Contaminate artifact property

[Lasty]

I just noticed a post about artifacts with the contaminate property, and I thought that I had some better ideas than the present implementation.

It seems to me that the property works best if the following things are true:

1) using contaminate artifacts should be a choice, and the choice should be one with possible drawbacks.
2) Stopping use of the artifact should not instantly prevent all consequences of use.

To that end, let me propose something that's in between the old, terrible version and the new, harmless/terrible binary version: When you are wearing an artifact w/ this property, you gain 1 contamination every 1000 or so turns if you don't already have contamination. If you do already have contamination, when that contamination would decay it instead has a (number of contaminated artifacts / 5) chance of instead increasing by 1. Taking off the contaminated artifact multiplies your level of contamination by 2.

As a result, wearing it for an extended period would probably eventually slide you into yellow glow sometimes, particularly if you already do a lot of things that give light contamination now and then (spell failure, certain charms, potions of resistance), but would mostly be fine. If you try to manage the contamination too much by removing the artifact after a small amount of contamination lasts longer than expected (a possible sign that it's climbing upwards), then you run the risk of doubling it into yellow glow territory, prompting people to avoid micromanaging contamination by wearing/removing the item frequently.

Thoughts?

[BlackSheep]

With the numbers you have now, your contamination would likely just decay normally. If you increase the amount of contamination, we're back to where we started only worse: Contam artifacts generate glow over time and if you take one off, you're doubly screwed.

[Psieye]

I don't see the problem with the current implementation of Contam items. It's essentially an alternate form of 'cursing equipment', except you don't get rid of it with a consumable. Is there a need to add more drawbacks?

[Lasty]

With the numbers you have now, your contamination would likely just decay normally. If you increase the amount of contamination, we're back to where we started only worse: Contam artifacts generate glow over time and if you take one off, you're doubly screwed.

The idea was to pick a number where it's possible but relatively unlikely to get yellow glow off a given instance of contamination; it becomes a possibility over the long term, but not a guarantee that renders the item unusable. Further, it gets more likely if the character is getting extra contamination, meaning that the user's behavior actually has influence on the outcome.

I don't see the problem with the current implementation of Contam items. It's essentially an alternate form of 'cursing equipment', except you don't get rid of it with a consumable. Is there a need to add more drawbacks?

The problem with the existing implementation is that either there is no drawback (in the case where you want to keep the item on at all times) or the drawback makes the item completely useless (in the case where you may need to switch it out ever). That makes it a much less interesting property.

[BlackSheep]

The old implementation gave you glow over time. This made items with Contam either completely useless, or you could get around it by swapping frequently. Your implementation just sounds like taking the worst of both implementations. You'd get contaminated over time and get hit with more glow when you unequip. You can tweak the numbers, but you will either guarantee glow over time or will guarantee that you can rest it off, meaning we'd either be back where we started or Contam would be a "don't caste haste/invis and you'll be fine" property.

[Psieye]

You get 7 glow for taking off a Contam artifact. That makes it likely you take 1 bad mutation attempt, with a 50% chance to block it with rMut and 100% chance to make it more manageable with being/becoming undead. A cost that needs consideration but not something that makes the item 'completely useless'.

[crate]

not sure why you're suggesting reverting to the old contam except worse....

[Lasty]

The old implementation gave you glow over time. This made items with Contam either completely useless, or you could get around it by swapping frequently. Your implementation just sounds like taking the worst of both implementations. You'd get contaminated over time and get hit with more glow when you unequip. You can tweak the numbers, but you will either guarantee glow over time or will guarantee that you can rest it off, meaning we'd either be back where we started or Contam would be a "don't caste haste/invis and you'll be fine" property.

The old implementation stacked contamination so fast that you'd always get glow if you wore it for a while. My proposal is designed to make sure that it is unlikely to give you glow over a given period of time, but any given instance of light contamination might conceivably get amplified, making it so that if you wear it for a long time the law of large numbers makes it likely that you'll probably get glow for a while eventually, and that chance goes up if you do things that increase your contamination amount frequently.

