[non-variant-specific] Items with Drawbacks

[Djabanete]

Time for a stupid idea. What would you do if you found this item, with whatever character you're playing right now (any variant):

Nuke-Resistant Armor [60, +40] (-4 to speed)
[super heavy, like 50 lb]
It provides resistance to fire, acid, lightning, sound, and shards. It decreases your speed and dexterity by 4. It cannot be harmed by the elements.

Intriguing, yes... but would it be worth using? Well, why not take it for a spin and see how it goes? Here's another lame example:

Ring of Undeadness
It provides immunity to nether. It attracts undead creatures.

That one's a little different. Maybe worth keeping around just for when you're fighting Azriel? How frequently do the undead appear? How bad would they even be if you're immune to nether? Maybe it's worth keeping around? No? As a swap?

To me, that kind of item is the most interesting kind, and I don't see enough of them. It's not that they're not out there; it's that when you find them, you've always got better options. That's no fun. Items with drawbacks create interesting equipment decisions, and sometimes they create interesting game situations. The perfect "grey item" (as I like to call them) is Calris. It has a lot of upside, but also a very significant downside which must be weighed. An example of a non-grey item would be the One Ring, which is so ridiculously powerful in every respect that you don't care if you aggravate and you don't care that you can't take it off. The Palantir of Westernesse lies somewhere in between those.

I'd like to see more things along the lines of Calris. Just as an experiment, I'm going to look at a couple pages of winners on the Angband ladder --- not the top score guys since they all have perfect kits, maybe one page in. (*looks*). 0K --- The One Ring and the Palantir (very rarely Calris) are the *only* items with any sort of drawback that winners use. Let's look at some mid-level characters. (*looks*). I didn't see any grey items at all.

Well, I just wish the grey items were better, that's all. That there were some items with *real* drawbacks that aren't just cursed crap. That there were items that create tough choices - and I'm not talking about the "Hmmm, Ringil or Deathwreaker?" kind of tough choice.

Anyway, sorry for the disorganized, late-night post. This just occurred to me and I wanted to share the thought and see what you guys have to say.

-Djabanete

[Magnate]

To me, that kind of item is the most interesting kind, and I don't see enough of them. It's not that they're not out there; it's that when you find them, you've always got better options. That's no fun. Items with drawbacks create interesting equipment decisions, and sometimes they create interesting game situations. The perfect "grey item" (as I like to call them) is Calris. It has a lot of upside, but also a very significant downside which must be weighed. An example of a non-grey item would be the One Ring, which is so ridiculously powerful in every respect that you don't care if you aggravate and you don't care that you can't take it off. The Palantir of Westernesse lies somewhere in between those.

I'd like to see more things along the lines of Calris. Just as an experiment, I'm going to look at a couple pages of winners on the Angband ladder --- not the top score guys since they all have perfect kits, maybe one page in. (*looks*). 0K --- The One Ring and the Palantir (very rarely Calris) are the *only* items with any sort of drawback that winners use. Let's look at some mid-level characters. (*looks*). I didn't see any grey items at all.

Well, I just wish the grey items were better, that's all. That there were some items with *real* drawbacks that aren't just cursed crap. That there were items that create tough choices - and I'm not talking about the "Hmmm, Ringil or Deathwreaker?" kind of tough choice.

Anyway, sorry for the disorganized, late-night post. This just occurred to me and I wanted to share the thought and see what you guys have to say.

Well said. I agree completely. This is one reason I never play without randarts - too many of the standard artifacts are no-brainers (Thorin, Dor-Lomin, Fingolfin etc.). When Chris Robertson & I re-wrote the randart.c code for V (back in 2.9.x - it never made it into V though it has been adapted for use in NPP, Un and others), one of our primary goals was to make "grey" randarts more viable. Unfortunately we were constrained to a small range of drawbacks (negative pval, minuses to-hit, to-dam or AC, light or heavy curse, aggravation), but that still allows a much broader range of items than the old randart code (which simply gives all powerful items aggravation). Adding some other drawbacks would make life much more interesting.

