Cloud Of Darkness HTBF

The thing is, SE told us that there are drop rates for items which vary by item and put them in categories.



This chart comes from a SE representative. This means that items have a rarity and that rarity determines how TH affects that drop rate. I suppose these could be "slot" rarities and not item rarities, but...

I don't really see how this jives with the idea of "ore eaters have 4 drop slots." Do they have 4 common drop slots, and each one just happens to load one of those 4 items...? What if Silver ore drops less often than Copper ore, does that mean the Silver ore is in the uncommon slot, and the other items are in the common slots?

Anyway, I don't claim to have answers to these, I'm just questioning how anyone would recognize the difference between the two systems, or how you would treat TH differently if it worked the other way.

For Lilith, Odin, & presumably this fight, we know a few things:
There is a cap on the number of drops they can give you
They can, with or without TH, drop multiple items
(I think in all cases, if you kill spite wardens) they can drop multiple of the same item

So...how do the slots work here, how does it interact with TH? How does it affect your decision to/not to apply TH?

When talking about getting "another roll" at the item drop, like a second 1/6 chance, or increasing a single chance from 1/6 to 2/6, that's functionally the exact same, it doesn't matter how the code executes it, IMO.

Tarutaru adds have Apoc always = Encumberance
Hume adds have Amanomurakumo always = Magic atk down
Elvaan adds have Bravura always = Atk down
Galka adds have Spharai always = Plague
Mithra has Kikoku/Mandau always = Avoidance down

Well on the subject of TH I can confirm I did a single VE run on THF and got Masuqe 1/1.

Mooching around this thread, an intermittent player at the moment, but I run without a THF only using TH4 for gear to tag TH on CoD doing Easy runs and got Masque on run 2, 5 and 12 for my characters. Only item I'm missing is 2 Shawls (so 10 drops in maybe 20 runs or so atm - I got a drop every single run for the first 7).

Man you got 3 masks and don't have all your capes yet? I can't imagine what that's like. I'm well over 200 kills now, most on VE, some on E, and a couple on normal, and I have yet to see a Masuqe drop. It's grinding my gears.

We're greatly off topic in here lol, would've been better to start a new thread.
Regardless, trying to simplify what you guys have been saying:


Drop Slot model

Drop rate model (categories)
I tried to keep things simple.
I don't see why you claim it's gotta be one model or the other.
Couldn't it be both?
But clearly applied differently according to target/content.
It's quite clear there are exceptions and that drops from special content and sometimes from special NMs are handled/coded in a slightly different way from the "default" way that is used for instance on most normal monsters in the open world.

I have one mask and need one neck in 2 chars. Might get another head before it drops too!

Now imagine having over 500 VD kills and only the 3 basic drops on Shinryu

This is...kinda my point (that neither model can be definitively proven, that is)? But also...if it could be both, then there's clearly not enough evidence to support either model. I think the slots model doesn't bring anything novel to the table, but brings a lot of extra (hidden from view) complexity.

It's possible that there are rolls applied to a table of items like /random 1-1000

1: Crep Cloak
2: Crep Knife
3: Crep Scythe
4: Crep Helm
5: Crep Body
6-500: Nothing
501-700: Crep Ring
700-900: Crep Earring
900-1000: Crep Pebble

Sure. But we know that it can drop multiple, so that means (if this model is true) it needs to roll this multiple times. At least up to n times, where n = the highest number of drops we've seen in a single loot pool. We also know that it's possible to get multiple of the same drop (for some mobs, not necessarily in this example, it's not important). We also know that items have a hard limit on how many can possibly drop, like Baobhan Sith drops 1-4 Cheviot Cloths. We know that Ore Eaters can drop Flint Stone, Copper ore, Zinc ore, & Silver ore. They can only drop 1 of each, but they could drop all 4 of them.

So my point is: Why would they make the roll on that table, then make a second roll on the same table, but exclude the item that just dropped, then make a third roll on the same table, but exclude the two items which have already dropped...rather than just giving each item a drop rate?