I'm not sure why you see that as being the same as "always get glow" or "never get glow", so I disagree with your analysis, but if you're right and it is "don't caste haste/invis and you'll be fine" it would still be more interesting than the current implementation.

[Lasty]

not sure why you're suggesting reverting to the old contam except worse....

I'm really surprised that everyone seems to be taking this interpretation. My proposal is not at all like the old one. My proposal makes this property give contamination a chance to escalate slowly, whereas the old version simply caused contamination constantly.

[BlackSheep]

My proposal is designed to make sure that it is unlikely to give you glow over a given period of time, but any given instance of light contamination might conceivably get amplified, making it so that if you wear it for a long time the law of large numbers makes it likely that you'll probably get glow for a while eventually, and that chance goes up if you do things that increase your contamination amount frequently.

And the obvious response to that is to hit 5 any time you have any amount of contamination, and to take off the item if you want to cast haste.

[crate]

oh nevermind with the numbers you are suggesting it is actually effectively the same as newcontam except you can take it off by pressing 5 a lot

[Lasty]

Pressing 5 a lot is not a good response, because waiting won't necessarily clear the contamination, and if waiting would clear it then behaving normally (aside from casting Haste/CBlink) will also clear it.

Taking it off becomes free as long as you aren't already carrying at least 3 contamination, meaning that you can freely take it off and put it on except in the case where you're already fairly contaminated.

[crate]

Contam clears at 1 point per 40 turns on average iirc.

You are suggesting adding 1/1000.

Compare those two numbers :p

[Lasty]

Contam clears at 1 point per 40 turns on average iirc.

You are suggesting adding 1/1000.

Compare those two numbers :p

Yes, I understand. The more important part of the proposal is that when you would naturally clear contamination by timeout, you have a 1 in 5 (or so) chance to instead add 1 contamination. So, once you're contaminated, your contamination value may wander up and down a little before clearing. If it wanders high enough (which should be relatively infrequent), you will get some yellow glow.

The +1 contam per 1000 turns or so is to keep it doing something for characters who don't normally pick up contamination.

[BlackSheep]

The point I keep hammering on is that a 20% chance of gaining 1 contam instead of losing 1 is meaningless in the long term. If you rest enough you'll eventually lose your glow unless you get very unlucky.

[Lasty]

The point I keep hammering on is that a 20% chance of gaining 1 contam instead of losing 1 is meaningless in the long term. If you rest enough you'll eventually lose your glow unless you get very unlucky.

Right, if you do anything enough (rest or not) you'll eventually lose your glow -- in most cases. It only has to increase instead of decreasing 5 times to get to yellow glow, and while that's not particularly likely for any given instance of contamination, over enough instances of contamination it'll probably happen once. Thus, there's a reasonable risk associated with leaving contam items equipped over a long time frame, instead of either no risk at all over any time frame (current version) or an unacceptable risk over even short time frames (old version).

Anyway, now I feel confident that you understand what I'm proposing, so I'll stop trying to explain it. If you still don't like it, that's fine.

[KoboldLord]

It only has to increase instead of decreasing 5 times to get to yellow glow, and while that's not particularly likely for any given instance of contamination, over enough instances of contamination it'll probably happen once.

Uh, you realize those odds are easy to calculate, and they are 1 in 3,125? How many instances of glow do you get in a typical game?

[Lasty]

Uh, you realize those odds are easy to calculate, and they are 1 in 3,125? How many instances of glow do you get in a typical game?

The odds aren't just 1/5 to the 5th, because if you make it to 2 contam, your next check will have a 4/5 chance to take you to 1 contam, not to 0, meaning that you still have a chance to climb back to 2, and from 2 potentially up further.

That said, I do realize that the odds are small, which is one of the reasons I believe it should also give you 1 level of contam periodically so that you are entered into that lottery more often. If the numbers 1/5 chance and 1000 turns aren't the exact right ones, that's fine -- they can be tweaked.

[Galefury]

Uh, you realize those odds are easy to calculate, and they are 1 in 3,125? How many instances of glow do you get in a typical game?