I suspect that someone will come along in a minute and say that this has all been done brilliantly if only we would try Un ... I will one day, honest ...

CC

[Nick]

Adding some other drawbacks would make life much more interesting.

Next version of FA has percentage resistances - and vulnerabilities. Not many grey items yet, but it's certainly in the plan.

[Mondkalb]

As for armor I think it is too easy to get rbase covered. Finally any item of elvenkind (or Resistance) will show up and that's it - for a while.
The player isn't forced to wear any too heavy armor usually (there are some rare moments I remember walking around in rusty elvenkindish things with -4 or such and also filthy rags of elvenkind with poor base stats).

How about making shield/robe/armor of elvenkind only cover up random three of rbase + one high extra? This would lead to some change at least for midgame.
The good high resistances like rPoison, rConfusion for example could be bound to only heavy armor or they would be only granted if the item had only two of rbase covered.

Similar things could be done with weapons - telepathy could always cause aggravation. *Slay* weapons could aggravate the undead.

[Bandobras]

Speaking for low-levels, I like the Shooters of Nazgul (whatever variant that was, I guess O) that grant SI at a cost. They are really nice for humans, especially if they could be auto-identified somewhat (ID for each {terrible} weapon is a bit too expensive for a starting char).

I think similar things should be done with FA (some big drawback too), and other vital abilities, so that players have the option to reasonably dive with some interesting drawbacks instead of scumming for the ability or risking madly. I think most of the cursed items in the game could be made more interesting by throwing in some big drawbacks and small but interesting bonuses, thus increasing interesting strategic trade-offs. As the game progresses the grey items could be exchanged with strictly positive items (except some really wonderful grey artifacts).

[Djabanete]

Adding some other drawbacks would make life much more interesting.

Yes, Definitely. Some drawbacks that I haven't seen in V:

Randomly summon certain kinds of monster (undead, animals, etc)
Be much heavier than a typical specimen
Require you to carry stuff around for some reason (like flasks of oil; why no artifact lantern?)
Reduce your light radius

I'm trying to think of stuff that's not too far-fetched for V. Bright ideas?

@Mondkalb: Agree on the Elvenkind armors, although it's already plenty annoying to have to *ID* them all. This would increase that necessity because you need to know your resistance holes! But I think *ID* should be a lot more commonly available anyway.

@Bandobras: Yes, those Nazgul weapons are exactly the kind of thing I'm talking about! You reminded me - my favorite Angband weapon was a Great Axe of Morgul that I enchanted up from (-17, -17) to (+8, +8); suddenly I had a weapon with See Invisible, Slay Undead, and Poison Brand, at the cost of Aggravation. I died soon after, but it was awesome

[andrewdoull]

I suspect that someone will come along in a minute and say that this has all been done brilliantly if only we would try Un ... I will one day, honest ...

CC

I've started it, but the approach taken by yourselves (and me) is I think the wrong one, if I may be so bold.

Currently, we evaluate a single artifact power level, which is the positive bonuses, minus the negatives.

We should be generating the negatives first, and heavily cursed items then allowed to rise above the current depth's maximum power level. This would make e.g. weapons of Morgul very tempting as they'd be 'ahead' of the power curve.

Andrew

[Magnate]

I've started it, but the approach taken by yourselves (and me) is I think the wrong one, if I may be so bold.

Currently, we evaluate a single artifact power level, which is the positive bonuses, minus the negatives.

We should be generating the negatives first, and heavily cursed items then allowed to rise above the current depth's maximum power level. This would make e.g. weapons of Morgul very tempting as they'd be 'ahead' of the power curve.

I would go even further, and suggest that a fresh approach to evaluating artifact "quality" (carefully avoiding "power") would be to quantify more than one distinct total. So instead of getting into the bizarre minutiae of "is rchaos worth more than +3 STR? What about +4 STR? What about +6 Speed? What about ESP?" etc. etc., you can have a more sophisticated approach. I thought about this last time round, but didn't get anywhere with it because I'm not clever and/or persistent enough. I'm probably still not, but the discussion of "grey" items, and their potential for improving the game, has prompted me to contribute what I can.

