Its hard to say how it works in instances, so i wont try to reflect from the following statement to how it works there.
But there are mobs in Blasted Lands called Shadowsworn Adpet's (Shadowsworn Adept - NPCs - World of Warcraft). They usually run around with a staff in their hands, as you can see on the screenshot of the wowhead link. Now if these mobs will drop any weapon, they will have it equipped already and you can see it in their hands instead of the staff they normally wield. This would lead to believe that, atleast outdoors, mobs will have their loot set once they spawn.
Reclaimer: I havent been to Blasted Lands since pre-TBC so it might have changed, but i wouldnt guess so.
I've seen this myself so I can second this. Additionnally there is the blue post that confirm that all loot is determined on spawn time (of the mob).
How this exactly works inside (raid) instances for mobs that spawn not at the creation time of the instance but later is a bit hazy but those who do those raid dungeon solo farm runs can certainly confirm that newly spawned in mobs do have loot later on. (Also respawned trash, of course)
Please note that the thread is a bit weird to read. Tseric first said that loot is determined on mob kill, was then asked about the Blasted Lands mobs and later added his "I could be wrong". Then he apparently checked with someone and finally edited his first post clarifiying that loot is determined before it dies. He didn't say it's determined on spawn time but that's what effectively happens.
I did this analysis for myself a little while ago, and some of my results don't agree with yours.
For Naj'entus I found that Boots of Oceanic Fury fell into pool B, making his loot split 6/8 instead of 7/7. Akama also seems to split 6/8, with Flashfire Girdle falling into your pool A.
Thanks! I will check this and if it match, i will correct my initiate post.
Originally Posted by Skulli
Well your example with rage isnt right atleast.
We got Chronicle of Dark Secrets, Cuffs of Devastation at the same time once.
Vut well that might explains why we never got cloth healing bracers from him in 18 kills.
Our main combo is deadly cuffs/reju bracers.
I disagreee. Looking at your DKP System you only got Deadly Cuffs & Chronicle of Dark Secrets at the same time. Every Rage Winterchill Loot fits perfect, except this kill, but only Chronicle of Dark Secrets is listed. So i can't check this kill.
The pool from which an item comes also determines which slot in the loot window it will take. In other words a pool B item will always show up as the 2nd item when looting the boss.
Common sense tells me that it is just a pseudo-random number at the kill or spawn time, and has nothing to do with raidleaders etc. But who knows? Maybe there is also a check that decreases the chance of the same item being drop the next week as the week before?
Atleast it should be quite easy to determine if it's the Raid Leader or the Master Looter setting the loot by running it though all the samples we have.
Common sense tells me that it is just a pseudo-random number at the kill or spawn time, and has nothing to do with raidleaders etc. But who knows? Maybe there is also a check that decreases the chance of the same item being drop the next week as the week before?
Nope there isn't. It's a computer generated random number (which we do not know how it is constututed), that eventually determines what loot a mob will yield when killed. And that happens when the mob is spawned, as you can read about in the previous post of me in this thread.
Atleast it should be quite easy to determine if it's the Raid Leader or the Master Looter setting the loot by running it though all the samples we have.
Not so easy if you calculate the amount of samples (of the same experiment, ie the same target mob) you'd need to have to get statistically relevant dataset.
People often fail to realize that he gave the answer in the second paragraph already. Mobs have loot tables, the loot they drop is therefor not entirely random, meaning something out of all items in the game, but limited to that loot table. That's what he hints at. But people read his statements as if he said there is this magic hidden thingie which we don't tell you that influences what you'll get.
Loot is randomly selected from the loottable from the specific mob. And randomly selected means using a PRNG which is seeded at some unknown time by some unknown value. If there is truely a way to influence this mechanism as a player, it's most likely only the random seed and there probably only a factor of it and not the seed in its entirety which means in normal speak that it's most likely useless even to try.
I personally have a deep suspicion that Blizzards' implementation and/or use of the PRNG has some weird bias now and then, but even if I were right, it would *still* be completely random.
Oh and changing master looter never had any influence at all, because for one, the mobs were already spawned once you're inside the instance and for the other, we tended to raid with FFA where there is no masterloot set at all (which could be viewed as the 26/41th setting of it, though).
If anything influences the seed or PRNG that is ultimately responsible for your Kael'Thas loot, it happens right when the first person enters the instance and spawns it with all mobs contained.
