|
Post by yune on Sept 20, 2021 12:14:18 GMT
The limbo P2 weapons don't seem to be able to tell what damage types to use, even on non-healing mobs. It often picks types the mob has absolute immunity to (elements that can't be inflicted on that mob by a Lash) and doesn't even always pick the same types consistently for multiple instances of the same mob.
Examples (all are after the types changed on attacking the mob):
Osyluth: Sometimes picks cold. Non-paragon Osys have absolute cold immunity. Pit Fiend: Picked Acid/Cold/Div. It should be Sonic/Cold/Div. PFs have absolute acid immunity. Pit Fiend case 2: Attacked one PF, it picked CED. All of those types damage it, but CSD is better. Without hitting anything else in between, swapped to a second Pit Fiend. It changed to CFD. They're fire immune, and CED wasn't even right to begin with. Pit Fiend case 3: Picked ACD. They're acid immune. Korachron: Picked Acid/Electric/Positive. They're immune to acid and electric. Other roaches picked other elements, but never sonic, which is the best element for them. Erinyes Vitarch: Picked ESD. They're elec immune.
Sometimes it does pick the right 3 types, but more often it doesn't. And it doesn't even pick the same wrong types consistently. But it is very consistent about picking the healing type for superior+ mobs in hells, as advertised.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 20, 2021 12:24:51 GMT
It picks what the mobs take best. Those immunities adjust some even with paragon iterations. They likely don’t factor in inflictions though. The current creatures in the game aren’t designed around limbo weapons existing rendering them all but useless because the things they take best, also cause them to heal. It’s unfortunate that entire classes of weapons are meaningless and limbo weapons just add to that list.
|
|
|
Post by yune on Sept 20, 2021 12:54:07 GMT
Prep, that doesn't explain why they pick things like acid against Pit Fiends. PFs have 100% acid immunity and their acid immunity can't be inflicted. And PFs never heal on it.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 20, 2021 13:30:44 GMT
The game likely registers the 100% acid as what they take best. That’s what I am saying. I didn’t code it. But there’s a whole slew of mobs that “take” stuff they have absolute immunity too. It’s registering it.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 20, 2021 13:53:03 GMT
None of that makes it ok. I agree 100% there is a problem with them. As I said originally, the mobs weren’t designed with these weapons in mind and there’s going to have to be adjustments to the current code, or the acceptance that the weapons are trash until mobs designed around how they work added. One or the other. Just add them to the pile. Here’s looking at you abo weapons, and red glaze and yellow glaze and the like.
|
|
|
Post by woqued on Sept 21, 2021 8:34:04 GMT
When you were testing this, were there inflictors and/or dr drops present? The weapons might not consider absolute immunity as part of the deal at all, and so absolute imm 100% mob cold inflicted might have the least amount of cold immunity somehow, while retaining absolute imm so making it negligible... Or something among those lines.
So, dr/imm drops? I recall on quick testing they were working fine under very "sterile" circumstances.
|
|
|
Post by chirality on Sept 22, 2021 0:09:37 GMT
The phrase "absolute immunity" doesn't exactly hold the same meaning as "100% immunity." Lashes just can't inflict a type with 100% immunity. It's not quite precisely equivalent. I understand that the context used is to point out that no lash infliction was skewing the imm% in question, but just to be clear, absolute immunity is not necessarily 100% and 100% is not absolute immunity, so it's probably best to avoid using the two concepts interchangeably and simply say a mob has 100% immunity and therefore can't be lash inflicted.
I remember reading/hearing about old bug where PFs healed on acid for some period, which got fixed. I don't know the details of the bug or how it was fixed, but I do know that there was also an engine bug of some unspecified nature involving immunity % that had to be fixed at some point long ago as well (maybe it was related to the PF thing, or to lash inflict, or maybe all three were related in some combination, but I'm not privy to the details, just old forum posts and chats). I recall a few different issues involving imm% and inflicts over the years, and also a few features that had to be tweaked to change the way things worked (like lashes inflicting heal type).
I believe there were various complications with immunity % calculations and the way different features or things interacted with them, even before healing comes into the equation. I think there's probably a fairly tangled web of code that was adjusted (or, likely even completely refactored) multiple times over the years in regards to imm% and things that affect or evaluate it.
Anyway, whether it is some debuff throwing off the numbers in a weird way, or paragon healinng/imm% having some impact, I definitely wouldn't be surprised if there was some oddities and edge cases for the weapons to pick the "right" damage types. We know that other special features which intelligently select a mob's "weakest" imm types (aka ele swarm) are generally reliable by default, except when healing imms would screw it up, so the check to not deal that type even if it's "the best type" exists to prevent healing.
DR may also play a factor, as woqued mentioned, or it may not; without seeing the implementation of the weapons, it's hard to know exactly how they work.
