Ashenzasi/Octopode Bug?

[Fungus]

I hope this is the right area to post this -- if not, could someone please direct me to where I should mention this?

Here's the scoop: I'm playing an Octopode of Ashenzari in 0.11 trunk and have cursed at least TWENTY scrolls of remove curse throughout the game. Every single one of them gives me curse jewelry. Every. Single. One. I have a huge stack in my lair so I know the numbers are not exaggerated. I've cleared the Lair rune branches, the Vaults, and and now on D:25. Every time I pray I think "maybe this time" and... nope. I have 8 cursed rings currently equipped and no other jewelry in my inventory.

Is this intentional behavior? Bad luck combining with some sort of weighting system for octopodes? My last Ash character (in 0.10), a Dr, didn't have this problem at all. I have a shield and a hat I've been wanting to curse all game! (For the record, I have six scrolls of curse weapon from the ground, so maybe the RNG just hates me).

[nicolae]

Is this intentional behavior?

As far as I'm aware, yes, converting remove curse into curse scrolls is biased towards curse jewellry for Octopodes, since they need more cursed jewellry to get to half-bound. Though not completely biased, that's the RNG's fault in this case, I think.

[BlackSheep]

Looks like all sacrificed scrolls are curse jewellery for felids and randomly chosen for non-octopodes. Octopodes have a 75% chance of just getting curse jewellery and 25% chance to get a random curse scroll (which has a 1/3 chance of being jewellery anyway). So the odds are very slanted. I am surprised you haven't found curse armour on the floor at all, though.

[Fungus]

Looks like all sacrificed scrolls are curse jewellery for felids and randomly chosen for non-octopodes. Octopodes have a 75% chance of just getting curse jewellery and 25% chance to get a random curse scroll (which has a 1/3 chance of being jewellery anyway). So the odds are very slanted. I am surprised you haven't found curse armour on the floor at all, though.

I did find one curse armour and used it early in the game (pre-Lair). I guess I should stop wasting my scrolls, since I have, what, only a 8% chance of getting curse armour? I don't know if my opinion really matters here, but that seems a bit harsh. I guess my only chance is Pan scumming.

[eeviac]

why can't we just let ash worshippers choose which scroll they want

[dtsund]

why can't we just let ash worshippers choose which scroll they want

I believe interface burden was the argument for not doing that.

[Bloax]

I can't imagine it being harder than the dropping system, I wouldn't mind it being exactly like the way you drop things - in fact.

[Fungus]

What about something like cursing 1 scroll uses current weight but cursing 2+ scrolls give you guaranteed variation? If I knew I could curse 3 scrolls and be guaranteed one of each type, that would be rather handy.

[eeviac]

Or how about instead of creating a cursed scroll, praying to ash burns a remove curse scroll in your inventory/on the ground to curse something of your choosing.

I mean the way ash works now isn't terrible, it just has the potential to create frustrating early-mid game situations like the one in this thread.

[eeviac]

I like that - remove the three curse scrolls in favor of a general 'scroll of cursing'. It could work like a mummy death curse when read, and could target anything in your inventory when under Ash. This change makes both Ash and the ID game better, IMO.

[galehar]

why can't we just let ash worshippers choose which scroll they want

Because it's a god gift. You don't get to ask your god for what you want. Also, there's a lot of RC and curse scrolls in the game, so it's rarely a problem.
But maybe this Oc hack could be tweaked. A human has 3 jewellery slots on 10 (30%). An octopode has 8 on 11 (72%). The human has a 33% chance of getting curse jewellery, the octopode 83%. Changing one_chance_in(4) to one_chance_in(3) here would change the chance to 77% which seems more reasonable.

[twelwe]

My best proposal for this incorporated mummy curses as well. As an Ashenzari worshiper, when you read any kind of curse scroll, or are cursed by a mummy, you gain an invocation ability to curse any item like how delayed fireball works. Ashenzari already absorbs mummy curses, but there is no benefit aside from just not being cursed. The remove curse -> curse foo scroll game seems silly to me. Mummies are plentiful later in the game and this actually makes Ashenzari cursing in the early game much more permanent.

[BlackSheep]

Don't Octopodes have 9 jewellery slots? (Amulet too)

[galehar]

Don't Octopodes have 9 jewellery slots? (Amulet too)

Haha, of course they do. It makes their jewellery slot ratio 9/12 = 75% and doesn't really change my reasoning.

