Sorry, I should have been more clear. I thought your testing was on a server that came up past BC's release and thus likely had no players who did AQ40 content.

While it's not what you wanted, I thought this had been proved before TBC came out. Searching "blasted lands mob drops weapon" on google returns multiple examples:

http://img191.imageshack.us/img191/2280/epic14vj.jpg
http://img191.imageshack.us/img191/6671/epic27wg.jpg

ImageShack - Hosting :: brightewoodstab3oa.jpg

http://img150.imageshack.us/img150/1...60018392ze.jpg

http://img511.imageshack.us/img511/3...tstuff1ox4.jpg
http://img252.imageshack.us/img252/5...tstuff2pf3.jpg
http://img512.imageshack.us/img512/8...tstuff3uf5.jpg

On the Illidian Kill Despawning point (in case it helps), and the loot generation.

This happened on our first kill last Thursday night. wipe in progress from 7%, manage to get him trapped, and according to combat logs, a fireball's ~30 damage dot component tick from a player killed him after we reached a 100% wipe. (unless he was actually already dead, and it was just still broadcasting actions). He went through the animations, collapsed, and then vanished about 5 seconds later. Contact from a GM took much shorter than expected, and loot was distributed through the mail about 4-5 hours after the kill.

One of the comments that stands out from the GM was: "Hope you get a warglaive!"... which indicates that he didn't know what dropped. Although he could be a peon-level GM that doesn't have access to the loot tables, it is also possible that they had to 're-generate' the loot table to identify what dropped.

This then would imply that they could, in-fact, regenerate the table. If the table is originally wiped (and not saved somewhere) when the corpse despawns, then it would help to confirm that the information in the loot table is based on a factor that can be artifically recreated to recreate the exact loot.

Of course, that is alot of 'if...then' and 'implies', and might simply be that they couldn't re-create the original table, but instead just generate a new table. Which would also indicate that their comments that they have to find out what dropped is misleading.

Sorry, this isn't a survey and is thus inconclusive. How come Kaz'rogal doesn't drop Cataclysm's Edge every time? Excuse me while I be snarky.

Anyway, a point I meant to convey in my previous post is, I hope that people doing tests realize that as far as we know, unlocking an item requires the boss to have dropped dead, not merely just spawned. There seems to be other factors that can 'unlock' an item. The PTR however seems to be relatively pristine, and loot so far from Sunwell could only be linked after the various bosses were dropped. Exceptions being the pieces that were available from the trade-in vendors.

I'm sorry, but that's exactly what I'm talking about. All of those pictures are entirely without context of all the other mobs around. There is no attempt to disprove that mobs that wield "Glowing Brightwood Staff" may not drop "Glowing Brightwood Staff". A statistically valid sample of data is not provided. You're entirely welcome to think what you want but it's all religion and superstition until proven one way or the other methodically. Anecdotes are the exact opposite of a random sample or survey.

And posting that somebody is wrong is the exact opposite of spending all the effort of trying it yourself.

I said they are wrong to make that kind of assertion without providing a reasonable body of evidence, not that they are wrong that the mobs do not in fact work this way. It has not been proven or disproven and I do not see a reason to waste my time doing so since I am not making a fantastic assertion. Countless individuals are willing to bet on all sorts of things (like "loot seeding") affecting their game experience but the only thread I've ever seen where any aspect of WoW's loot system was analyzed methodically in order to find any sort of pattern was a recent-ish one that showed that drop tables appear to be segmented rather than one uniform table on the mob where N items are picked randomly among M.
Link. A. Survey.

You know what -- forget it. Have fun with your hearsay and anecdotal pseudo-science and arguments about how the loot tables are "seeded" and how to influence them. I'm sure you'll be guaranteeing your Warglaives drops in no time through ingenuity

Considering the mobs have a default weapon graphic that changes in the event of a weapon drop (Many images linked in this thread), it seems there is enough information presented for you to at least not troll for the sake of trolling.

