Issues on current master

[Estie]

Silly idea: the current slots have average power ratings and slot biases, right? Like randart boots are more likely to have speed, randart body armor to have CON, etc. Each game, switch up the average power ratings and biases for each slot. So in one game, shields might be more likely to have INT/WIS and gloves to have Speed, while helms get the +to-hit/dam that gloves normally have.

Since egos are effectively "fixed" currently, this increases the variability of randarts vs. egos. The most notable problem I see with it is Boots of Speed/Elvenkind, which would now be able to stack with whichever slot got the "randart boots normally get speed" bias.

The problem remains: for randarts to be used over egos, they need to be more powerful, and since the standarts are pushed so close to the egos, there simply is no room to keep the power in line. You will end up with super toons.

But that aside, I like the idea and will continue to playtest everything you try with randarts.

[Derakon]

So here is the task: create an algorithm that makes items with some variance, for example boots with stats and hat of speed as opposed to always boots of speed and stats from the helm. How would you go about it ?

Silly idea: the current slots have average power ratings and slot biases, right? Like randart boots are more likely to have speed, randart body armor to have CON, etc. Each game, switch up the average power ratings and biases for each slot. So in one game, shields might be more likely to have INT/WIS and gloves to have Speed, while helms get the +to-hit/dam that gloves normally have.

Since egos are effectively "fixed" currently, this increases the variability of randarts vs. egos. The most notable problem I see with it is Boots of Speed/Elvenkind, which would now be able to stack with whichever slot got the "randart boots normally get speed" bias.

[Estie]

@Estie--I don't understand the rationale of your suggestion. Are you claiming that big-3 ego weapons are too powerful? What about Dwarven PDSM? and Elvenkind<+5,+9>? For most slots, there are egos or ordinary articles (=CON,"Weaponmastery, "Trickery) that are commonly used instead of even most standard artifacts. Not sure this is bad.
Changing them because it messes up randart balance is the cart before the horse.

Yes I think egos have become too powerful compared to artifacts (standarts, too). MoDs (HA), while they did in theory exist before, used to be a rarer find than Ringil. Similar for ego PDSMs; they are just too common, and speed boots have (unfortunately) always been uncontested in slot, whether they have stealth on top or not.

Having egos in some slot isnt bad at all, in fact I would be worried if at some point only artifacts were used. But the quality of what kind of ego you can expect to have after a few GVs has gone up dramatically, while the artifacts have been nerfed (!). Previously, you might be using Calris or Aule or Aglarang, but these days you are likely to replace it soon with a great ego weapon which is also likely to remain your best option till the end. Surely you must have noticed this ?

Now to reverse this, removing great weapons is not the only way, as noted before; Derakon suggested a way to effectively make them rare again, while my post was going to reduce the power of the top base weapons. I am no longer sure my idea is better than Derakon´s.

If you want to keep all as it is for standarts and just adjust randarts, that is certainly possible, but I ask that you find a way that leads to variations in the item set used, because thats the only friggin point of randarts.
And since the egos are fixed, the only way to do this is to ensure that the randart generator creates stuff that tops the egos _on average_ , which leads to super powerful toons.

So here is the task: create an algorithm that makes items with some variance, for example boots with stats and hat of speed as opposed to always boots of speed and stats from the helm. How would you go about it ?

[Gwarl]

Regarding the latest build;

Looking through some spoiler files for artifacts from a few games, I get the impression the artifacts being generated are distinctly weaker than standarts.

There also seems to be a bug where a randart is rolled with a negative power rating.

[Nick]

It seems there is a good chance that rings pick up the "no melee and bad at casting" flag from rings of escaping, which makes even high randart rings all but useless. This should probably not happen, or be reduced to the extent of badly cursed artifacts.

What's happening here is that randart rings and amulets are based on base ring and amulet types, and you've manged to score two based on =Escaping in one game. I think that's just really unlucky, but I'll check there's not some sort of bias

Also I had gauntlets with a curse that gave negative modifier to dexterity; I could remove the curse, but the dex penalty remained.