[BlackSheep]

I agree, that sounds like a good change.

[sanka]

As my favorite god is Ash, I would really like more to use remove curse scrolls directly for cursing or merge the three scrolls or something than to tweak the chances. If it sounds too powerful, make skill bonuses a little bit smaller in exchange, or reduce piety gain for being highly bound, or both, or whatever.

Currently Ash is the most unpredictible in the sense that you have a hard time guessing when she will be finally useful. You not only need to find equipment for the slots to fill, scrolls to use on them, but you also need some luck to get the right combination. I had characters who was fully cursed/full piety before Lair, and learned a lot of high level spells very quickly. I also remember characters who reached the vaults (after clearing lair+orc) with only the weapon hand bound. I know crawl supposed to be random, but this usualy just feels annoying to me (It's annoying enough you can't get full bonus because you can't find any hats/gloves for a long time - the curse mechanic just makes this worse occasionaly).

[jejorda2]

Merging the scrolls makes fewer bad scrolls, so the scroll ID game gets easier. That's probably a bad idea.

Adjusting the chances of getting curse jewelry probably helps, but it still makes it easy for the RNG to make the game frustrating. I think it encourages scumming to leave it random, since you can always eventually get the scroll you want.

Letting sacrifice of remove curse scrolls directly curse an item might give the player too much control of the god, but I'm not sure that deciding when to spend scrolls to get Ash's curses is any worse than deciding when to use food to get Trog's berzerk.

Could the chance to get a particular curse scroll be made dependent on the number of uncursed slots and the number of unused scrolls in stashes? So if you have two uncursed armour slots, three uncursed jewelry slots, and an uncursed weapon, for six uncursed slots, you have a two in six chance of geting curse armour, a three in six chance of getting curse jewelry, and a one in six chance of getting curse weapon. If there is a curse jewelry scroll in the temple and two curse armour scrolls in shops, those chances get adjusted to no chance of curse armour, two in three chance of curse jewelry, and one in three chance of getting curse weapon.

[BlackSheep]

Could the chance to get a particular curse scroll be made dependent on the number of uncursed slots

This is the best suggestion I've heard for reducing frustration with Ash sacrifices. I wouldn't eliminate the chance of getting a scroll for a fully bound slot, but maybe apply a weighting so that at fully bound and fully uncursed, the chances for each scroll are the same as now (or the same as they will be after galehar tweaks the octopode numbers), but tilting the weighting away from cursed types otherwise.

[Blade]

Merging the scrolls makes fewer bad scrolls, so the scroll ID game gets easier. That's probably a bad idea.

It is virtually impossible for the scroll id game to become easier than it is right now. The curse foo scrolls are utterly insignificant in essentially all non-Ash games.

[quote='jejorda2"]Could the chance to get a particular curse scroll be made dependent on the number of uncursed slots?[/quote]
I definitely don't think the items currently available in the dungeon should have any effect on the gifting, but significantly increasing weight of scroll types with uncursed slots sounds like a very good idea.

[galehar]

Could the chance to get a particular curse scroll be made dependent on the number of uncursed slots

This is the best suggestion I've heard for reducing frustration with Ash sacrifices.

So, let me use this ?RC to clear my foo slots to skew the weighting toward curse foo before sacrificing this big stack of ?RC. Spoilery?

[BlackSheep]

Could the chance to get a particular curse scroll be made dependent on the number of uncursed slots

This is the best suggestion I've heard for reducing frustration with Ash sacrifices.

So, let me use this ?RC to clear my foo slots to skew the weighting toward curse foo before sacrificing this big stack of ?RC. Spoilery?

That is kind of scummy behavior, yes. It's also risky behavior since you'd have a chance of not getting the scroll type needed to recurse those slots and losing your skill bonuses for a while. Personally, if I'm relying on the bonus for min. delay or spell success, I never uncurse something unless I already know I can recurse it.

[smock]

What if Ash gave "scrolls of curse equipment" that could be used to curse any item? Other cursing scrolls could be retained but could curse only the type of item specified.

That way (1) the scroll ID mini-game would not be affected, (2) the player cannot choose the gift, (3) scrolls of remove curse to be generated less frequently (I'm assuming that Ash is why there are still so many of them generated; perhaps not and this point is moot), (4) issues regarding species-specific weights (which seems lightly problematic in terms of theme and Crawls' stated principals) could be avoided, (5) is not spoilery and cannot be influenced by player behavior.