Second point: if there are different drop systems for some content (we know there are, like KB, Baobhan Sith, Fu, Mee Deggi, etc.), how do you determine this? The ones we know of have a distinct pattern which is never broken: They have guaranteed drops (1 Cheviot cloth, a Dring or a Pixie Earring, a guaranteed scale + 1 other picked from a list) and never deviate from them.

Third point: If you're trying to apply this "drop slot" model to Cloud of Darkness (See, this is actually a relevant tangent!), could you explain how it works, how you determined that it works that way, and what, if any, impact it has on how TH works in here?

As far as I can tell, there's no guaranteed drop, there's no slot that only applies for a single item, there's a chance for multiple of the same item, and all of these states can happen with or without TH. I'm really not seeing anything that implies this is anything other than random chance.

I'm inclined to believe that since the items drop off the NM, TH applies like it does to basically everything else in the game. That is to say: each item is assigned a rarity. When the mob dies, it randomly determines if that item dropped. If you have TH, it applies a bonus to that random chance equal to the rarity of the item in question. This check is done for each item, and then the ones that you "passed" the roll for appear in the treasure pool. Presumably there's a cap on total items that can drop (you could call this "drop slots" if you want), so for CoD R/E items it's 2. Once you reach 2 items, stop rolling for the other items. Just like the Ore eater, it can only drop up to 4 total items. Narasimha can drop up to 6 manticore hairs, up to 2 manticore fangs.

I don't think this is controversial; I believe this is the baseline of our understanding of how TH works for the vast majority of content in this game. If someone wants to provide an alternative I'd like for any proponent of that system to explain how it works, how they observed this behavior, and how it changes the way TH works. In CoD or elsewhere.

Terms:
Treasure Entry - One entry in the governing reward table
Drop Entry - A type of treasure entry that rewards a single item
Slot Entry - A type of treasure entry that rewards one of several possible items

Each treasure entry has a rarity within the table, effected by TH. Guaranteed(100%) is an additional rarity not shown in your chart.

Each treasure entry has either a single item (Drop Entry) or a set of potential items (Slot Entry) defined.

When the monster dies, each Treasure Entry is rolled to determine if it should be awarded. Once it's rewarded, if it is a Slot Entry, a secondary roll decides which item within the distribution is given.

Unlike the 'cap on drop slots' and individual rarity you suggest, each slot has independent probability. This model is simpler to explain when looking at things like Gaes Fete. Let's take Urmahlullu, for instance. I would posit that he has 9 Treasure Entries.

Entry 1(Guaranteed) - Slot (33% skormoth mask, 33% eschan stone, 33% onca suit, 1% chastisers)
Entry 2(common) - Vulcanite Ore
Entry 3(very common) - Slot (33% skormoth mask, 33% eschan stone, 33% onca suit, 1% chastisers)
Entry 4-9(guaranteed) - Slot (20% Eschalixir +2, 70% Eschalixir+1, 10% Eschalixir)

A model that rolls for each item then applies a cap is much harder to explain in the context of monsters that have many independent potential drops.

With this model, TH would help to gain additional slots (original HTBC and many other pieces of content have a guaranteed drop slot and a possible drop slot). It would not help to obtain a rarer item within a specific shared slot. This model also explains the long-agreed upon stance that treasure hunter does not help on items like defending ring.

If each item is being rolled in a preset order until a cap is achieved, the item being checked first should gain substantially to the detriment of others when TH is involved. Using Urmahlullu again, since we know the items are roughly 1/3, they must be very common. If you start killing with TH9, you'd see 72%~ on that roll and whatever item is rolled first would skyrocket compared to the others.

I haven't observed this, but I haven't done a controlled test either. Wouldn't be very hard to do..

Anything that has actual data for it though...? I haven't killed a ton of Urmahlullu, but it seems like we've already got the answers for his drops, as we do for a ton of other enemies:

Eschan Stone (C)
Onca Suit (C)
Skormoth Mask (C)
Chastisers (VR)
Vulcanite Ore (C)
Eschalixir/Eschalixir+1 (4-5)
Eschalixir+2 (G)
Eschalixir+2 (?)

