Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Does group leader affect drops?

22 views
Skip to first unread message

Ham Pastrami

unread,
Dec 5, 2005, 11:31:09 PM12/5/05
to
Despite some other questionable design decisions on Blizzard's part, I
cannot bring myself to believe that they would be so dense as to tie
instance drops to who the group leader happens to be. And yet I keep hearing
this rumor, unsubstantiated as yet, that the group leader has some
determination in what kind of loot drops. Does anyone know of official word
on this, to either confirm or deny? Failing that, does anyone know of an
article, official or otherwise, that discusses this topic in depth?


Brian

unread,
Dec 6, 2005, 12:23:01 AM12/6/05
to
"Bother!", said Pooh, as he read Ham Pastrami's latest post to
alt.games.warcraft.

Blizzard has repeatedly denied it.

Fact is that their random number generator is *extremely* prone to
"streaks". There's too much evidence, from too many guilds. Like my own
guild, that got 4 out of 5 drops the same on two successive kills of
Lucifron/Magmadar. Anecdotal evidence from many sources seems to indicate
that changing the raid leader will sometimes be enough to re-randomize the
drop tables.

All random number generators require a "seed" value in order to operate.
In the case of WoW, the best theory I've seen is that the seed value for
loot is very strongly influenced by the instanceID. And the instanceID is
generated at least in part by some sort of ID number attached to the
character that leads the group/raid. Changing raid leaders is not always
sufficient to change the instanceID, though, in the case of "saved"
instances such as Molten Core. There was a lot of issues a while back with
improper lockouts and people not being able to enter the correct instance,
because of the instanceID not being reset properly, and being linked to the
raidID.

Short version, Blizzard denies that the raid leader affects the loot
tables. Strong anecdotal evidence suggests that they are either lying, or
once again their player base understands their game better than they do.
No one has, as of yet, done any sort of controlled study to decide the
issue with certainty, and I'm not sure that the necessary degree of control
over the variables is possible, from the player's end.

Brian
--
ICQ#: 68214833 | AIM: LineNoise54
.
Allow me to introduce my selves.

Devast8or

unread,
Dec 6, 2005, 1:52:06 AM12/6/05
to

"Brian" <brian...@pobox.com> wrote in message
news:cj7ap1h6gt1gesqq1...@4ax.com...

(snip)


>
> Fact is that their random number generator is *extremely* prone to
> "streaks". There's too much evidence, from too many guilds. Like my own
> guild, that got 4 out of 5 drops the same on two successive kills of
> Lucifron/Magmadar. Anecdotal evidence from many sources seems to indicate
> that changing the raid leader will sometimes be enough to re-randomize the
> drop tables.
>

(snip som more).

Very interesting, I've never heard of this in WoW (in D2 someone had a
similar theory about drops - same randomization code perhaps :)

One question though: You mention the drops being tied to thr group leader.
Do you mean just that the same group will get the same drops, or do you mean
that if the group is caster, there will be a greater chance of a caster item
(like a wand vs. 2H mace)?

Devast8or


Michael Vondung

unread,
Dec 6, 2005, 1:52:20 AM12/6/05
to
On Mon, 05 Dec 2005 21:23:01 -0800, Brian wrote:

> Short version, Blizzard denies that the raid leader affects the loot
> tables. Strong anecdotal evidence suggests that they are either lying, or
> once again their player base understands their game better than they do.

I hate to buy into this, but it is our experience also. We had hardly *any*
priest drops in MC until we changed the raid leader when we got a fresh ID.
Ever since the drop situation has massively improved. If we go back to the
old leader, we once again get only warlock and rogue stuff.

Then again, we get crap in Zul'gurub no matter who our leader is. *g*

M.

Rene

unread,
Dec 6, 2005, 2:46:01 AM12/6/05
to

In short: yes.

Raidleader = Group ID, GroupID = InstanceID. As soon as the first player
enters (spawns) a new instance, it appears that the loottable is somehow
chosen. Depending on the GroupID now, the table is either very well
balanced (can happen, these are later the players who claim this is all
hogwash and there isn't anything like that) or it can be slightly or
extremly tilted.

Up to this point Blizzard never really confirmed nor denied this but there
are lots of guilds who have special players who, when they open MC, can
force a very high drop chance on the legendaries. We for example have one
person that causes a huge amount of warlock drops (also 5 out of 6 times
earthshaker at magmadar which would have a normalized drop chance of only
~20%). We also have one other person that causes the perdition's blade to
drop twice in a row. We're still rotating to find some player who causes us
to drop things that we're missing up until now.

I didn't believe this first but we've experienced some very odd things and
since we're rotating, our drops are much more distributed instead of always
the same things over and over again.

If you're lucky, your standard leader has a good balanced table but if not,
you'll soon find out anyway ;)

As I said, Blizzard is very quiet about this and it IS being "exploited" by
many guilds/groups to get specific drops. There's no article about this I
believe and all discussions in the official forums are full off people who
diss' each other because about half say it's not true and the other half
believes in it. I've seen too many odd things before we started to rotate
that I'm very convinced that it is - unfortunately - all true. The
RNG-system for Raiddrops is majorly broken.