Just to re-tldr some earlier points: paragons really impact imm%s; many healing paragons have 0% imm and 0 resist, but not all; as prepared mentioned, these stats also vary across different paragon rank, and not always in simple increases to imm and dr from the lesser paragon rank to the next.
|
|
|
Post by yune on Sept 22, 2021 4:43:28 GMT
Hell mobs don't have elemental/exotic DR, and lash inflicts all elements equally, so it should never skew the element choice. Most of my testing was on non-paragons. It seems most of the issues are from picking types that the hell mobs have 100% immunity to, which should never happen no matter how many inflicts are on it.
|
|
|
Post by chirality on Sept 22, 2021 15:31:32 GMT
a) I was assuming that you weren't inflicting anything with a lash, so my post wasn't based on the idea that you were--I was just pointing out that "absolute immunity" is a threshold below which cannot be inflicted (eg 60%, 70% on some xDD color type), whereas 100% imm is just a value that a lash cannot inflict. I only even went on the tangent because you specifically defined absolute immunity as an element that can't be inflicted by lash. That's not really an accurate definition, because "an element that can't be inflicted by lash" would simply be defined as "100%", whereas "absolute immunity" as a concept and term is entirely independent of 100% or lash infliction. b) Of course lash inflicts all elements equally...well, except for any that are 100%... c) I did forget that outdated data from paragons having idiosyncratic DR changes (i.e. sometimes DR was added between ranks, other times lowered) is no longer relevant after mob DR redux. Nonetheless, DR aside, I wouldn't be surprised if imm% wasn't much different from how it used to also shift in asynchronous and unexpected ways across paragon ranks (i.e. sometimes a higher rank would actually have lower imm% than earlier rank to given type(s), and sometimes healing type imm% would increase or decrease between ranks as well). The point there wasn't to say that you were missing something about KV data that would explain why the weapons were picking stupid types, but more just to say that sometimes mobs have weird stats, like some healers have 0% to their imm type at p2, and never changes, some have some % at p2 and goes down, some healers have 30% imm to their type and a high multiplier, etc. I was just trying to say that healing imm types are weird. All that was more of a tangent than a rebuttal to the OP, so my bad if that came off like "here's why you could be wrong..." d) I think your intuition is correct that healing type and/or 100% immunity are the two things that are borking whatever algorithm the weapons use to select the "best types," so hopefully whenever Funky has a chance, maybe the code could be looked at just to double-check for any obvious simple bugs or other things that may be making them not work right (although some more elaborate and documented testing would probably help immensely there). I assume the implementation is a lot of greater-than/less-than/equal-to checks and similar if-then checks based off of arithmetic comparisons and true/false operations (things like "if type is 100%, then do this"), and even when code like that is written bug-free, there is plenty of room for things to work differently in practice, especially on edge cases, due to how mathz works on compooters. Then, there's room for very simple errors or omissions in the code, too. Anyway, I wasn't trying to say it's all working perfectly and you're just wrong. I think the opposite is more likely--there's some flaws in the algorithm that don't account for certain things that didn't come up in testing, and/or there's some bugs that didn't get caught. Good on you for trying to test and report it. I wouldn't be surprised if some inflicts were involved to explain some oddness, because I think the weapons were tested fairly extensively and seemed to work ok, but I also wouldn't be surprised to discover that those tests weren't extensive enough, or that there were some considerations left out or bugs un-noticed, which are causing the unpredictable behavior you've seen. I think it's definitely the case that those weapons were never used in practice enough to document problems, given the existence of XR weaps kind of making the p2s obsolete by default, regardless of them ever working or not in the first place.
|
|
|
Post by yune on Sept 23, 2021 0:00:52 GMT
The other super weird thing is that the p2 weps aren't consistent between multiple instances of the same mob. Look at all the different selections it picked vs p0 Pit Fiends.
|
|
|
Post by woqued on Sept 23, 2021 18:05:31 GMT
On very brief testing in Oinos, something fucky is happening indeed.
The pt2 weapon seems to consistently pick the right damage types vs monsters such as Nabassu, Bulezau, Nalfeshnee (killed several).
But almost completely random vs some PoM mobs like Piscoloth, Mezzoloth and Nycaloth, as well as Erinyes (didn't kill several of all; only mezzoloths and nycaloths in multiples). They were easy since it was possible to get multiples of them to show up at once. I don't know about their elemental immunities, but whatever the weapon picked seemed completely random and definitely not what they were taking the most.
The only obvious constant here seems to be that the above category doesn't have 100% to any elemental immunity, while the bottom category does. Testing was done by greater restoring the monsters to remove any infliction possible when there were several around, and constant swapping of targets to force trigger the swapping of types on the weapon.
Very low effort testing with large margin for errors, but eh. Cba to do more; but regardless the weapons aren't working as they should.
|
|
|
Post by starcore on Oct 13, 2022 2:42:15 GMT
It seems like these may be treating absolute immunity as 0%... its always picking fire and cold against primal fire elementals (in elemental plains). Cold is correct, but they take 0 from the fire.
|
|