So...it seems like none of these items share a slot.

If someone can provide, for any individual NM, two items which "share a slot" AKA which cannot be contained in the same pool, then you can update the list above.

I think items that share slots are exceedingly rare and most of them are known about.

In either case, I fail to understand how any of this changes the way TH is applied, or its value. The only way I could see it being relevant, which I think was alluded to earlier, is that TH isn't as good because the slot is guaranteed to drop something anyway (ala KB), and the only "random" element is WHICH item it drops. Except...we've all seen pools with no item in it, despite the fact that someone in the party is missing something.

So...I understand the IDEA of it, but I'm still really not seeing the actual practical, real world evidence that it works this way, or how it changes the value of TH.

Once again bringing it back to CoD: Does anyone have any evidence that these items share "slots" with each other? We've certainly seen 2 Masques, we've seen 2 Loops, 2 Belts, 1 Belt 1 Loop, and probably every other permutation. We know that drops aren't guaranteed like Ou, Mee Deggi, or KB...so...why would we assume that sharing a slot changes the effects of TH?

Actually proving how it's coded is very difficult, but there are some things that are true:

-You always receive one R/EX drop from Urmahlullu. [I have farmed 50k beads for over 400 characters I've sold, my farming characters don't keep any of the rewards besides chastisers and the leech chars never cap on them so I'm not masking drops by owning them.]

-You receive a second drop occasionally, and sometimes it's the same item as the first drop.

-The order of the drop pool is always the same (the second r/ex comes after the vulcanite ore and before the eschalixirs). This isn't particularly meaningful on its own, but it does indicate the order the items are being rolled. If the roll was based on an item rarity and max quantity, they should be adjacent when they are the same item.

Again, looking at your model, this is very hard to fit. If they're all common, you have a .76^3 * .99 = ~43.458% chance of not receiving any R/EX piece. Do you have a potential explanation for this?


[FWIW, the same is true of basically all Gaes Fete NMs. The rarity of the second slot varies, WoC/Kirin have 2 guaranteed slots and a possible slot, and reisen T4 have 2 guaranteed slots and no possible slot.]

I'm not sure of the details of Urmahlullu, it's entirely possible that the categories on the page are incorrect and that it's actually more like 1 of: Stone, Suit, Mask (G) and then 1 of: Stone, Suit, Mask (C). Maybe the Chastisers shares a slot with either of those items.

The GF NMs certainly seem to generally have a guaranteed drop and then a chance at a second drop, that's a pattern I've noticed.

I think this is an abnormal model in FFXI in general and don't think it applies to most enemies. Even when it does apply, I don't think it's outside the realm of possibility that TH could still be applying to the items within that shared slot as well.

Let's take Lilith for another example. If you kill her without killing any Spite Wardens, let's say she has 1 drop slot, and it's shared among all her gear.

Maliginance Pole (2%), Malignance Sword (2%), Malignance Earring (2%), Malignance armor (1% each).

Go down the list, rolling each item until one of them hits the lucky number and drops. Then stop rolling the rest of them. You can, if you like, randomly decide the order in which to roll the items (to avoid a bias towards specific items). Or you can put less valuable items up higher on the list to allow them to cockblock the items you want to be rarer (we see this in many fights).

If you have TH applied to the mob, increase each of their %s by the number dictated by the chart above. Follow the same process.

In the case of rare items sharing a (guaranteed) slot with less rare items (Aurix, Dynamis [D] wave bosses, Fu, Ou, Mee Deggi, etc) I don't believe TH does anything, because that's a guaranteed drop, it's just picking which of the two to give you. It MIGHT still apply, but I don't think it does.

In the case of items sharing a slot, but the slot isn't guaranteed (like CoD and Lilith), I think there's a very good case to be made that TH still applies as normal.