CU

René

--
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service $9.95/Month 30GB

ASKF

unread,
Dec 6, 2005, 2:58:15 AM12/6/05
to
Brian ytrede sig i <cj7ap1h6gt1gesqq1...@4ax.com> med
dette:

I recently did more than 10 runs in a row of RFC (was hgelping a mage
making a movie) and the boss drops changed a lot. Since we had the same
group leader every time, I tend to believe Blizz is right.

I'm also doing boss-runs at SM all the time, together with a warrior,
and the boss loot is having the same drop-stats as when I was
practically living there between lvl 37-42. It doesn't matter at all,
wether my shaman or the warrior is the leader. It doesn't affect the
loot chance either, if we bring a low lvl guildie along (to make sure he
gets the blue drops he wants/needs).

There might be differences between instances though, but I don't see why
the dev's should have used different algorithms.
--
Allan Stig Kiilerich Frederiksen
"When you try to change a mans paradigm, you must keep in mind that he
can hear you only through the filter of the paradigm he holds."
-Myron Tribus

valpozzo

unread,
Dec 6, 2005, 3:09:22 AM12/6/05
to

Rene ha scritto:

We started to disenchant warlock and hunter stuff, while shamans are
still waiting the #2 or the #3 earthfury from MC ... this is not
working "as intended" i suppose...!

Mel

unread,
Dec 6, 2005, 3:45:33 AM12/6/05
to
Michael Vondung a Ă©crit :


We've been reconsidering renaming our Guild to the Zulian Shields :-)
Randomized drops doesn't account for the amount we've had.


I don't belieive that drops are really random, not having seen the
Magister Headdress in more than 20 Scholomance runs and the skinning
knife in about the same number of UBRS runs.

And class drops in MC for us are at about 2 per person per class except
priests at less than 1 per person. And those damn shammy bracelets keep
dropping - stop Blizzard we don't want any more!!!!

Rene

unread,
Dec 6, 2005, 3:50:10 AM12/6/05
to
"valpozzo" <killbill...@hotmail.com> wrote:
> Rene ha scritto:
>
> > "Ham Pastrami" <nom...@dot.com> wrote:
> > > Despite some other questionable design decisions on Blizzard's part,
> > > I cannot bring myself to believe that they would be so dense as to
> > > tie instance drops to who the group leader happens to be. And yet I
> > > keep hearing this rumor, unsubstantiated as yet, that the group
> > > leader has some determination in what kind of loot drops. Does anyone
> > > know of official word on this, to either confirm or deny? Failing
> > > that, does anyone know of an article, official or otherwise, that
> > > discusses this topic in depth?
> >
> > In short: yes.
> >
> > Raidleader =3D Group ID, GroupID =3D InstanceID. As soon as the first
> > pla=

> yer
> > enters (spawns) a new instance, it appears that the loottable is
> > somehow chosen. Depending on the GroupID now, the table is either very
> > well balanced (can happen, these are later the players who claim this
> > is all hogwash and there isn't anything like that) or it can be
> > slightly or extremly tilted.
> >
> > Up to this point Blizzard never really confirmed nor denied this but
> > there are lots of guilds who have special players who, when they open
> > MC, can force a very high drop chance on the legendaries. We for
> > example have one person that causes a huge amount of warlock drops
> > (also 5 out of 6 times earthshaker at magmadar which would have a
> > normalized drop chance of only ~20%). We also have one other person
> > that causes the perdition's blade to drop twice in a row. We're still
> > rotating to find some player who causes =

> us
> > to drop things that we're missing up until now.
> >
> > I didn't believe this first but we've experienced some very odd things
> > and since we're rotating, our drops are much more distributed instead
> > of alwa=

> ys
> > the same things over and over again.
> >
> > If you're lucky, your standard leader has a good balanced table but if
> > no=

> t,
> > you'll soon find out anyway ;)
> >
> > As I said, Blizzard is very quiet about this and it IS being
> > "exploited" =

> by
> > many guilds/groups to get specific drops. There's no article about this
> > I believe and all discussions in the official forums are full off
> > people who diss' each other because about half say it's not true and
> > the other half believes in it. I've seen too many odd things before we
> > started to rotate that I'm very convinced that it is - unfortunately -
> > all true. The RNG-system for Raiddrops is majorly broken.
> >
> > CU
> >
> > Ren=E9

>
> We started to disenchant warlock and hunter stuff, while shamans are
> still waiting the #2 or the #3 earthfury from MC ... this is not
> working "as intended" i suppose...!

Our warlocks also are all either 8/8 or 7/8 while the most items our
pallies have is 2. Priest vambraces random-drop never dropped once. Since
August, when we started. Also during the first 6 or 7 MC runs, the instance
was always opened by the same raidleader save once when he wasn't there. We
had 0 rogues drops during that time.

The odd thing is that if you view the drops over all raids over all realms
it seems pretty evened out. But then again, if you give each raid only
drops of one class and every raid another class this also evens out overall
but sucks for the specific raid.

Rene

unread,
Dec 6, 2005, 4:00:50 AM12/6/05
to

There are differences. For example, in blue instances you can get the
shaman set as Alliance and the paladin set as Horde. The items are not
class bound so there is some use for them. The epic sets however, are class
bound. There is another mechanism that ensures that Alliance players never
get a loot table that allows to drop shaman sets while Horde players never
get Paladin set items in MC, BWL and from Onyxia (and also Zul,
supposedly).