It. Is. Gone.

It's from the old EJ forums. It's that old. It has been fact for a long time now. It's nice to challenge things that have been held fact for a long time, but as with the case of anyone who wishes to disprove something held as fact, it is up to the challenger to provide the proof.

Back on the topic of forcing instanced loot to spawn. Seems the easiest way to check would be to be on a server that never formed an Atiesh and try to link a [Staff Head of Atiesh]. If it crashes you, then head into Naxx. Since it is a 100% drop rate if you have the quest then KT is guaranteed to have it on his loot table. If you can then link it without crashing, that would be some strong evidence towards loot appearing on the server as soon as it is on the mob's loot table rather than when it is actually dropped.

Well if the mobs drop a weapon only when wielding a non-standard weapon graphic fine, it's been proved. So it's easily demonstrable. Someone give us coordinates and mob type. I presume the non-weapon-dropping mob will be wielding nothing.

Unfortuntely, said mob also will drop weapons, so we're going to have to discard his drops. The weapon-wielding drop mobs will have to drop weapons of the "wielded" type at 100% or else I doubt we're see much of anything other than the attribution bias. Should take 10 people about 15 minutes to verify.

By the way, Habanero, "anecdote" isn't a bad word. It's the singular of data.

All that said, there is absolutely no way whatsoever this scheme is going to yield warglaives. Knock yourself out trying, by all means, but it's not going to work. I refer you to the EJ home page and recommend you keep killing Illidan.

Huh?

anecdote - Definitions from Dictionary.com

an·ec·dote (ān'�*k-dōt')
n.
A short account of an interesting or humorous incident.
pl. an·ec·dotes or an·ec·do·ta Secret or hitherto undivulged particulars of history or biography.



The singular of data is datum.

On topic, a bit of a wrench can be thrown in the idea of being able to use item linkability as a basis for whether or not the item is going to drop.

I'm on Echo Isles and the highest guild clears ever here are up to Gorefiend, but I have been able to link every single piece of loot in BT for quite some time. That even includes a period where the one guild that could enter BT at the time had left the server and thus no one would have been in BT at the given time.

I'm not sure if this lends towards that theory being incorrect or to my observation being skewed somehow, but for what its worth, when I was on Thunderhorn, we didn't have any guilds enter BT for at least the first few weeks, but the loot from Naj and Supremus could be linked before then.

On a further note, I don't think that it much matters how the loot is decided. If, for instance, we were to determine for sure that the loot was decided upon the boss's spawning as opposed to his death, we would still not be able to effectively figure out from the RNG what numbers would be necessary to cause a glaive to drop. Why? Because there is only one ID for the mob, and that ID doesn't correlate to one item, but rather to a grouping of items, e.g. ID 345678910 causes both Glaives and the staff, not just any single one of those items. We would have to compile an excessively large collection of every boss ID and what items dropped with that ID, and then establish the similarities between all of those IDs until we got a table that might, keyword might, give us some guess as to when a glaive will drop. Even given that, we still run a low chance of being able to manipulate the numbers enough to cause a specific drop.

The item cache is by server cluster, not just server. There there is set of hardware that house all of our servers and they draw from the same database. Malfurion experienced this with the Warglaives when one dropped in our cluster before anyone in our server had killed Illidan.

The clusters are posted somewhere, but you can infer it from the published Battlegroups, as the Battlegroups are composed of groups of server clusters, which is why you get some imbalanced Battlegroups like Shadowburn which are like 80% PVP servers.

Long story short, you can't necessarily use the item cache for anything, as there are more variables than you can control outside of your server.

Correct, I believe this was changed when originally it was by server and people were linking warglaives in arena games causing the opposing team to disconnect when clicking the link.

The things you're stating are true, and have been previously mentioned in the thread.. you can link items seen in a battleground that have NOT yet dropped on your server, so for all intents and purposes, once something has been seen by your battlegroup it should be safe to link on your server.