The same model is used for all HTBC prior to selbina, vagary, delve, Ou. If you do those older HTBC on lower difficulty, you have sub-100% slots. Do you think they implemented a different set of drop code just for lower tier versions of the same fight?

Ultimately, it's a matter of interpretation and the only thing that 'really' matters is roughly how common a drop is. But, I find it hard to lend any credit to your model if you can't come up with an explanation for how it works in a pretty wide swath of the game's content.

I would say that TH applies to the chance the slot pays out, not the distribution of what's within the slot. This would be tough to test, since you'd need an item that's rarer than the alternatives like the chastisers setup, but not so rare that data collection was impossible.

I'll refer back to a few posts prior, though. If you look at the second possible drop slot for Urmahlullu, it's roughly equal chances of Eschan Stone, Onca Suit, and Skormoth Mask with a much lower chance of Chastisers. If you were to propose that TH effects the slot, you should see one of those 3 items leap far ahead of the others(whichever is rolled first).

Chastisers shares both slots, I've logged it appearing both before and after the Vulcanite Ore. I've also logged it dropping alongside a Skormoth Mask in both orders(Chastisers before Mask and Mask before Chastisers), which seems to prove that the items are rolled more than once. I'm pretty sure I've even seen a screenshot of a double pulse drop from one of those NMs.

The simplest explanation is that they created a slot definition for each gaes fete nm and linked it twice (once as guaranteed, once at a rarity tied to NM tier).

It is trivial to apply my model to every enemy in the game. Here you go:

Items have a drop rate. Let's look at a normal enemy, Ore eater:
Roll each, award the drops that hit their %.

OK, let's look at Aurix:
KB:
This logic can be applied to everything.

Baobhan Sith:Urmahlullu:
For every single one of these items, if TH applies, apply it to the % drop rate for that item. It's dead simple and works on every enemy in the entire game using the exact same, very simple, data structure.

Your implies that there are (at least) 3 different kinds of drop systems.
Or that back in 2003 when they first developed Tiny Mandragoras, they already implemented a slot system that treats items as being in different shared slots even though that system wasn't used by anything for 15 years.
Or that they overhauled the entire TH and drop system when GF came out

I'm not so sure I believe this. Vanishingly few enemies had shared drops before GF, and the ones that did were largely (exclusively?) 100% drop of one-or-the-other.

I really think it's a pretty simple process to go from the existing TH system to one that allows items to share a drop rate and with the implementation of Mee Deggi, KB, et al, it's already baked in, all you need to do is change the %s or the order the items are sorted in and BAM, you've got a whole "new" system of drops which works with your existing code and applies TH. For any given enemy, all you have to do is fill out the table with itemid, drop rate (or category), group number, and roll order. If group number is null, roll it normally, if it's not, it rolls with its mates. If roll order is null, it's randomized; if not, it rolls in that order.

You're still ignoring that under your proposed system, every monster with a guaranteed slot would receive a bias toward the first option to be rolled; this hasn't been observed. It should be glaringly obvious when using TH on Gaes Fete if your system was the one in use. A 24%(there is no 35% rarity) drop becomes 70.5% with only TH8. Since the other rolls don't happen if it's hit, this is recognizable within a very small handful of kills.

There don't have to be 3 different types of drop systems. Both equal(torques from BC60, relic armor from dynamis, salvage 25 pieces) and unequal(pixie/dring, fuma/strider, etc) versions of shared slots have existed since zilart. There just has to be a mapping of drop slots. One roll based on rarity level to decide if slot is awarded. Second roll to decide which item within the slot is granted.

And, honestly, it makes more sense to have each drop slot set up like that. If you have a structure that holds an item distribution and the monsters all link to the structure, you can change the structure to update all the monsters at once. Think about dynamis; each monster should link back to the same 2 drop slots(pre-toau jobs, and toau/wotg jobs) rather than copy/pasting all 20 jobs.