Of course I cannot say if they use the same mechanism, I'd have to see the
code for that. I and many other are however seeing too many oddities and we
experienced that we can, indeed, influence in a limited fashion of what
drops and what won't drop. Much is speculation based on oddities that we
observed. It is almost impossible to perform a real scientific proof due to
the limited number of samples we're able to produce but even at this rate
the effect is very noticeable fast if you happen to be unlucky and had an
extremely uneven loottable up until you start to rotate.

By the way, there is already two different kinds of drop-systems. That out
of instances and that inside instances. Those out of instances decide the
drops on kill, while those in instances seem not to. Also remember that
drops already are player influenced. If you have a quest that requires you
to collect X, mobs of the target kind suddenly are able to drop that for
you, whereas they weren't before you had the quest and stop dropping it
after you completed the quest.

Hornet

unread,
Dec 6, 2005, 4:21:33 AM12/6/05
to
I'd have to agree Brian. Their generator is either hosed or raid leader
does affect it, or both. I've seen the same leader get the same drops
from several bosses for three weeks in a row at which point he was
fired. I've seen 3-4 raids get identical loot from the same boss within
minutes of each other. My current guild has never gotten a legendary
drop, and 1 perd blade, despite the fact they are one of the leaders on
the server for pve and have been farming MC for a very long time. There
is another guild on the server that got eye of sulfuras on their last 2
runs, and has been on Rag for under 2 months(anyone from ATF is welcome
to correct me). There's probably a lot more to it than we know.
Personally I think there's a lot of factors that make the random number
gen, not random at all. Who's logged in after a restart and had some
really kickass drops, only to see nothing once the population gets back
up?

Brian

unread,
Dec 6, 2005, 4:21:16 AM12/6/05
to
"Bother!", said Pooh, as he read Devast8or's latest post to
alt.games.warcraft.

>"Brian" <brian...@pobox.com> wrote in message

Just that if you continuously run the same instance with the same raid
leader, you're going to see bad (non-random) distributions. In our case,
most of our early Molten Core runs were done with a Warrior (our MT) as the
raid leader, and we had a veritable shitload of Felheart drops. So it
doesn't care what class anyone is. It's just an imbalanced random number
generator.

Brian
--
ICQ#: 68214833 | AIM: LineNoise54
.

Don't like my driving? Then quit watching me.

Brian

unread,
Dec 6, 2005, 4:24:41 AM12/6/05
to
"Bother!", said Pooh, as he read ASKF's latest post to alt.games.warcraft.

>I recently did more than 10 runs in a row of RFC (was hgelping a mage
>making a movie) and the boss drops changed a lot. Since we had the same
>group leader every time, I tend to believe Blizz is right.
>
>I'm also doing boss-runs at SM all the time, together with a warrior,
>and the boss loot is having the same drop-stats as when I was
>practically living there between lvl 37-42. It doesn't matter at all,
>wether my shaman or the warrior is the leader. It doesn't affect the
>loot chance either, if we bring a low lvl guildie along (to make sure he
>gets the blue drops he wants/needs).
>
>There might be differences between instances though, but I don't see why
>the dev's should have used different algorithms.

There's a significant difference in the 40-man instances, because a lot of
the loot weirdness that has been observed has been tied into some of the
same bad coding that gave us lockout bugs and split instances. There's
never been such a thing as a split instance in UBRS, because the mechanism
for being saved to an instance, and the instance lockout timer simply isn't
present in those cases.

Brian
--
ICQ#: 68214833 | AIM: LineNoise54
.

Peace, Love, and Cable Modems for everyone...

Jack D

unread,
Dec 6, 2005, 4:30:32 AM12/6/05
to
On 2005-12-06 09:45:33 +0100, Mel <mel...@pasdespam.chezmoi> said:

> And class drops in MC for us are at about 2 per person per class except
> priests at less than 1 per person. And those damn shammy bracelets keep
> dropping - stop Blizzard we don't want any more!!!!

Same with us, Felheart drops like crazy. A quick check shows me we had
2 times as much felheart drops as any other tier 1 set. I think Bliz
has put in a system to reward Guild Masters :)
--
http://www.dagon-roots.com/
Nerghal - Undead Warlock lvl 60 - Bloodscalp EU
Chasey - Undead Priest lvl 23 - Bloodscalp EU
Tomganks - Troll Rogue lvl 10 - Bloodscalp EU
Chasey - Human Mage lvl 12 - Twisting Nether EU

Dave

unread,
Dec 6, 2005, 4:32:51 AM12/6/05
to
Brian wrote:

> Just that if you continuously run the same instance with the same raid
> leader, you're going to see bad (non-random) distributions. In our case,
> most of our early Molten Core runs were done with a Warrior (our MT) as the
> raid leader, and we had a veritable shitload of Felheart drops. So it
> doesn't care what class anyone is. It's just an imbalanced random number
> generator.


We also had a shitload of Felheart, like every other Guild who is
raiding on our server. We and many other Guild tried repeateldy to
switch the Raid Leader, to no effort, Felheart just keeped dropping in
the beginning.

Soon Felheart drops got replaced by a shitload of Cenarion, again like
every other Guild on the Server.