The first quantity to measure is the artifact's effects on your six main stats. This is probably the simplest, because they're all numbers already. There is some debate to be had about whether +4 STR is worth twice as much as +2, or four times, or whatever, but in the end you can come up with an algorithm which measures the item's contribution to improving your stats.

The side issue here is the "spell stat". Personally I favour rating items according to their value to the current character (ie. WoG is {uber} for a priest, while Raal's is worthless), though I know this is a big departure from current item generation. With this idea of a more sophisticated rating system for randarts, I think there is mileage in saying that the current char's mana stat is worth more than the non-mana stat.

The second thing to be quantified is the item's contribution to your offensive power: to-hit, to-dam, slays, brands, extra shots/might/blows. This is the bit that Chris Robertson did really well, with the monster rating code and weighting of brands/slays according to how many monsters resist them. The only enhancement I wanted to add was a more detailed assessment of the value of +to-hit, taking monster AC and criticals into account - this is not really significant in V but would be very important in O-combat variants.

Anyway, the third thing to be quantified is the item's contribution to your defensive capacity: +AC, resists, immunities, FA, hold life, regen, feather fall, bonus to saving throw in variants which allow that. You could arguably dump all the other possible attributes in here, it all really depends on how detailed you want to get.

Then there are super useful things like +speed, +stealth, SI and ESP, which really deserve their own consideration. There are also a bunch of not so useful things like +searching, slow digestion, +tunnelling, +infravision, plite. If you wanted you could group these into detection-related (SI, ESP, searching, infravision, light) and then just have speed, stealth, tunnelling and slow digestion as separate, but again it depends on your penchant for detail.

Finally there is the activation, if any. So you have at least five distinct contributions to an artifact's value, maybe more (we haven't touched on item weight, for example). Sure, you can just add them all up and end up with a power rating not dissimilar to the current code, but you could do more than that. You could have separate depth distribution for each element: certain totals or combinations of stat enhancements could be restricted to one set of depths, while combinations of offensive or defensive enhancements could be found at different depths. You could develop Chris R's idea of "supercharging" and create items which have only offensive capabilities, or only defensive, etc.

So what does all this have to do with grey items? Well, my thinking is that if you have a more granular system for assessing and creating items, you can treat drawbacks in the same way. Multiple pvals (as used in S) would be a huge help, because at the moment in V you must have ALL pval-related stuff attached to the single pval (stats, speed, stealth, infravision, searching, tunnelling, shots, might, blows). To do grey items properly you need to be able to create items which boost one stat but lower another, boost speed but lower stealth, etc. etc.

Even without multiple pvals you could concentrate the drawbacks in a particular area (lots of offensive penalties on an item with a lot of defensive enhancements, for example). Another thing that would broaden the scope is the addition of vulnerabilities (negative resists) - at the moment the only drawbacks you can have are negative to-hit/dam/AC/pval, curses and aggravation (ooh, and xp drain and random teleportation, I forgot). Introducing vulnerabilities, or something like Z's anti-magic or anti-teleportation would give more combinations (what about a mind***k attribute that cancelled ESP? Fine if you don't have ESP yet, but very painful if you do ...).

I'd better get on with some work. Sorry this hasn't been much other than a bunch of undeveloped ideas, but maybe I'll do something with it one day. None of it was intended to disagree with the main point - I agree that grey items being generated 'ahead' of the power curve is a good thing.

CC

[Djabanete]

We should be generating the negatives first, and heavily cursed items then allowed to rise above the current depth's maximum power level. This would make e.g. weapons of Morgul very tempting as they'd be 'ahead' of the power curve.

Do you think you could elaborate a little on this? I don't really understand what you're suggesting. Are you thinking of applying this to ego items? Randarts? How is generating negatives and then adding positives different from generating positives and then adding negatives? Sorry, I'm probably totally missing the point here. =p

@Magnate: Yeah, I think a smart randart generator would do well to consider those various aspects so it can make artifacts more intelligently. Then you can produce really interesting grey item randarts (e.g. an item that enhances senses but reduces defense ---> SI, ESP, light radius, but AC penalty; that would be pretty clever for a randart, and I think that kind of thing would be simple to generate).