Yes, randarts can get negative modifiers when they get a curse. That probably doesn't really fit with the new curse system; I'll have another look at that

The phial still looks like dark red *

That sounds like a bug with how I'm handling the special artifact objects.

[Nick]

I'm in favour of doing both of these together. Now we've got rune-based ID, what I'd really like is to break the sort of rigid, predictable progression you get of single resist -> Resistance -> Elvenkind -> artefact, etc., and replace all the fixed-trait egos with templates that just specify how many of each random trait they should get. e.g. An armour type that gives one random base resist and one random higher resist; or one with two random higher resists and +1 to a random stat; or a weapon type that gives two random slays and a random ability, etc.

Not sure if we should really have higher-level egos with as many bonuses as Holy Avenger etc. anyway, but even just randomising the existing ones so that instead of guaranteed Slay Evil/Undead/Demon and See Invisible you sometimes get one with Slay Animal/Orc/Dragon and Feather Falling would go a long way to making the really useful ones rarer.

Or maybe you could free it up even further, and just have a list of allowed traits for each slot, and pick completely randomly from those according to the level of the ego. So for instance, for a given slot you might have allowed lists of "good", "great", and "superb" traits. (Say, for an ego helm, "good" might be a bonus to Infravision or Searching, "great" might be See Invisible or pBlind, and "superb" would be ESP.) A level 1 ego would get a single pick from the "good" list. A level 2 ego would get either two picks from the "good" list or a single pick from the "great" list. A level 3 ego might get one "superb" trait, one "great" + one "good" trait, or three "good" traits. Etc.

I really like the way you're thinking here, and I think there are a number of ways it could be done. I'm attracted to the idea of reducing the number of ego items, but increasing their variability.

[Pete Mack]

@Estie--I don't understand the rationale of your suggestion. Are you claiming that big-3 ego weapons are too powerful? What about Dwarven PDSM? and Elvenkind<+5,+9>? For most slots, there are egos or ordinary articles (=CON,"Weaponmastery, "Trickery) that are commonly used instead of even most standard artifacts. Not sure this is bad.
Changing them because it messes up randart balance is the cart before the horse.

[Estie]

Holy Avenger is an ego weapon with Slay Evil/Undead/Demon, which makes it quite viable in the late game if you get it on a big weapon. It's not that infrequent that you see a winner wielding a Scythe of Slicing (or MoD/BoC) of Slay Evil, especially in randart games. And HA is strictly better than Slay Evil.

The simple fact of the matter is that for weapons, damage is king, and you'll usually find at least one big-damage ego weapon in the game, while randarts are much less reliable. You could "fix" this by restricting the damage output of high-end ego weapons. I believe there's no restrictions currently and indeed that if the game attempts to make an artifact for a weapon type that has no artifacts left (which includes all three of the "big three" weapon types) then it just throws a bunch of +to-hit and +to-dam onto an ego instead.

Something like this:

while (ego weapon fails to pass a check to allow it to be more powerful than intended) {
  weapon to-dam is decreased by 1;
}

should do the trick.

This would put a (soft) cap on the power of egos, I dont like that at all.

If Fizzix´s policy of "only high egos at depth" is to be maintained, then the high base damage weapon types with high egos give the base that good artifacts need to top. You could remove the great weapon types, which would make something like a Lochaber axe (HA) the best ego and Ringil tops that easily (unlike a MoD (HA)).

Or you could go back a little on the policy and mix up the ego pool again, allowing lower egos to be created. It would have the side benefit to actually make it possible to find something like a branded MoD. Then again, it doesnt really solve the issue.

I think I would simply remove the "great" types, leave Lochaber axe (3d8) as best polearm, remove the weird executioner´s sword as well and change twohanded sword 3d6 -> 4d6, and great hammer 8d1 -> 2d12 (why no spread in the first place?). That should do the trick.