You can see a pattern but i do not believe that changing RL does help
anything at all.

And by the way, in the end it event out for every class so no worries :)

Rene

unread,
Dec 6, 2005, 4:45:03 AM12/6/05
to
Jack D <jack_2...@Mhotmail.com> wrote:
> On 2005-12-06 09:45:33 +0100, Mel <mel...@pasdespam.chezmoi> said:
>
> > And class drops in MC for us are at about 2 per person per class except
> > priests at less than 1 per person. And those damn shammy bracelets keep
> > dropping - stop Blizzard we don't want any more!!!!
>
> Same with us, Felheart drops like crazy. A quick check shows me we had
> 2 times as much felheart drops as any other tier 1 set. I think Bliz
> has put in a system to reward Guild Masters :)

Nah, not tied to Guildmasters. At least not in our case. I'm still unsure
wheter it is the accountID, charID or playerClass as Raidleader who
influences the loot table. We now had two different Priests with a Warrior
one week interspersed open the instance. When it was a Priest's instance,
we had quite some Paladin drops (Paladins have had the fewest drops in our
group up until now so we try to increase it currently) This friday it will
either be me as third Priest or a Druid since we never had a Druid open the
instance. So in one or two week I should have a guess wheter the class is
important or the char.

Oh and btw, if you're reading this and are new to MC, this is how you
randomize:

- If no one is in the instance, simply select a random person and transfer
the raid leadership to that person. Then let people walk into the instance,
thus spawning it. As soon as some are in the instance, the random person of
the raid transfers the (L) back to the usual leader. It doesn't matter once
the instance has been spawned. Players must not leave the instance now
though until you get the "You are now saved to this instance" upon killing
the first boss.

- If there are already people in the instance and you want to reseed, let
two people leave the instance (if already in) and leave the raidgroup. Let
them form a party, then upgrade to raid. One of them has the (L) now.
They'll (re-)enter the instance, thus spawning a new one. That is to say,
now there should be a few players in the "old" instance and at least two in
the newly seeded instance. Now let all players leave the old raid and join
up in the new one. Let all walk out and immediately back in the instance
and everyone should be together again in the new instance. Transfer (L)
back to the usual raid leader. Of course this won't work on the second day
if you already have killed a boss and are bound to the instance.

Actually this above is just a variant on how you farm the blue instances.
For example for DM where you run the two bosses, get out, reform the party
and thus reset the instance and get back in. I suppose just as with normal
instances, that the reseeding only works up to five times in an hour,
though you generally only need this once per week if someone already walked
into the instance. Tell your people to wait outside to ensure proper
seeding, it's easier if you don't have to do this.

Kai Scholz-Starke

unread,
Dec 6, 2005, 5:03:16 AM12/6/05
to
On Tue, 06 Dec 2005 04:31:09 GMT, "Ham Pastrami" <nom...@dot.com>
wrote:


Hmm, having read the other replies to that thread, i really wonder
what a reliable n - number of drops - would be if you really want to
say that the distribution of drops among the classes is imbalanced.
Any fancy mathematician around?

Kai

Jack D

unread,
Dec 6, 2005, 5:08:25 AM12/6/05
to
On 2005-12-06 10:45:03 +0100, Rene <inv...@email.addr> said:

> Nah, not tied to Guildmasters. At least not in our case.

Was mainly kidding, being the GM/Warlock myself. Though I do appreciate
having 6 Felheart drops when we ran MC on my birthday :)

> Oh and btw, if you're reading this and are new to MC, this is how you
> randomize:
>
> - If no one is in the instance, simply select a random person and transfer
> the raid leadership to that person. Then let people walk into the instance,
> thus spawning it. As soon as some are in the instance, the random person of
> the raid transfers the (L) back to the usual leader. It doesn't matter once
> the instance has been spawned. Players must not leave the instance now
> though until you get the "You are now saved to this instance" upon killing
> the first boss.
>
> - If there are already people in the instance and you want to reseed, let
> two people leave the instance (if already in) and leave the raidgroup. Let
> them form a party, then upgrade to raid. One of them has the (L) now.
> They'll (re-)enter the instance, thus spawning a new one. That is to say,
> now there should be a few players in the "old" instance and at least two in
> the newly seeded instance. Now let all players leave the old raid and join
> up in the new one. Let all walk out and immediately back in the instance
> and everyone should be together again in the new instance. Transfer (L)
> back to the usual raid leader. Of course this won't work on the second day
> if you already have killed a boss and are bound to the instance.

Once I get my full Felheart set, I will inform my guild of this discussion :)

Chris

unread,
Dec 6, 2005, 6:53:29 AM12/6/05
to
Warlock & Druid are the least popular class however. Although their
stuff seems to drop more often isn't it just because their are less of
them?

Rene

unread,
Dec 6, 2005, 6:57:10 AM12/6/05
to

I'm no mathematician but had enough statistics done to have a basic
understanding in the field. I'd say you could start making really profound
claims after around 100 or so samples. Since every boss can approximately
be killed once a week (slightly more) and the game is only out a year now
in the US and 3 months later in the EU. So if you want the mathematical (or
rather, statistical) proof of these claims, you'd have to wait at least
another year.