Every miscast, every cast of haste, plus every cast of cblink or invis, plus every controlled teleport, plus zot traps, and the new 1/1000 turns chance. This adds up to a lot of instances of glow. Many of these also give more than one point of glow, so the chance to get to yellow (or stay at yellow glow for much longer than usual) would be much higher.

I don't think it's a good idea, but this version of contam would probably be relevant for most chars who cast spells.

[KoboldLord]

I use Haste when I have it, but I don't use it three thousand times during the course of a game. That's, like, an individual Haste for every skeletal warrior in Crypt. I don't even always find Haste until I'm ready to collect my Runes and get out, and what's more I'd have to find it at the same time as I found an artefact with a particular uncommon negative property that is also worth using in spite of that negative property. Old contam and new contam were both, at the very least, not something you could nearly ignore.

And then we can thin out the interesting even more by the fact that plenty of instances of yellow glow end up going away without anything more than a trivial amount of damage, and then thin out the interesting even more by the fact that many of the malmutations you can pick up this way aren't even bad enough to qualify as an inconvenience.

The proposed contam is mostly a call for mashing 5 a lot whenever you get contamination so it goes away. Once or twice per game contamination might actually spike up randomly for no good reason with no player action, but a couple malmutations isn't usually all that bad anyway. Even if the player draws one of the bad ones, potions still exist in the game and if Jiyva doesn't that's the player's own fault. Do players really need to mash 5 an extra time the 20% of the time the contamination decay counter randomly decides to increment instead of decrement? Is mashing 5 an extra time 20% of the time really a meaningful drawback for an artefact property?

[njvack]

OK, I just went ahead and simulated this:

  Code:

#!/usr/bin/env ruby

contam_rate=ARGV[0].to_i || 1000
contam_level=0
peak = contam_level
levels = [0, 0, 0, 0, 0, 0]
i = 0
while contam_level < 5 do
  contam_up = (rand(contam_rate) == 0)
  contam_down = (rand(40) == 0)
 
  contam_level += 1 if contam_up
  contam_level -= 1 if contam_down
  contam_level = 0 if contam_level < 0
  levels[contam_level] += 1
  peak = contam_level if contam_level > peak
  puts "#{i} #{contam_level} #{peak} #{levels.join(' ')}"
  i += 1
end


With a 1/1000 chance, this essentially never gets over 3 contam (over the course of millions of iterations), and you spend the vast majority of your time (96%) at 0. With a 1/250 chance, it reaches 4 usually pretty quickly and 5 in maybe 250K-1.5M turns, spending maybe 85% of your time at 0. With a 1/125 chance, you get to 5 in thousands to tens of thousands of turns. So there's some numbers.

Feel free to pick apart my simulation as critically flawed for any number of reasons; I didn't think too hard about it.

I think this generally means "wait off your contamination" or "this will mutate you pretty soon" with this implementation. Maybe there's an interesting magic number, but I suspect not.

Making {contam} just set your contam floor to += 1 in addition to its existing effect might be more interesting and simpler.

[Psieye]

Making {contam} just set your contam floor to += 1 in addition to its existing effect might be more interesting and simpler.

That or just amplify whatever glow you take on, i.e. you'd get half as many cBlinks before yellow glow and if you cast haste/invis you are really likely to hit yellow glow. Though I'm still not convinced Contam needs to be changed again.

[Lasty]

I think the model is wrong in that (if I'm reading it right) there's only a 1 in 1000 chance of increasing contamination instead of decreasing it. Here's a Python implementation:

  Code:

from random import randint
contam = 0
turns = 0
glow_level = 6
contam_increase_time = 1000
contam_timeout = 40
contam_increase_chance = 4
turns_glowing = 0
times_glowing = 0
turns_wearing_contam_item = 50000

while True:
   if turns % 40 == 0 and contam > 0:
      if randint(1, contam_increase_chance) == 1:
         contam += 1
         if contam == glow_level:
            times_glowing += 1
      else:
         contam -= 1
   if turns % contam_increase_time == 0:
      contam += 1
      if contam == glow_level:
         times_glowing += 1
   if contam >= glow_level:
      turns_glowing += 1
   turns += 1
   if times_glowing > 0 and contam == 0:
      break
   
print("Total times glowing: " + str(times_glowing))
print("Total turns spent glowing: " + str(turns_glowing))
print("Turns taken: " + str(turns))

With this implementation, it was fairly random when the first case of yellow glow would occur, but it generally spanned a range from 25,000 turns to 2,500,000 turns. Mostly the glow would last for 40 turns, but sometimes it would last longer -- in some cases the player would even leave the glow state and then return to it before contamination fully subsided.