You can sometimes see that the loot is generated when the mob spawns and not when it dies. For example I have seen shade of aran holding a big sword and everyone commenting on it. Sure enough on death there was a green sword in his loot which when equipped had those same graphics. I recall something similar in burning steppes with mobs working around with the non standard graphics since they were equipping an item in thier loot table.
Hmm, if I remember correctly, there have been people reporting that the same bosses dropped different loot after server crashes and rollbacks and when the mob respawned after death (Moroes, Lurker..) due to a bug.
So "spawn" in this case would mean the actual spawn of the game object (the mob) and not bound to the instance id. This would also mean that a soft-reset, also resets the currently assigned loot, since a soft reset probably frees the server memory and removes the instanced objects, which are newly instanciated when someone enters the dungeon after the reset.
It makes perfectly sense if you look at it from a programmer's perspective in object oriented programming, so the constructor of the spawning mob could already assign the loot.
I recall hearing the story of a GM that was fired for revealing the loot drops to a BWL raid before they got to the boss.
However, it was 2nd hand and quite a while ago.
I think that is right. We had a Pumpkin Bag drop back in Naxx from a trashmob after the halloween event. The only possible explanation for this is that the Instance was created (and the mob spawned) when the event was still in progress and his loot was determined accordingly. (Instance created Wednesday, Event ended Friday, Pumpkin Bag dropped Sunday).
It took our guild until mid-naxx to get our first thunderfury in guild.
Other guilds on our server had them all over the place.
Kept thinking how nice that proc would have been for BWL/AQ40.
We've pretty much accepted the fact we won't see a single warglaive until the next expansion. But thank god we get that haste healing cloak from illidan!
There is an old thread when we first figured this stuff out for all the MC/BWL and AQ bosses but I cannot seem to find it. I think it was some time in 2005 though so I doubt any of the data in it is still relevant.
But to comment on your thesis there are 100% definitely loot pools. We had people from dozens of guilds digging through their dkp/loot trackers and no one could find exceptions to the rule.
Rare drops(war glaives) are in an additional loot pool and represent an additional loot when they drop.
Loot inside the pool does not need to have the same chance of dropping (though recent dungeons it seems like they maybe do) Back on nef each pool had 4 20% drops and 2 10% drops, I remember that pretty clearly.
The set items will just pick 1-2 tokens from the same even distribution pool.
I think that is right. We had a Pumpkin Bag drop back in Naxx from a trashmob after the halloween event. The only possible explanation for this is that the Instance was created (and the mob spawned) when the event was still in progress and his loot was determined accordingly. (Instance created Wednesday, Event ended Friday, Pumpkin Bag dropped Sunday).
That theory stems from a Something Awful thread ("Ask a WoW GM").
As previous posters have stated, killing Moroes twice in the same Kara clear can yield different loots. Hell, the opera event isn't even determined until you talk to the doorman. We've wiped intentionally and switched the raid leader just to get a change to a particular event. Loot is either on spawn or death, and really, who cares.
Also, there's not a super secret way to influence the loot dropped. It's all random.
Anyone with unlucky streaks of loot that think they can somehow influence it is fooling themselves, because you would never know if you were actually affecting it or not.
No offense, but the discussion on whether loot is generated at instance load, mob spawn, or mob death is all kinda pointless, a la Schrödinger's cat [0]. Unless Blizzard's programmers are complete morons (and learned nothing from EQ or Diablo 2), the loot isn't going to be available to the client until the mob dies. Thus, there will be no WoWView for finding rare spawns or resetting instances to get that coveted rare drop.
That said, the discussion on mutually exclusive loot (i.e., figuring out the exact loot subtables) is interesting and can actually be figured out with enough statistics from different guilds.
[0] If you're feeling philosophical, call it a "loot waveform" that gets collapsed into epix when the mob's ghost is released to the graveyard on death. :-)
I'm sure i'm completely wrong, but it would seem dangerous for the loot to be decided and loaded in before a boss is actually killed. I seem to remember talk of a program in Everquest that would tell you what every mob in an area was loaded with.
ShowEQ.
It worked in one of two ways. The first was running on the same PC as EQ, the second running as a packet sniffer on a second computer on your local network (In this mode it would read the network traffic sent, decrypt the packets. Since it was a on a second PC and only reading traffic it wasn't detectable in this mode).
