Artifact generation - what exactly do you care about?

[fizzix]

I'd say over the past month or so I have not really looked at V and have been focusing energy in getting pyrel dungeon generation working properly. The upside is thanks to many suggestions from other people, the tricky parts are almost all up and running already.

[Magnate]

Off the wall idea... Randarts are generated in advance. Egos are essentially lesser randarts. Why not generate egos in advance (assuming we have the computing power to create and store 1000's of possibilities) and ensure ensure that at least "on paper", that no ego would rival a similar artifact (there will always be situational superiority). In lieu of that, just compare on the fly generated egos to artifacts or randarts of a similar base type and nix the ones that are too powerful (perhaps throw in some sort of DL weight so that deep ego daggers, etc. can exceed the early artifacts).

Begin rant, again (seems like all I do anymore). Escalation of power is a problem, and the removal of cursed items didn't help at all. IMO, decent starting equip should be store bought because everything the the very early dungeon would almost certainly suck, and I mean suck bad, unenchanted maybe, cursed likely, and go from there. The fact that finding a powerful ego in the first 10% of the dungeon isn't uncommon IS a problem and I suspect that the long term solution will be to ramp of the power of everything else to compensate. It's just dev instinct, or so it seems. No offence intended to anyone in particular.

My new near resolutions: Start own variant, kill self. Not to worry though, it's a new year resolution. It's never gonna happen.

Don't worry, no offence taken. Pretty much all of us are done trying to fix V, with the honourable exception of fizzix (and myshkin's noble quest to fix the cocoa port). I'm working exclusively on pyrel, takkaria is completely gone, d_m is around but busy enjoying RL and doing funky stuff like inventing new languages (at least, I think it's something like that - way over my head anyway).

I love how all the best ideas occur to lots of people. I too had thought of comparing ego power to artifact power in order to ensure that pyrel doesn't generate ego items that are too powerful. But before I can play with object power I need to implement monster power, and before I can work that out I need to implement combat, which means I am currently getting my head around pyrel's proc system. So far I can define them and load them, so I'm doing well. I've defined area attack and ranged attack procs. Next step is to use one successfully ... that's proving a little harder ...

[buzzkill]

Off the wall idea... Randarts are generated in advance. Egos are essentially lesser randarts. Why not generate egos in advance (assuming we have the computing power to create and store 1000's of possibilities) and ensure ensure that at least "on paper", that no ego would rival a similar artifact (there will always be situational superiority). In lieu of that, just compare on the fly generated egos to artifacts or randarts of a similar base type and nix the ones that are too powerful (perhaps throw in some sort of DL weight so that deep ego daggers, etc. can exceed the early artifacts).

Begin rant, again (seems like all I do anymore). Escalation of power is a problem, and the removal of cursed items didn't help at all. IMO, decent starting equip should be store bought because everything the the very early dungeon would almost certainly suck, and I mean suck bad, unenchanted maybe, cursed likely, and go from there. The fact that finding a powerful ego in the first 10% of the dungeon isn't uncommon IS a problem and I suspect that the long term solution will be to ramp of the power of everything else to compensate. It's just dev instinct, or so it seems. No offence intended to anyone in particular.

My new near resolutions: Start own variant, kill self. Not to worry though, it's a new year resolution. It's never gonna happen.

[Magnate]

Well I havent played pyrel, but what you describe is exactly what I was worrying about.
If egos offer alot of extra dice, sides and whatnot, they might be no longer inferior to artifacts as you say.

The design problem that current V has solved is this: How can artifacts be better than egos, but egos still be wanted once you have found an artifact ?
It has done this by having high peaking base items, so low base + high artifact = high base + low ego.

There is of course nothing wrong with having ego and artifact on the same level and removing the base item peak - but is that what you want ?

Well the whole concept of ego has changed, so we're not comparing like with like. In V, an ego is a one-shot thing, and you get whatever that ego brings. In pyrel, ego is made up of a random number of affixes, each of which alone is a small boost but in combination can create some interesting and powerful things. This is why we don't need powerful base items.