However there are simply effects that don't require such a large sample
space to show how squewed the loots are. The "proof" is however purely
anecdotical. I can tell you what happend to us and your interpretation of
that event can be hugely different from mine. It's just not possible to
proof it really at this point. But I don't feel like I'd have to. It's what
happend to us and it's what happens to lots others (but not all). Even if
this were statistically "normal", it sucks. Hard. And we have the control
to make it less sucky, so we do that now, once we found out that we indeed
have a little control over it.

So enough preamble, that's what happend to us: We got 5 weeks in a row the
earthshaker from magmadar. If you go to thottbot and look it up, you'll
find the dropchance overall observed by thottbot. Let's give the doubt to
the values and let's say the dropchance is 20%. Getting this item 5 times
in a row therefor has a chance equal to 0.2^5 which is 0.032%. Now if this
were the only oddity I'd accept it as a statistical outlyer. They happen.
But it's absolutely NOT the a seldom event. During the same period, 5 out
of 6 random drops where warlock drops. The naked chance is 1/8 for each of
the 8 classes.

So interprete this as you want, it is no proof of anything save that we're
a bit unlucky (and unhappy). But later I tumbled upon a thread in a forum
describing the same problem (with other items) and someone suggested the
"seed randomizing". We tried it too and it helped us. I tried to search for
the reason of why this is so, and all I've written about this topic is as
much as I can come up with. I have no hard statistical data and I cannot
back anything up with anything else than our stories, those of people I
talk with and of other people's reports. It's all anecdotical. In fact, one
of our 6 raidleaders still does not believe it either and thinks it's
purely random. He's not convinced - yet :-)

But all in all, I don't really care if YOU believe it or not. I personally
believe it. It's hard to impossible to proof and Blizzard is not willing to
discuss this, I may say so, rather widespread problem. In patch 1.9 we're
getting a new system for raid-lockout to prevent the cascading. It may well
be, that they change it silently. However the effects I see are there since
1.6 where we started. Maybe there are changes all the time, maybe there
aren't.

To summarize, this is how I believe it works at the moment. I haven't got
ANY proof to support this, so take it for what it is, a theory:

- There are different loot tables, each loot table has a probability for
each item that is able to drop in the instance. The probability for any
specific such item may be anything from 0% (shaman stuff for alliance..) to
??%. These probabilities are fixed per loot table, they don't change.

- The person who is raidleader at time of instance spawning seeds the
random number generator that picks the loot table for the instance. There
very likely is another component in the chosing of the loot table, because
even if always the same person opens, it seems not to be always the same
loot table. However this other component is certainly not a dominating
factor. Might even be the day number which would cause the factor to be
constant if the instance is always spawned at the same day like we and lots
of people do it.

- When a boss is killed the item drop is a random roll but based upon the
probability of the current loot table, items are chosen. Since the loot
tables seem to have very varying and uneven probabilities, this makes the
occurrence of the much observed streaks quite common.

To summarize: There is ALWAYS a random factor. It's just that the
probability distribution can be VERY skewed and that for a very very long
time, causing streaks of the same items dropping over and over for many
weeks in succession. Drops ARE random, the distribution per loottable
however is not and you CAN influence which loottable is being chosen.

That is how I see it how it works. I cannot give proof that this is really
so. I might be far far away from the truth. But this is my best shot at
what I can observe from my POV. It does coincide with other people's
observations but by no means with all other people's observations.

Rene

unread,
Dec 6, 2005, 7:22:40 AM12/6/05
to

Warlock Druid AND Priest are the most underplayed classes - worldwide and
also when you only look at the EU. Still we have too much warlock stuff, ok
druid stuff and slightly too few priest items. Yet WE have far too few
paladin drops, other raid groups have zero to no priest drops but too many
Paladin drops. Other groups have excessivly many Hunter drops but lack
other classes. Viewed over the server it might even out, I don't know. But
many individual groups have one or two classes that have it all, while one
or two classes are swimming in points to spend.

We swap BoEs with other groups but you can't do that with the BoPs.

Of course it gets worse the longer you raid. You start to disenchant stuff
while others only have at most 2 set items. It sucks.

Olaf

unread,
Dec 6, 2005, 8:24:16 AM12/6/05
to
I dont know much about programming, or enough about math to be able to prove
anything. But in my experience as a guild leader, their random loot is not
very random. Granted, our samples are small relatively speaking. But when
you lead every raid and see the same shit DEd over and over again, it wears
on you.

Like pretty much everyone else, we are swimming in Felheart. I think this
is partially just a Blizzard being stupid problem, and not adjusting drop %s
for the least played classes. We get an uncanny amount of Giantstalkers
BOEs. Basically ever other BOE drop we get is GS piece, usually a belt.

We have also had a problem with Onyxia loot. In over 30 kills we have seen
over 15 Ancient Cornerstone Grimoires. We also have over 15 Priest hats.
At the same time we have seen just one weapon ever, a Deathbringer.

They need to implement some kind of a token loot system, ala Velious in EQ.

olaf


Adam Russell

unread,
Dec 6, 2005, 12:20:19 PM12/6/05
to

"Ham Pastrami" <nom...@dot.com> wrote in message
news:hu8lf.31950$tV6....@newssvr27.news.prodigy.net...

