poison recovery for players that are out of danger.

Things that could be tweaked in this proposal:
* Whether resting ameliorates poison when the player is adjacent to monsters.
* Whether resting ameliorates poison if the player takes damage (or is attacked?) during the 10 AUT of resting time.
* How much each resting action reduces poison units.
* How much damage poison does per time by number of poison units.

[TwilightPhoenix]

The one nifty thing about deterministic and visible-ahead-of-time poison damage is there's definitely a certain horror in seeing all of your HP bar turn purple and knowing, unquestionably, that you will die if you don't act. I like it.

[Sandman25]

Perhaps poison should decrease movement speed, then it will be different enough from both rot and direct damage.