The numbers could be tweaked to make it more or less common, but I think that's a reasonable spread.

That said, I also like the other ideas: Making {contam} just set your contam floor to += 1 in addition to its existing effect or making {contam} double the amount of contam you receive when you receive contam.

The reason I think that something of the sort is needed is that the current implementation is both boring or binary: either it's worth using, or it isn't. It's basically a curse that you remove with potion of cure mutation or scroll of vulnerability, but it's also a curse that you never enter into except intentionally (unless there are a lot of people wear-iding artifacts for some reason).

[njvack]

The reason I think that something of the sort is needed is that the current implementation is both boring or binary: either it's worth using, or it isn't.

This describes basically every item in Crawl.

This change would make it so {contam} artefacts would either be worth using or they wouldn't. And I think it would add a bunch of resting to try and remove accumulated glow -- because you are *usually* at 0 contam, you can probably cast Haste and wait for the contamination to go away. Because this is stochastic, though, sometimes you'll get glowed for "no reason" which would not be very fun.

I am also not convinced {contam} needs further changes.

[crate]

Helmets are boring. You always want to use them. Remove helmets.

[KoboldLord]

I don't mind mashing 5, but I do find that I do an entirely adequate amount of it and would not like to do more.

That's the worst problem. The other problem, which isn't quite as bad but still not good, is that most turns are spent resting or walking around in relative safety. Accordingly, almost all instances of severe contamination will take place when the player character is under no particular threat anyway, even if magic numbers are found that makes it happen often enough to matter but not so often that the item is de facto unusable. Inflicting a merely annoying or mildly troublesome effect on the player is generally only a good idea in situations that are actually dangerous.

[Lasty]

The reason I think that something of the sort is needed is that the current implementation is both boring or binary: either it's worth using, or it isn't.

This describes basically every item in Crawl.

Fair point; I expressed that badly. What I meant to say is that the property does essentially does nothing, except in the case where the item is only sometimes worth using (weapons/jewelry that will be switched, artifact armour that is good, but not good enough to take through Zot). Within that relatively small design space of items you want to take off, the property changes the item from "sometimes useful" to "never useful".

Given how little this property accomplishes, in most cases it might as well not exist, and when it is meaningful that it exists, it just makes the item unusable.

I think that's worth trying to change, but it sounds like others don't, so I'll drop it.

[crate]

Not being able to swap your weapon or jewellery is pretty significant. Armour obviously isn't a big deal but it's an improvement from the previous case where armour with contam might as well have just not existed because it was entirely useless.

[Lasty]

Right, it's a huge, relevant drawback for the infinitesimal minority of weapons/jewelry that are have this property and are good enough to use but not good enough that they shouldn't be swapped out. For those minority, it basically changes them from "sometimes useful" to "never useful unless you have an abundance of cMut potions, scrolls of vulnerability, or Jiyva".

[crate]

Um, I swap out things that are really good quite often. I don't know about you but I like magical staves, and rods, and rings of protection from fire etc...... Even if I'm wielding a +9 triple sword of speed I will swap to my staff of wizardry an awful lot in a normal game. You can't do that with contam. Similarly I will swap out my AC rings or slaying rings if I need a resistance against whatever I'm fighting. Can't do that with contam.

[yogaFLAME]

Where's ebarret's sig when you need it

[njvack]

aaaaugh, being in that sig is my greatest shame

[Mankeli]

In the crawl world of surprising stepdown functions (dexterity's effect on EV) and mysterious damage formulas even the greatest players dont' always seem to understand (ranged combat damage) do we really need more mechanics that you have to simulate in order to get the slightest idea what it actually does in Crawl?? Not to mention that the devs already implemented a solution that is better than it was before.