It's bullshit. People always say stuff like this without any specific
theory to back it up. Never does anyone get down to specifically saying
something like "If its a priest leader then more felheart will drop".
Because if they said something specific then it could be immediately
disproven in just a few runs. Instead they always say something
non-specific like "If we change the raid leader we 'might' get different
drops". Which of course is true except that you also get different drops
all the time with no change in raidleader. Bottom line though is the best
way to prove a theory is to test it, but no one yet has come up with a
theory thats testable.


Caeryn Dryad of Whisperwind

unread,
Dec 6, 2005, 1:31:20 PM12/6/05
to
QFT

I have spoken personally with the person in charge of the
randomization. They have acknowledged that there is such a perception
in the player base and have gone above and beyond to test their results
and found no such connection as Brian pointed out.

However, I am also one of who believes in the raid seed concept having
seen my raid get the same crap over and over until we changed raid
starters.

Hornet

unread,
Dec 6, 2005, 2:48:56 PM12/6/05
to
If blizz changed drop % based on how many ppl play that class or were
in your raid they would wipe out a really nice aspect of certain
classes. One of the advantages to being a druid or a warlock is that
you get geared out very fast. This makes up for the (former) suck of
druids and the perceived suck of warlocks keeping players away by
having some start thinking "well I will have an easy time finding an MC
run and getting purples." The counter to this, is people are less
inclined to roll warriors, rogues and recently, hunters because despite
being very powerful, there are dozens of unguilded 60 rogues and the
newer guilds are jammed with war/hunters.

I rolled an alt warrior knowing full well that it'll take a long time
to get geared, but when I do a warrior in many epics>>>>>>lock in
felheart. Felheart is scrub gear, you have nice stats but good luck
with the damage(+96 or something for the set) and in BWL I'm only using
2 pieces(1 nemesis) even though the whole set is at my disposal.

In short, gear/guilds/grouping is something many people do consider
when rolling new toons, and its fine the way it is.

Christian Stauffer

unread,
Dec 7, 2005, 3:35:16 AM12/7/05
to
"Adam Russell" <adamr...@sbcglobal.net.invalid> wrote:

> It's bullshit. People always say stuff like this without any specific
> theory to back it up.

Hrm.... So, what you say is:
A connection between the raid leader and the drop rates is bullshit,
because there is no bullet proof evidence.
Do I assume right you have proofs that there is no connection? If so,
please feel free to explain them here. If not, I'd be more careful
before labelling something as bullshit, it put's you in a very bad
position.

Chris

--
[WoW] Wildcard - Treehugging Tauren (60) on EN Sunstrider [PvP]
Lonewalker - Striding Tauren (22) on EN Sunstrider [PvP]
Jazrah - Brutal Troll (16) on EN Sunstrider [PvP]
Jivarr - Charming Troll (12) on EN Sunstrider [PvP]

Babe Bridou

unread,
Dec 7, 2005, 5:16:29 AM12/7/05
to

Christian Stauffer a Ă©crit :

> "Adam Russell" <adamr...@sbcglobal.net.invalid> wrote:
>
> > It's bullshit. People always say stuff like this without any specific
> > theory to back it up.
>
> Hrm.... So, what you say is:
> A connection between the raid leader and the drop rates is bullshit,
> because there is no bullet proof evidence.
> Do I assume right you have proofs that there is no connection? If so,
> please feel free to explain them here. If not, I'd be more careful
> before labelling something as bullshit, it put's you in a very bad
> position.
>
> Chris

He has a point though. Nobody has come up with any theory that's
actually testable.

Really, do forget about this connection, as it has no relevance.

Here's a theory that is actually testable:

[following is complete garbage]
Technically, the rolls are at very least determinist. Change *one*
thing in your raid strategy, such as switching the order in which you
kill bosses and mobs, having the mob reset through a wipe or an evade
trick, get one more or one less core hound pop in the middle of your
run and there are chances that you'll see much, much different drops.
This is testable in instances, but you won't get any info over such
small sample experiment sets. It has been tested that raid leader
doesn't influence the drops - but I am convinced that team strategy and
timing on the other hand, do. Depending on how many raidmates you have,
which add do you kill or banish and in what order, how many hits does
it take for you to kill a boss, etc, etc, you'll have many differences
in drops. There are patterns in random rolls within the same seed, it's
true, after like a million rolls you'll start seeing things such as
546789465467892134546789 with "546789" being a pattern that happens
often... so if your strategy is similar everytime you reset the
instance, you'll have good chances to fall inside that pattern
lifespan, so much more chance to roll a 5, a 4, 6, 7, 8 or 9 than 1, 2
or 3. But the fact is, these patterns appear and disappear within a
"random" sequence after a while. Get half that amount of "rolls" and
the pattern isn't here yet, get double that amount of "rolls" and the
pattern is no more visible. That doesn't mean you take more time or
less time to do the instance, simply that you force more rolls or less
rolls on the server side before the "drop roll" arrives. Switching from
two 1.3 speed daggers to one 2.6 speed one hander with shield is a way
to divide the amount of attack rolls you personally do by four, but you
also add a block roll if the enemy strikes you. Get within AOE range
and you'll get a resist & damage roll. Remove one % of criticals and
you'll do 1% less damage, leading to more rolls. Lastly, if one roll in
the pattern leads to a certain critical hit or spell, depending on the
exact time and sequence everyone acts in the instance, you'll either
have a critical from the boss, a critical hit on the boss or a critical
heal on the tank, with all of these leading to a lot of variation in
the aftermaths of the roll: healers will start healing more (more
rolls)... etc etc etc bla bla
[end of complete garbage]

In the end it's random. If you're unhappy with your drops, just try to
farm it in a much different way. At least it could be a more fun way to
stash nexus materials for 1.9.