You also mentioned randomizing roll order, but this actually runs counter to your stated interpretation. If you randomize roll order for something like Urmahlullu, you make TH work against the rare drop. Using TH4:
25% chance that you start with the VR role (2.5%)
25% chance that it ends up second(0.36 * 0.025 = 0.009)
25% chance that it ends up third(0.36 * 0.36 * 0.025 = 0.00324)
25% chance that it ends up last(0.36 * 0.36 * 0.36 * 0.025 = 0.0011664%)
The overall chance of receiving chastisers ends up below 1% because the chance of your common items being hit earlier has reduced the possibility you'd even get to roll for them.

So, if TH effects the rolls in this manner and the distribution of the 3 more common items isn't altered, some TH values will considerably lower your chance of getting the rare item. To make it work the way you're wishing it would, you'd have to randomize part of the order and not the rest. It's vastly more complicated than a system where the slot is awarded then the contents are decided.

Now, if you did randomize part of the order but not the rest, leaving the rare items first.. you should be able to get Chastisers, Tartarus Mail, Omen Bodies, etc.. up to 8% by using TH12. I'd love to see some evidence of that. Even using TH8 should get you a cool ~1/20.

You can keep adding theoretical conditions (randomize only equal droprates, prioritize rarer items for first roll, whatever). Eventually your model can probably be made to fit data. But, the model of slot(rarity limited)->distribution(more flexible) easily fits all data in the game without jumping through hoops. Since we have no way of knowing, and the accepted knowledge seems to be that TH does not help on replacement drops, I don't see any reason your model is relevant or helpful beyond your own stubbornness.

You're still ignoring the thing I said. Your failure to understand/implement my system, doesn't actually make it worse, sorry. Randomize the order of the items. If you have 6 items with equal probability of dropping (Let's just say Kammavaca has equal chance of dropping Kali, Ichigohitonofuri, Midnights, and the 3 abjurations).

1/6 of the time, Kali is selected to roll first, so it has a 70.5% (your math) chance to drop.
1/6 of the time, Ichigohitonofuri is selected to roll first, so it has a 70.5% chance to drop.
....

What's the chance of getting a Kali, if you fight Kammavaca 6 times?


BCNMs don't apply TH and are completely different to enemy drops, Salvage didn't exist in Zilart, I have no idea what shared drops you're referring to with relic armor...? and shared slots are still easily explained under my system.

How about items that don't share a rarity? Under your system, you said

For Urmahlullu's second drop slot, do you have equal chance of getting Chastisers as the other items within that slot? I think it's clear that you don't. So this means that the SLOT has to have a rarity (let's say 6%) and THEN the item has to also have a rarity within that slot. How does this work, since you understand the system so well? Does TH apply to the slot? Does it apply to the second roll you made up?

Explain Odin's drops to me please. TH increases the chance of a slot dropping, but those items also have different rarities and appear at different frequencies. So, TH increases the chance of getting something, but doesn't affect which item drops when you do get something?

Is there any way to actually prove any of what you're proposing? What is the actual implication of it? Have you actually proven any of it?


The accepted knowledge is that it doesn't work on guaranteed drops. I don't believe it is at all widely accepted that it doesn't work on any drops that share a slot, which is an entirely different matter.


At some point in the system, you still have to list all 20 items somewhere so I'm not really sure how you think this is saving space. You can have a drop pool shared by all 100 enemies in Dynamis - Beaucedine, if it's appropriate; they all point to ID 123456, which has all 20 of those pieces of armor in it. No copy-pasting necessary, and when you add another job, all you need to do is add 1 row to that ID and you're done with the updates to every mob, now it drops from them. This is not complex or difficult to maintain, it's dead simple.


It's not vastly more complicated, I already explained the mechanism by which this works, did you miss it? If you want Chastisers to have a 1% drop rate all you do is make it order #1, and the others order blank. There, you're done. Chastisers have a 1% drop rate and all the others are equally distributed.


Where on the ultra rare graph do you see TH12 = 8%, I'm curious? The axes are labeled quite clearly and I'm not seeing that line get very close to 8%, but maybe my eyesight is going.