It doesn't sound like you have tried v4 - try it and you'll see what I mean.

[Estie]

This rather depends on quite a lot of other things. Personally I don't think your last statement holds true at all in v4 or pyrel, and wouldn't in any affix-based system. v4 and pyrel create a lot of ego items with extra dice, sides, slays and other bonuses that can rival artifacts. In fact the danger is the opposite - it becomes difficult to ensure that artifacts are truly special if ego items can be so good. This difficulty is particularly acute for weapons, because you can't use the "not normally found in this slot" trick that makes armour artifacts interesting.

But pyrel allows for Crawl-style unique abilities, so I don't anticipate this being a problem ...

Well I havent played pyrel, but what you describe is exactly what I was worrying about.
If egos offer alot of extra dice, sides and whatnot, they might be no longer inferior to artifacts as you say.

The design problem that current V has solved is this: How can artifacts be better than egos, but egos still be wanted once you have found an artifact ?
It has done this by having high peaking base items, so low base + high artifact = high base + low ego.

There is of course nothing wrong with having ego and artifact on the same level and removing the base item peak - but is that what you want ?

[Magnate]

Thanks for the comprehensive explanation

Something that occurs to me:

Currently the best bases are a) much better than the next best base type - and I am talking purely about damage dice and not about "of X" which might or might not be tied to base items - and b) extremly rare compared to the next step bases.

This leads to the situation that, assuming you do not find the artifact of the #1 base type (Deathwreaker), you can either use a top tier base item with ego ("MoD of slay evil") or a lesser base, but excellent artifact (Aule). Thus egos are less powerfull than artifacts, but still viable and in use.

If you reduce the step in quality from #2 to #1, youll want to squelch all non-artifacts once you find Aule.

This rather depends on quite a lot of other things. Personally I don't think your last statement holds true at all in v4 or pyrel, and wouldn't in any affix-based system. v4 and pyrel create a lot of ego items with extra dice, sides, slays and other bonuses that can rival artifacts. In fact the danger is the opposite - it becomes difficult to ensure that artifacts are truly special if ego items can be so good. This difficulty is particularly acute for weapons, because you can't use the "not normally found in this slot" trick that makes armour artifacts interesting.

But pyrel allows for Crawl-style unique abilities, so I don't anticipate this being a problem ...

[Estie]

Thanks for the comprehensive explanation

Something that occurs to me:

Currently the best bases are a) much better than the next best base type - and I am talking purely about damage dice and not about "of X" which might or might not be tied to base items - and b) extremly rare compared to the next step bases.

This leads to the situation that, assuming you do not find the artifact of the #1 base type (Deathwreaker), you can either use a top tier base item with ego ("MoD of slay evil") or a lesser base, but excellent artifact (Aule). Thus egos are less powerfull than artifacts, but still viable and in use.

If you reduce the step in quality from #2 to #1, youll want to squelch all non-artifacts once you find Aule.

[Magnate]

Couldnt you reduce the chance for BoC to wanted level for Doomcaller, and set the probability for any BoC to become an artifact to 1 ?

Out of curiosity: if you remove BoC completely, how does the best sword base (executioners sword ? that should probably also go for the same reason BoC does) compare to the best polearm/hafted base ?

Ok, stepping back a bit: there has long been a major anomaly in V, with the best weapon of each class being much more akin to an ego item than a base item. (I'm talking about BoC, Scythe of Slicing and Mace of Disruption.) There are several ways to deal with this, one of which is to give them the "OF_GOOD" flag and use that to put them in a separate allocation table.

In v4, anything with "of" in its name was turned into an affix - so you can still generate MoD and SoS, and they look exactly like their V counterparts - and they can get ego powers. This manages their rarity in the same way as other powerful bonuses. The BoC is the odd man out because there is no base item called "blade". The simplest solution is to create one, but v4 doesn't do this. It allows the "of Chaos" affix to be available to the top three base swords (katana, zweihander, executioner).