BombayMix

unread,
Dec 7, 2005, 7:01:47 AM12/7/05
to
>In the end it's random. If you're unhappy with your drops, just try to
>farm it in a much different way. At least it could be a more fun way to
>stash nexus materials for 1.9.

Reading all the above and from my knowledge of mmog there's nothing
here that can't be explained by probability and the limitations of
computer based random algorithms.

It's human nature to see patterns in things. People often say, "we
got x dropped twice in a row that can't be random" but there's a
probability you will get it twice and a probability you won't. Just
because the probability very slim doesn't mean it won't happen. The
probability you will win the lottery very remote yet people do win it.

It comes down to how random Blizzard's random algorithm is. Making a
truly random algorithm is very hard to do and I'll be amazed if
Blizzard went to that length. All computer random algorithms are prone
to streaks and can be analysed to determine patterns. But you are going
to need a degree in maths to work them out.

The "randomness" of the algorithm Blizzard uses is undermined due
to that fact items are not designed to drop evenly. You only have to
play the game for a few minutes to see you get more of one item then
another. This can cause streaks in drop rates.

And then there's also the possibility of bugs in their code causing
drop rate problems. In my experience unless its a bug or designed to
work that way, most theories like "if you do x you get item y"
aren't true.

So basically everyone might be right! ^_^ But it's impossible to
tell exactly who's more right since we don't know the exact nature
of Blizzard's algorithm.

Joe Bryant

unread,
Dec 7, 2005, 8:02:49 AM12/7/05
to
> It comes down to how random Blizzard's random algorithm is. Making a
> truly random algorithm is very hard to do and I'll be amazed if
> Blizzard went to that length.

A common trick is to include a time-related variable in the
calculation. For example, you could seed your random number generator
with the number of milliseconds since midnight. It's highly unlikely
that you'll consistently start your raid at the same number of
milliseconds past midnight.

--
Joe

Adam Russell

unread,
Dec 7, 2005, 1:42:14 PM12/7/05
to

"Christian Stauffer" <wildc...@bluewin.ch> wrote in message
news:43969ec3$0$1145$5402...@news.sunrise.ch...

> "Adam Russell" <adamr...@sbcglobal.net.invalid> wrote:
>
>> It's bullshit. People always say stuff like this without any specific
>> theory to back it up.
>
> Hrm.... So, what you say is:
> A connection between the raid leader and the drop rates is bullshit,
> because there is no bullet proof evidence.

No, that is not what I said.

> Do I assume right you have proofs that there is no connection? If so,
> please feel free to explain them here. If not, I'd be more careful
> before labelling something as bullshit, it put's you in a very bad
> position.

In a game where stats are thoroughly tested to the nth degree, no one yet
has proposed a theory to the effect of "If you have a priest raidleader then
mostly lock stuff drops but if you have a warrior raidleader then mostly
priest stuff drops". The lack of a testable theory in this case *is*
evidence of a lack of relation. It is my very humble opinion that the
theory that "there is some kind of AS YET UN-SPECIFIED link between the
raidleader and the drop types" is bullshit.


adamc...@fastmail.us

unread,
Dec 7, 2005, 5:18:57 PM12/7/05
to

BombayMix wrote:

> It's human nature to see patterns in things.

The rest of your post was well-written, informative, and insightful,
but this one sentence really hits the nail on the head.

If you run MC a dozen times, you will see a streak of _something_. It
may be something awesome like the Talisman of Aggro. It may be
something terrible like Shardstrike. More likely it'll be something in
between. But it's a statistical near-certainty that you will see a
streak. That's just the way random numbers are.

Suppose you have five bosses in an instance, and each boss gives one of
five possible drops per kill. So, for boss X, you have a 20% chance to
get drop Y. You run the instance once, get one piece of loot from each
boss. You run the instance a second time, and each boss has a 20%
chance to drop the same item he dropped on your first run. Out of the
five bosses, that means there's a pretty good chance at least one of
them will drop the same item he dropped on the first run. It's not
evidence that the random number generator is broken; it's evidence that
it's working just fine. (The chances of repeating one item get even
higher when a single boss drops two or more items at once.)

To take one specific example, the poster who says his/her guild got
five Earthshakers out of six Magmadar kills. That's an impressive
streak. Did you also give five Chokers of Enlightenment from Lucifron?
How about five Aurastone Hammers off of Garr? Five Drillborer Disks?
Five Cauterizing Bands? Five Brutality Blades? I'm assuming not,
because if you had gotten those you would have mentioned it. The point
is that if you list every one of the 100+ items that drops in Molten
Core, you had a streak on _one_ of them. That's not surprising.