[XuaXua]

Do not do it randomly. Re-adjust the chance weights based on usage. edit: reworded from "Balance it based on usage"
Distribute a percentage of the received scroll type to the percentage chance of all unreceived scrolls types, where the percentage value distributed = 100% / # of slots.
Reduce the received scroll type to a minimum of percentage chance.
If a scroll type is received and its current percentage chance is greater than its base percentage chance, reset it and distribute the percentages accordingly to all other types.

Base percentage chances for Octopode:
90% chance of Curse Jewelery
10% chance of Curse Armour

Pray, gain one Curse Jewelery; 10% is distributed to other types (Curse Armour)

80% chance of Curse Jewelery
20% chance of Curse Armour

Pray, gain one Curse Jewelery; 10% is distributed to other types (Curse Armour)

70% chance of CJ
30% chance of CA

Pray, gain one Curse Armour; CA is reset to 10% and the 20% is redistributed to all other types (CJ)

90% chance of CJ
10% chance of CA

At its lowest, it would get to
10% chance of CJ
90% chance of CA

which means you still are not guaranteed CA by this time, but the likelyhood has increased greatly.

[XuaXua]

With the above, let's say you're a human with 10 slots (6 armour slots, 3 jewelery slots, 1 weapon slot).
Percentage distributed is 10%

Base percentage for humans:
60% chance of Curse Armour (CA)
30% chance of Curse Jewelery (CJ)
10% chance of Curse Weapon (CW)

Pray, convert one RC to a Curse Armour; distribute 10%

50% CA
35% CJ
15% CW

Pray, convert one RC to a Curse Armour; distribute 10%

40% CA
40% CJ
20% CW

Pray, convert one RC to a Curse Armour; distribute 10%

30% CA
45% CJ
25% CW

Pray, convert one RC to a CJ; reset CJ to base 30% and distribute 15%

37.5% CA
30% CJ
32.5% CW

Pray, convert one RC to a CJ; distribute 10%

42.5% CA
20% CJ
37.5% CW

Pray, convert one RC to a CJ; distribute 10%

47.5% CA
10% CJ
42.5% CW

Pray, convert one RC to a CJ; CJ is already at minimum 10% do nothing.

47.5% CA
10% CJ
42.5% CW

Pray, convert one RC to a CA; distribute 10%

37.5% CA
15% CJ
47.5% CW

Pray, convert one RC to a CJ; reduce to minimum 10%, distribute 5% overage

40% CA
10% CJ
50% CW

[dtsund]

Do not do it randomly. Balance it based on usage.

This is both abusable and spoilery.

[BlackSheep]

Usage probably wasn't the best choice of words. If you read his posts, he's looking for a weighting based on prior sacrifices.

[dtsund]

Usage probably wasn't the best choice of words. If you read his posts, he's looking for a weighting based on prior sacrifices.

Oh, whoops! Disregard my nonsense.

[XuaXua]

Usage probably wasn't the best choice of words. If you read his posts, he's looking for a weighting based on prior sacrifices.

Oh, whoops! Disregard my nonsense.

Also it will never go to 0%, so there's always a chance you get the same scroll you received previously, just (usually) less of a chance of it happening.

[galehar]

Distribute a percentage of the received scroll type to the percentage chance of all unreceived scrolls types, where the percentage value distributed = 100% / # of slots.

The end result would be completely unnoticeable. A lot of coding to try to emulate what statistics do just fine.

[Fungus]

Any merit to the idea that if you drop exactly three scrolls of RC you get one of each?

On a side note, one thing that was a bit frustrating playing a mage Octopode of Ash vs. other mages of Ash was how long it took to even GET 8 rings.

[smock]

Distribute a percentage of the received scroll type to the percentage chance of all unreceived scrolls types, where the percentage value distributed = 100% / # of slots.

The end result would be completely unnoticeable. A lot of coding to try to emulate what statistics do just fine.

I think this would be noticeable. At present, you get the occasional annoying game in with many scrolls of curse weapon are generated but few other types are. The proposal doesn't change the mean probability of getting a scroll of curse foo. Instead, it would reduce the variance without changing the mean. Whether that effect is worth the coding effort is another question entirely.