[Derakon]

Some of the egos actually predate angband and I'd like them to remain simply for continuity purposes. Personally I feel like most of the higher egos are sadly junk compared to artifacts, to the point where even stuff like HA is usually hardly worth condisering.

I would have ego items start appearing sooner so they get a look in vs artifacts.

Holy Avenger is an ego weapon with Slay Evil/Undead/Demon, which makes it quite viable in the late game if you get it on a big weapon. It's not that infrequent that you see a winner wielding a Scythe of Slicing (or MoD/BoC) of Slay Evil, especially in randart games. And HA is strictly better than Slay Evil.

The simple fact of the matter is that for weapons, damage is king, and you'll usually find at least one big-damage ego weapon in the game, while randarts are much less reliable. You could "fix" this by restricting the damage output of high-end ego weapons. I believe there's no restrictions currently and indeed that if the game attempts to make an artifact for a weapon type that has no artifacts left (which includes all three of the "big three" weapon types) then it just throws a bunch of +to-hit and +to-dam onto an ego instead.

Something like this:

while (ego weapon fails to pass a check to allow it to be more powerful than intended) {
  weapon to-dam is decreased by 1;
}

should do the trick.

[Estie]

http://angband.oook.cz/ladder-show.php?id=20713

It seems there is a good chance that rings pick up the "no melee and bad at casting" flag from rings of escaping, which makes even high randart rings all but useless. This should probably not happen, or be reduced to the extent of badly cursed artifacts.

Also I had gauntlets with a curse that gave negative modifier to dexterity; I could remove the curse, but the dex penalty remained.

The phial still looks like dark red *


edit: unbarring ways now works with trapped chests, great!

edit2: or does it ? summoning rune went off this time.

[takkaria]

Gah I still have a character alive from the last build. I just found two scrolls of deep descent though so I'll see about fixing that

Savefiles should be compatible between recent dev versions.

[Gwarl]

Some of the egos actually predate angband and I'd like them to remain simply for continuity purposes. Personally I feel like most of the higher egos are sadly junk compared to artifacts, to the point where even stuff like HA is usually hardly worth condisering.

I would have ego items start appearing sooner so they get a look in vs artifacts.

[Nomad]

- randomize the egos in some way
- nerf the egos and keep the artifacts as is, so you end up using more randarts.

I'm in favour of doing both of these together. Now we've got rune-based ID, what I'd really like is to break the sort of rigid, predictable progression you get of single resist -> Resistance -> Elvenkind -> artefact, etc., and replace all the fixed-trait egos with templates that just specify how many of each random trait they should get. e.g. An armour type that gives one random base resist and one random higher resist; or one with two random higher resists and +1 to a random stat; or a weapon type that gives two random slays and a random ability, etc.

Not sure if we should really have higher-level egos with as many bonuses as Holy Avenger etc. anyway, but even just randomising the existing ones so that instead of guaranteed Slay Evil/Undead/Demon and See Invisible you sometimes get one with Slay Animal/Orc/Dragon and Feather Falling would go a long way to making the really useful ones rarer.

Or maybe you could free it up even further, and just have a list of allowed traits for each slot, and pick completely randomly from those according to the level of the ego. So for instance, for a given slot you might have allowed lists of "good", "great", and "superb" traits. (Say, for an ego helm, "good" might be a bonus to Infravision or Searching, "great" might be See Invisible or pBlind, and "superb" would be ESP.) A level 1 ego would get a single pick from the "good" list. A level 2 ego would get either two picks from the "good" list or a single pick from the "great" list. A level 3 ego might get one "superb" trait, one "great" + one "good" trait, or three "good" traits. Etc.

[Nick]

- nerf the egos and keep the artifacts as is, so you end up using more randarts.

This looks like the best choice to me

Thanks for your thoughtful reply.

[PowerWyrm]

Looking at the new changes for confuse monster, the duration was changed for "confuse monster" (wand + spells) and "confuse monsters" (staff), but not for other effects using MON_CONF: Totila activation (MON_CONFUSE) + Bedlam spell.

Also for the "wonder" effect (PROJ_MON_CONF).