Except when SE added TOAU/WOTG job relic to Dynamis, they didnt "add one row to that ID", they made a new drop slot for those jobs relic armor. Whatever their reasons were, they deemed it easier to not touch anything with the original data and simply add a second drop slot instead of skewing the data in the original drop slot. And this wasnt maintenance mode XI, nor was this skeleton crew "working on XVI" XI, this was prime FFXI.

I wasn't aware of this drop behavior and haven't experienced it (getting an original job relic armor + a TOAU armor from the same mob?) but it's still extremely easy to include using my system.

The drop pool has two groups, group 1 and group 2. Group 1 includes all the original jobs, group 2 has the TOAU jobs. For each item in the list, roll to see if it dropped. Stop when you hit one. It's built right into the framework with no changes needed.

Add 3 new items to the list. For these rows, set the group number to 2. You're finished with the updates to the drop table.

All this grief over th discussions throughout the years because even when the dev team "explained" th, they didn't really explain it... I personally lean towards Thorny's slots model. As far as Maletaru's model goes I don't know why you would bother programming progressive sequencing of rolls when a single roll would handle each slot, once and done.

The real issue is with stupid stuff like this the dev team can't be accessible and clarify stupid stuff in a 20+ year old game.

My Treasure Hunter tldr:
It does stuff, sometimes, but not always, and never when you want it too.

Good job, you just completely altered the original drop rate of the previous relic.

Good job, you just completely failed to understand what I'm saying.

If both of them can drop from a single mob (IDK how it works, I've never seen this) then they're in different groups. Each group rolls only against its partners.

If they can't both drop off the same enemy, then no matter what the system is, the drop rate of old relic was reduced by introducing TOAU/WOTG jobs.

According to Thorny's theory, you roll for the slot, then roll for which item you got. The second roll would be 1/20 instead of 1/15. This is a lower drop rate for THF.

Yes, the slot contains a distribution (your 1-1000 split, if you want to view it that way). The slot itself has a rarity(one of the 8 categories SE outlined).

Yes, you can get one pre-toau drop and one post-toau drop from the same mob. Because they are in seperate groups/pools.

https://forum.square-enix.com/ffxi/threads/56550-Development-Q-A-March-2020?p=624996&viewfull=1#post624996
Read the chart from the official source if the graph is too hard to interpret. There's no way Chastisers or Omen bodies are ultra rare (0.10%). I regularly get 1-3 chastisers in ~46k beads (~240 kills) with no form of TH whatsoever. I could accept omen bodies being 'Super Rare' in the absence of clear data, but Chastisers are absolutely 'Very Rare' if governed by this system directly. I've gotten enough Omen bodies with and without TH to be confident they are not ultra rare and not effected by TH.

I cannot prove whether they do some complex sorting to decide what order to roll in to see the results we get. I don't see how any reasonable person would conclude that is a simpler implementation than having a secondary distribution for the item slot. If the end result is exactly the same, it doesn't even matter how it's coded.

For that matter, how do you make your model fit Dring/Pixie or Okote/impact? You seemingly accept that shared drops that are unaffected by TH exist, but don't accept that many modern drops are shared. That adds yet another layer of complexity on top of the sorting. If I'm a dev making a shared drop with your model, I just set okote/dring to Rare, put impact/pixie in the same group as Guaranteed, and set your flag to ensure they aren't randomized. Only problem is.. if it was done that way, TH would work on these items and the folks who poked KB to death with single wield thief's knife would've gotten their rings a whole lot faster.

Ultimately, what matters is whether TH effects the drops, and your model is just plain not in line with lived experiences of players. If you can boost 'Very Rare' items to 4.75% with TH8 and 8.25% with TH12, a reasonably small sample would show it.

My system is the absolute best one... Put TH even on things that don't get affected by it and go: "If it drops it drops..."

I've gotten enough Omen bodies to know TH doesn't change anything.






... because they only exist in Trove

Omen bodies from bosses could be ultra rare, but maybe only super rare (0.5% with no TH).

My money is on drop slots because SE *** hates THF and TH

and it also lets them control how slow people get things without any interference