This speculation you're referring to is all based off the simple notion that before Illidan had died on any server, the Warglaives had been datamined, but were NOT linkable until supposedly some point during the night of Nihilum's Illidan attempts. Now, since Illidan despawns between attempts, the theory is that upon respawn he has re-rolled his loot tables, and would yield different loot than if you had killed him the previous attempt. This seem to mean that at some point during the Nihilum's attempts, one of the Illidan spawns had rolled to drop a glaive, and that THIS is what made the glaive linkable, even though Illidan didn't die that attempt.

As far as I can tell, It remains to be proven that spawning a mob makes the items he will drop linkable if they have never dropped before, even though the game has already decided to drop those items. It seems more likely to me that blizzard would want to pull RNG values at the time of mob spawn, but that these values aren't actually applied to a loot table (and thus determine a corresponding itemID) until mob death.... meaning that the loot has been technically predetermined, but you can't trawl for what the boss is going to drop by simply trying to link items that haven't been seen yet. This is, of course, totally unsubstantiated and largely unprovable until never-before-dropped loot arrives (sunwell) and we can attempt to link IDs that we know for certain have never been seen. Of course, this raises the question that if indeed this hypothesis is right, I'm then at a loss as to how to explain the linkability of warglaives before Illidan had died, unless it was through some Devloper presence/testing to ensure the item wasn't bugged if it did indeed drop.

Does this help people get glaives? Nope. Though it's still at least an interesting discussion in the spirit of trying better to determine at what point loot "exists" on the server. (For that matter, it was my understanding that the server's item cache is only purged during patches, not merely for reboots/maintenance, but perhaps someone has more insight on that. From my experience, post-patch days are the only days I have historically had issues linking items that I hadn't "seen" before.)

I remember a thread here a while back where someone did testing and loot seemed to drop in "sets". For example if A drops, then B never drops on the same kill. How would this idea tie into that one?

I'm not entirely sure you can bet the farm on the GM's word. It's possible he said "Hope you get a warglaive!" while he was typing in a query for the system to retrieve the loot information (i.e. he said that just to pass the time) OR he said that in context of "I don't know what Illidan dropped, server doesn't know what he dropped because we just purged the loot logs, and I'll just roll the dice and pick x amount of items from Illidan's loot table for you to have."

I have one anecdote too:

I was in heroic Mech a few months after TBC released, we had just killed one of the minibosses that carried half the Legion's Cache key. One of our members d/c-ed just as we killed him, and it was set to Group Loot. Unfortunately, our d/ced member was the one that had loot rights on the guy.

So I filed a GM ticket, GM answered while we were killing Panthaleon. GM took a few minutes to "verify" what happened (he/she used that exact word, I have no clue what/if they were verifying). GM tells me 1x Mote of Shadow dropped and half of the Legion Cache key.

Here's the interesting part: The GM tells me to log out, so he can put the items in my bags. I log out, and I stay at the Character Select screen. I'm on Vent, so I can hear my friends saying that they see my character logging back in (the GM controlling my toon obviously), and then logging back out. I log back into my character, lo and behold, I have the "lost" items in my bags.

I'm of the opinion that the system stores a list or a log of some sort of what dropped and what was looted. I believe heroics work more or less on the same system as raids, so more than likely what happened in my case would apply to 10man-25man raids as well.

It seems rather a waste for a system to store a seed number for the loot, and then have a GM "regenerate" the loot if needed. How many 25man raids are going on daily? Simply storing a short list of drops wouldn't waste any space.

--------------------------

Actually, a new thought just hit me as I was typing this. Storing a list of drops wouldn't waste ANY space.

Think about it. We know from past raids that loot is created at dungeon spawn time. As soon as you kill a boss mob, the loot from that boss is available immediately; there is no "loot generation time" from the time you kill a boss to the time you view the loot. That implies that the server is storing that boss's drops by each item's name, not storing a seed number and then generating the loot when the boss dies.