In pyrel we don't have this limitation at all. The 'subtype' field is irrelevant for artifacts - they're not chosen with reference to it at all. So we can call Doomcaller a 'blade' without ever needing to generate any other blade. We also use of Slicing/Disruption/Chaos as affixes, like in v4.

So the answer to your question is that the top base items are the executioner's sword, lochaber axe and the two-handed great flail, which are all roughly comparable. (In pyrel they are 2d12, 3d8 and 2d10 respectively.)

[Estie]

Couldnt you reduce the chance for BoC to wanted level for Doomcaller, and set the probability for any BoC to become an artifact to 1 ?

Out of curiosity: if you remove BoC completely, how does the best sword base (executioners sword ? that should probably also go for the same reason BoC does) compare to the best polearm/hafted base ?

[Magnate]

As I understand it, randarts have always taken the power of their base item into account. In other words, Bladeturner is a powerful artifact largely because it's based on PDSM; the randart based on Bladeturner should thus be similarly powerful even if it has a different base item (though it should be equivalently rare as well).

I think perhaps Estie's point was that the randart generator treats all power on an item as the same currency, which means that if Bladeturner turns into a base item significantly less powerful than PDSM, it gets a ridiculous number of other bonuses.

In fact the randart generator *does* try to ensure that it doesn't turn into anything wildly different in base item power - which is why it turns into top-end items like mithril shields and blades of chaos rather than daggers and sandals. But PDSM is way, way more powerful than any other base item, so there's a limit to what this can achieve.

In fact, I'm quite keen in pyrel for artifacts not to be tied to base items at all. Doomcaller should still be a "blade of chaos" even though such a base item no longer exists. We couldn't do this in v4, but we can in pyrel.

[Derakon]

One reason I can see for this is that an artifacts value can be written as sum of base item and additional modifiers. Currently afaik it isnt done that way, but the problems with randomizing Bladeturner in the randart generator could have been avoided if it was.

As I understand it, randarts have always taken the power of their base item into account. In other words, Bladeturner is a powerful artifact largely because it's based on PDSM; the randart based on Bladeturner should thus be similarly powerful even if it has a different base item (though it should be equivalently rare as well).

Where you potentially get into trouble is with going from defensive abilities to offensive ones (e.g. if Vilya ends up as a weapon this time around), because offensive abilities tend to snowball more than defensive ones -- having more makes you exponentially more powerful.

[Estie]

I agree with this. I don't see any reason to tie artifact generation chance to base item generation.

One reason I can see for this is that an artifacts value can be written as sum of base item and additional modifiers. Currently afaik it isnt done that way, but the problems with randomizing Bladeturner in the randart generator could have been avoided if it was.

Suppose someone wants (maybe for a mod) reduce the frequency of PDSM drops. If artifact generation isnt tied to base item generation, he probably also has to change the frequency of Bladeturner lest it become more common than the unenchanted base; if the check for artifact is done after the base item is determined, its easier to tune frequencies.

[Derakon]

Would you do that during the lootTemplate's constructor, or would you construct all three templates and then choose at item gen time?

I think I could probably implement this without too much difficulty.

During the constructor, I think.

[Magnate]

Yeah, the tree approach is certainly a possibility, and a good way to encode hierarchical loot systems. I feel that trees ought to be declared at the LootTemplate level -- i.e. a LootTemplate that says "20% of the time, pick from this template; 60% of the time pick from this one; the remainder of the time, pick from this third one".

Would you do that during the lootTemplate's constructor, or would you construct all three templates and then choose at item gen time?

I think I could probably implement this without too much difficulty.

[Derakon]

Yeah, the tree approach is certainly a possibility, and a good way to encode hierarchical loot systems. I feel that trees ought to be declared at the LootTemplate level -- i.e. a LootTemplate that says "20% of the time, pick from this template; 60% of the time pick from this one; the remainder of the time, pick from this third one". That would also handle the "artifactChance" system without needing procs to re-weight every item's commonness: you'd just have a LootTemplate that is "all the artifacts" and another that is "everything else".