I tried it out once. Pretty impressive. It would display a map of the entire zone you were in. Since Sony made the mistake of broadcasting _everything_ in the zone to each person in the zone you'd see a dot on the map for every NPC and item. (Imagine being able to see every herb in SMV without having to move and every rare spawn [and rare spawns that drop good items])
It also took advantage of a second mistake that Sony made... NPC drops were precomputed at spawn time and equipped on the NPC model. You could select an NPC across the zone and it would show you weapons/shields/etc that it would drop.
These are reasons why WoW only informs your client of NPCs & Items within a short visual range of your character and never shows equipped items on the NPC model.
Edit:
W.R.T. loot determination @ npc-spawn or npc-death: It makes no difference from the client perspective and is impossible to test (You'd need control over the server's random seed @ mob spawn & mob death).
People going crazy to find the "secret sauce" of switching the raid leader, mesuring the moon phase or whatever need to ask themselves this.
You can get an effectivelly random # by looking at the 1/100ths of second on the computer's clok after you requesting one or simply running a counter as a background process that goes up to 1000, resets and does it again really quickly. Want a random? Call Process(WhatAreYouAtInYourCount). Take the number and use it.
Or you can go through some insane machination that is the function of a raid leader, the number of paladins dancing in Ironforge, how many AV victories alliance had today or whatever. Nothing makes sense at all except for a variant of the first one.
The "loot pools" explain a lot about the so-called extraordinary frequency of same drops. If there are only 4-6 items in a loot pool, seeing the same one over and over is hardly exceptional from a chance perspective. If there are two dice in your hand, rolling a pair of 3s doesn't make you go "ZOMG TWO THREES THESE DICE ARE LOADED."
And while everyone who understands the difficulty with computng truly random numbers has a point about the lack of them, it's irrelevant here. Blizzard doesn't need an exceptionally good random number generation for most loot to remain random. It needs an incredibly basic one. There are so many dice rolls, so many kils, so many everythings...
The reality is that even if there were a massive bug that resulted in streaky loot, you'd never come remotely close to seeing it. You are simply not privy to enough dice rolls.
The notion that the game somehow is secretly storing the dice rolls from independent raid IDs (i.e. your previous weeks' raids) and getting "stuck" in giving you the same loot is absurd. Why would they do this? Why wouldn't they simply re-seed the RNG each week with an obscenely simple method of approximate a die or pair of dice in your hand? Well, of course they do the latter.
Period.
And everything else is a conspiracy theory of the highest order. When you're researching Schrodinger's cat, take a look at Occam's Razor.
I always find the discussion interesting but mostly because it's an excellent example individuals not understanding what "random" really means along with not understanding how to code something that tries to emulate "random"
Anyway, this thread just took an off-topic turn so I'll drop the comments about random loot with this old Tseric quote:
I'm convinced that the entries in the loot tables are emptied or become less likely when items from those tables drop. The percentages are either reset or recharge over time (with say, raid tables being reset much later in the week, blue tables hourly, and green tables say every 15 minutes). And its not just group or raid-wide, it would be server wide.
Back before everyone hit 70, you could farm mana's at a reasonable rate in Netherstorm from Kirin'var, but eventually the drops would slow down. You would almost always get generous streaks of motes of mana from the mana thistle in the zone after this occured, and when that slowed down, the ghosts in Kirin'var would be back up to their original rate (or the mana worms and elementals would have them in generous quantities for a while).
Motes of Fire? I don't think their drop rates were really nerfed as much as people think, I believe they are shared with mining nodes, and when BC came out the mining nodes out there were picked clean for weeks as people leveled up their tradeskills. The first few people in elemnetal plateau got riduculous amounts of fire, which I believe were due to the motes getting pushed off the mining node tables and accumulating onto the fire elmental's tables.
This would make sense, to promote exploration (spread people out by making the resources more optimal to acquire that way), and as an anti-farming measure. Farming would work but the drop rates would be reduced.
I've done a lot of the grind quests on an alt recently, and some of the more hellish ones when BC came out dropped a fair bit faster, which is either due to blizzard tuning the drop rates (possible but I really think they don't bother unless its seriously broken) or less competition for the mobs by players.
I'm convinced that the entries in the loot tables are emptied or become less likely when items from those tables drop. The percentages are either reset or recharge over time (with say, raid tables being reset much later in the week, blue tables hourly, and green tables say every 15 minutes). And its not just group or raid-wide, it would be server wide.
No.
Please read what has been posted. From a design/programming perspective this is the equivalent of building a car with its wheels on the roof and the engine in the driver's lap. It is not how the system works. It is random.