A similar but different way to come up with grey randarts would be to draw positive *and* negative correlations between all the possible attributes; and then to produce one "starting" attribute and go from there. So you could link "+stealth" to "combat penalty", "ESP" to "aggravate"; and each link has a chance of being applied to the generated randart. As you got deeper, the links would stay the same, but the numbers would get bigger; so the generated randarts would get even more grey (big +stealth, big -combat, for example).

Anyway, I'm not a coder, and I have no idea if that's practical or even original, but the concept seems pretty simple and effective.

[Nick]

A similar but different way to come up with grey randarts would be to draw positive *and* negative correlations between all the possible attributes; and then to produce one "starting" attribute and go from there. So you could link "+stealth" to "combat penalty", "ESP" to "aggravate"; and each link has a chance of being applied to the generated randart. As you got deeper, the links would stay the same, but the numbers would get bigger; so the generated randarts would get even more grey (big +stealth, big -combat, for example).

I like this idea, and intend to steal it

[Djabanete]

I'm flattered. =D

[Magnate]

[FONT="Courier New"]Do you think you could elaborate a little on this? I don't really understand what you're suggesting. Are you thinking of applying this to ego items? Randarts? How is generating negatives and then adding positives different from generating positives and then adding negatives? Sorry, I'm probably totally missing the point here.

If I understood him correctly it's not the order of application that's important, it's the fact that the positives are larger than they would otherwise be at that depth (because of the negatives). Hence you get more powerful items at a given depth, but with drawbacks.

A similar but different way to come up with grey randarts would be to draw positive *and* negative correlations between all the possible attributes; and then to produce one "starting" attribute and go from there. So you could link "+stealth" to "combat penalty", "ESP" to "aggravate"; and each link has a chance of being applied to the generated randart. As you got deeper, the links would stay the same, but the numbers would get bigger; so the generated randarts would get even more grey (big +stealth, big -combat, for example).

As Nick said, this is a good idea - one more reason to try FA ... the thing I found difficult about a randart generator was making any kind of smooth progression with depth. As you go deeper, items should generally get a little better - still with drawbacks, maybe, but overall slightly better. So in your plan the positives should get larger slightly quicker than the negatives. Maybe some of the positive links could come into effect only after certain depths, giving a small boost without affecting drawbacks. Also, you could have deeper and more powerful items starting with two or three basic mods, rather than one. Of course, if all the links come up negative, you'll have two or three drawbacks as well, making it a very grey item ...

I also think it could be tricky to balance the correlation becoming tedious (ie. anything which boosts stealth ALWAYS gives combat penalties) with not really being a correlation at all. I had one idea on this one, which is to structure the links in tiers. So while there is always a chance that +stealth items will penalise combat, often they won't. If they don't, they might compromise your detection instead, or your AC, or whatever. That would provide more variety and help avoid tedium without spoiling the feeling of correlation (or "theme" as we called it).

CC

[Djabanete]

So there's the possibility of instant run-off, allowing any particular drawback (or attribute) to be less commonly associated with another. That seems like it would work really well.

As for getting deeper, I was kind of thinking that the drawbacks would increase at more or less the same rate as the positives; otherwise you end up with "Ringil that gives -3 to tunneling" or something. If a grey item is interesting at all, it's because it enhances the things you care about and nerfs the things you care less about; so an item that REALLY enhances what you care about and REALLY nerfs what you don't care about, is still more desirable than an item that only provides slight enhancements/nerfs. Thus it would be found correspondingly deeper... if that makes any sense.

EDIT: Although it should definitely be true that at low depths, grey items should have more positives than negatives, or else they're pretty much junk. So in that sense I agree with you. What I was thinking was that if a deep grey item decreases strength, it should _decrease_ your strength (like -4 or -5), otherwise a high-level character doesn't care at all.

And yeah, now I see what Andrew Doull was saying about first generating drawbacks. Makes sense to me, although you don't want to have to *ID* every single {worthless} item you find, so there should be some sort of rhyme or reason to it.