What harm would it do to remember that boss's drops until the next raid reset, or server maintenance, or "garbage collection" or whatever? After all, the server has to do that anyway for the bosses you didn't kill, just because the server doesn't know if you'll be back to kill the next bosses in the future.


You're taking it to an extreme. Non weapon dropping mobs will wield their DEFAULT weapons. Dwarf Riflemen will have dual axes and/or a gun, BE guards have that vertically-long shield and a sword, Nethergarde casters will have a staff, etc. I honestly can't think of a mob that doesn't wield some sort of weapon (excluding maybe ogres in various dungeons).

Pretty independent notions. If a mob had 3 loot "pools" from which to pull, it would ostensibly request a random number for each pool to determine what item drops from that pool. As it pertains to warglaives, its pretty much irrelevant since they occupy an additional loot slot.

Theoretically, I suppose you'd have a roll to determine whether you get the extra loot slot, and then upon success, another roll to determine MH or OH glaive. Though certainly they could determine both the "extra slot" and "MH vs OH" in a single roll, but I'm not sure it really matters either way.. its just a matter of methodology.

The "loot pool" idea you're referring to is pretty much established and accepted at this point, but doesn't really tie in to the discussion at hand except in the sense that it implies that each boss spawn requires more than one RNG poll to determine what drops from each loot pool.

Each raid boss has several loot tables that they draw loot from. Illidan, for example, has 4.

Tier6 Chestpieces (which is rolled on twice)
Loot 1 (weapons/shields/maybe trinkets)
Loot 2 (random armor)
Warglaives (None/MH/OH)

While that may matter if you're tracking a pattern between GUID and items he drops, it shouldn't have any bearing on an item being linkable prior to the boss dying.

Edit: Too slow. (lousy work getting in my way)

I had the opportunity to talk to an ex-GM today and so I asked him a few question regarding this.

First I asked him if every GM is able to see what a bossmob is going to drop or if only selected "higher level" GMs can do that. He told me that every GM can look into mobs and what they are going to drop. He said something along the lines of that this is only a specific visibility setting on their specific clients. Whatever that means.

I asked him then if he'd ever looked into a boss after he respawned to see if the loot changes and he said he never did that and thus doesn't know.

We talked a bit more and I mentioned the blasted lands mob that display the weapon they are going to drop. He said then that this is an old idea that's been around for a long time and that should finally appear after 2.4 "officially". He also specifically meant that this is not going to be restricted to weapons only but there are some limits or dependencies which I didn't fully understand. He said that there will be individual looks on mobs (ie the mobs of one type look more diverse) and they then drop the armor / weapon in that specific look.

@Habanero, I've confirmed the behaviour myself about 2 years ago. It's not a study but I'm not going to spend a few hours on a youtube video for your pleasure. I've seen it myself and lots of others did. If you can't accept that, that's fine. Stop arguing though about this minor tangent to the discussion what essentially later has been confirmed not once but twice by blue posters that loot is being determined on mob spawn. This is a tangent because it has nothing directly to do with the OP's posting but I brought it up because it's still often contradicted out of a gut feeling.

Are we even sure linking items is consistent enough for this? I can think of times where the ability to link an item did not seem to match the criteria.
Before arena talk was nerfed people were linking items from illidan to get other people to disconnect. However I was able to link the item myself without disconnected, but clicking on the item in arena caused a disconnect. Also my own server had no killed Illidan at all at the time.
Back pre-bc I had a lootlink with Atiesh of the priest version, which I was unable to link on my own server, but I was able to link the warlock version. Other people could link the priest version with no problems. This happened with a bunch of other new items as well.
Im not exactly sure how the game determines what is available for you to see in the item cache. I would assume somehow it is tied to battlegroups who have to see other people in BG/arena. The sunwell patch might be a good time to test this when it goes live.