Animal brains are designed to find patterns. We therefore naturally see patterns where there are none. Rabbits in clouds, lucky numbers and yes, imaginary loot seeds.
That theory stems from a Something Awful thread ("Ask a WoW GM").
As previous posters have stated, killing Moroes twice in the same Kara clear can yield different loots. Hell, the opera event isn't even determined until you talk to the doorman. We've wiped intentionally and switched the raid leader just to get a change to a particular event. Loot is either on spawn or death, and really, who cares.
Also, there's not a super secret way to influence the loot dropped. It's all random.
That doesn't affect the prevailing "determined with seed based on when the constructor is called" theory, though.
If people really really want a conspiracy theory involving loot distribution, I'll give you a good and believable one: The rumors about loot distribution, the average player's ignorance about their class, and the .999... != 1 phenomenon all have the same source!
People are bad and math. Moreover, they don't realize how bad at math they are. Anyways, here's an industrial-strength open-source PRNG algorith: Mersenne Twister
Just for reference, there are around 10^80 to 10^90ish particles in the visible universe. A period of 10^6000 is cryptographically secure long after Galaxy of Starcraft MDCLVXI expac 5 comes out.
Well your example with rage isnt right atleast.
We got Chronicle of Dark Secrets, Cuffs of Devastation at the same time once.
Vut well that might explains why we never got cloth healing bracers from him in 18 kills.
Our main combo is deadly cuffs/reju bracers.
No, you haven't. I just checked your DKP page and went through your Rage Winterchill history. Incidentally on 9/2/07 your DKP admin recorded three drops for Winterchill, which I'm assuming was an error.
edit: looks like Kapuras already brought this up, sorry.
A quick browse around for Solarian (including our own data for tonight)
Table A - Star-Soul Breeches, Worldstorm Gauntlets, Star-Strider Boots, Girdle of the Righteous Path, Vambraces of Ending
Table B - Ethereum Life-Staff, Wand of the Forgotten Star, Trousers of the Astromancer, Heartrazor
Table C - Solarian's Sapphire, Boots of the Resilient, Void Star Talisman, Greaves of the Bloodwarder
Atm this is more then likely completely inaccurate, as the sample was quite small (11 guild killlists). Our loot tonight was Girdle of the Righteous Path, Wand of the Forgotten Star, Solarian's Sapphire. Either way, the old lists would need to be reworked for SSC/TK in regards to the 3 drop change
Ye can agree its inaccurate due to Table A: Star-Soul Breeches and Star-Strider Boots dropped on our last kill of her which was thursday, 3rd loot was Void Star Talisman.
I think with the 3 items the old tables may be rather useless now so I be very interested in seeing updated tables, if that is possible.
I'm convinced that the entries in the loot tables are emptied or become less likely when items from those tables drop. The percentages are either reset or recharge over time (with say, raid tables being reset much later in the week, blue tables hourly, and green tables say every 15 minutes). And its not just group or raid-wide, it would be server wide.
Back before everyone hit 70, you could farm mana's at a reasonable rate in Netherstorm from Kirin'var, but eventually the drops would slow down. You would almost always get generous streaks of motes of mana from the mana thistle in the zone after this occured, and when that slowed down, the ghosts in Kirin'var would be back up to their original rate (or the mana worms and elementals would have them in generous quantities for a while).
Motes of Fire? I don't think their drop rates were really nerfed as much as people think, I believe they are shared with mining nodes, and when BC came out the mining nodes out there were picked clean for weeks as people leveled up their tradeskills. The first few people in elemnetal plateau got riduculous amounts of fire, which I believe were due to the motes getting pushed off the mining node tables and accumulating onto the fire elmental's tables.
This would make sense, to promote exploration (spread people out by making the resources more optimal to acquire that way), and as an anti-farming measure. Farming would work but the drop rates would be reduced.
I've done a lot of the grind quests on an alt recently, and some of the more hellish ones when BC came out dropped a fair bit faster, which is either due to blizzard tuning the drop rates (possible but I really think they don't bother unless its seriously broken) or less competition for the mobs by players.
I'm pretty sure I remember reading a blue post saying that they adjust the amount that motes drop to keep the prices at the AH at about the level that blizz thinks that they should be. This would have nothing to do with loot tables being emptied. I'll go see if I can find it somewhere.
10/12/07, 4:39 AM
Similar Threads