[Magnate]

[FONT="Courier New"]As for getting deeper, I was kind of thinking that the drawbacks would increase at more or less the same rate as the positives; otherwise you end up with "Ringil that gives -3 to tunneling" or something. If a grey item is interesting at all, it's because it enhances the things you care about and nerfs the things you care less about; so an item that REALLY enhances what you care about and REALLY nerfs what you don't care about, is still more desirable than an item that only provides slight enhancements/nerfs. Thus it would be found correspondingly deeper... if that makes any sense.

I take your point about high-level characters not noticing minor drawbacks, but I think they also have higher requirements of their deep items (i.e. needing to cover more than one resist etc.), which is why I meant the positives should be better on deeper items. I'll try and clarify what I mean with an example:

At 500' an item that gives +3 stealth and (-3, -3) to combat is an interesting grey item.

At 2500' an item that gives +6 stealth and (-6, -6) to combat is an interesting grey item, but it's likely to lose out to something that gives, say, rblind and rnexus instead of the +6 stealth.

Plus the simple fact that the deeper you go the more items you have to choose between, meaning that grey items will have much stiffer competition, so the drawbacks mustn't be too severe ... like all design decisions, it's a careful balancing act.

And yeah, now I see what Andrew Doull was saying about first generating drawbacks. Makes sense to me, although you don't want to have to *ID* every single {worthless} item you find, so there should be some sort of rhyme or reason to it.

Hmmm. I think this whole discussion has quite significant implications for pseudoID. I don't think any grey item should be {worthless}. In fact the whole point about generating them is to replace the no-brainer cursed items that you just ID and destroy. So pseudo would need reworking along the lines of

{average} means non-magical
{magical} means one or two mods (which may or may not be drawbacks)
{arcane} means three to five mods (ditto)
{uber} means six or more mods

(You could still have {cursed} if one of the mods is the CURSE flag.)

CC

[andrewdoull]

I take your point about high-level characters not noticing minor drawbacks, but I think they also have higher requirements of their deep items (i.e. needing to cover more than one resist etc.), which is why I meant the positives should be better on deeper items. I'll try and clarify what I mean with an example:

At 500' an item that gives +3 stealth and (-3, -3) to combat is an interesting grey item.

At 2500' an item that gives +6 stealth and (-6, -6) to combat is an interesting grey item, but it's likely to lose out to something that gives, say, rblind and rnexus instead of the +6 stealth.

Plus the simple fact that the deeper you go the more items you have to choose between, meaning that grey items will have much stiffer competition, so the drawbacks mustn't be too severe ... like all design decisions, it's a careful balancing act.

That's basically what I was suggesting, but put a lot more clearly. You'd want to generate the negative effects of the item, and boost these up to the allowed 'object_level' so that you'd find progressively worse and worse negative effects as you got deeper.

Then you'd take the item and add its positive effects up to the allowed object_level + the value of the negative effects. e.g. a heavily cursed item may be allowed a power of 20 at level 10, whereas a normal magical item could only have a power of 10.

Hmmm. I think this whole discussion has quite significant implications for pseudoID. I don't think any grey item should be {worthless}. In fact the whole point about generating them is to replace the no-brainer cursed items that you just ID and destroy. So pseudo would need reworking along the lines of

{average} means non-magical
{magical} means one or two mods (which may or may not be drawbacks)
{arcane} means three to five mods (ditto)
{uber} means six or more mods

(You could still have {cursed} if one of the mods is the CURSE flag.)

CC

Why not just generate all cursed items/artifacts this way? This would make every cursed item worth investigating, which would potentially increase the total pool of useful items. All cursed weapons armour etc would be the 'grey' objects that you're talking about.

e.g. you might want to use gauntlets (+20, -20) {cursed} if you're an ranger or (-20, +20) {cursed} if you're a warrior.

The curse / heavy curse / perma curse level would be evaluated based on the total curse level of the item rather than being randomly selected. This would allow you to find truly powerful perma cursed items that might be worth using.

Of course, allowing multiple pvals would improve this significantly.

Andrew