Adam

Erick Pelden

unread,
Dec 7, 2005, 5:59:32 PM12/7/05
to
Rene <inv...@email.addr> wrote:
> So enough preamble, that's what happend to us: We got 5 weeks in a row the
> earthshaker from magmadar. If you go to thottbot and look it up, you'll
> find the dropchance overall observed by thottbot. Let's give the doubt to

Thottbots drop rates are a bad data basis. Blizzard is known to change
drop tables between patches and since Thott shows all collected data
since the beginning, the numbers can't be exact.

But 5 weeks in a row the same loot IS rare.

--
Erick

Rene

unread,
Dec 8, 2005, 2:19:28 AM12/8/05
to
adamc...@fastmail.us wrote:
> BombayMix wrote:
>
> > It's human nature to see patterns in things.
>
> The rest of your post was well-written, informative, and insightful,
> but this one sentence really hits the nail on the head.
>
> If you run MC a dozen times, you will see a streak of _something_. It
> may be something awesome like the Talisman of Aggro. It may be
> something terrible like Shardstrike. More likely it'll be something in
> between. But it's a statistical near-certainty that you will see a
> streak. That's just the way random numbers are.

Indeed and I would not have said a thing, except for the simple fact that
some groups experience THE VERY SAME STREAK WEEK AFTER WEEK AFTER WEEK.
sorry for shouting but I believe you (not just you, Adam) just don't
realize the extent of the problem. If you get Bindings of the Windseeker 5
weeks in a row (droprate according to thottbot ~ 3%, legendary quality) AND
other streaks concurrently, would you still cite that seeing patterns is
human and "normal" ?

Btw. when everyone says normal I expect people mean to say random uniform
distribution (or at least distribution according to the probability of
thottbot's recording - both is simply not happening).

Heck for one second, just take a look at the drops from Razorgore:
http://www.thottbot.com/?n=685299

Is there ANY sense in those drop rates? Why shouldn't all class gear be at
1/8 drop chance ? (except paladin and shaman of course, because they are
faction dependend). As a normal, non statistical-casual WoW player you
would expect the drop rate to be near 1/8 for each of those, no? What if
they are but the loot-drop system is so broken this is the result thott
sees?

Look, drops ARE random. The distribution is just very non-uniform. And you
can influence the distribution. That's what I'm saying and that's my
experience. I cannot go and test it, because my Raidgroup would kill me at
this point if we would go back to the same (L) opening the instance every
week that already caused all our warlocks to be fully equipped and caused 0
paladin and 0 rogue drops during that time (and 5 earthshakers). Please if
you're sooo inclined in your statistical data proof YOUR claim first.

I find it very odd to be answered with a "you can't proof it, therefor your
claim is invalid" and at the same time not willing to proof the other claim
"it works as intended" without any proof and without even the
acknowledgment that this, too, would need a proof of the same scope...

It's actually a prime example of hypocrysis. The "seeing patterns or
streaks is normal" is just as anecdotical as all the stuff I have said for
this topic.

And at the end of the day, playing advocatus diabolii: why is there a token
system in Zul'gurub and not a pure drops system and why is the same coming
for Ahn'Quiraj? Does Blizzard not trust its own drop system anymore?

> Suppose you have five bosses in an instance, and each boss gives one of
> five possible drops per kill. So, for boss X, you have a 20% chance to
> get drop Y. You run the instance once, get one piece of loot from each
> boss. You run the instance a second time, and each boss has a 20%
> chance to drop the same item he dropped on your first run. Out of the
> five bosses, that means there's a pretty good chance at least one of
> them will drop the same item he dropped on the first run. It's not
> evidence that the random number generator is broken; it's evidence that
> it's working just fine. (The chances of repeating one item get even
> higher when a single boss drops two or more items at once.)

And what are the chances if you get the f.... same item 5 weeks in a row
and it's "base drop probability" according to thottbot is below 20% ? THIS
happend to us, and "coincidentally" at the same time EVERY SINGLE RANDOM
DROP was felheart. Out of the 8 classes. We get around 2 to 3 random epic
drops per run. Still not enough? Well Magmadar dropped more than 6
flamewalker legplates for us in about 20 runs. Go and look up THAT drop
chance. Still not enough? I have more oddities to report. But, frankly, I'm
sick of it. We all were. We changed it now and drops are significantly
different now. I can't go back using the "old" person for instance seeding
now, my raidgroup would hang me to the nearest tree.

> To take one specific example, the poster who says his/her guild got
> five Earthshakers out of six Magmadar kills. That's an impressive
> streak. Did you also give five Chokers of Enlightenment from Lucifron?
> How about five Aurastone Hammers off of Garr? Five Drillborer Disks?

We got more than 4 Drillborer disks, so yes. I don't look up every single
drop from them so I don't know what really ought to be extremely seldom. We
can make perdition's blade drop from ragnaros if we want to, though.

> Five Cauterizing Bands? Five Brutality Blades? I'm assuming not,

Three brutality blades, though the last one dropped when we were already
rotating so it's no longer the same person causing it.

> because if you had gotten those you would have mentioned it. The point
> is that if you list every one of the 100+ items that drops in Molten
> Core, you had a streak on _one_ of them. That's not surprising.

That's exactly my point, as I already wrote. I'd discount it if it were
only ONE streak. It's many concurrently, for weeks (as long as we had the
same person opening the instance). It sucks.

Babe Bridou

unread,
Dec 8, 2005, 2:48:25 AM12/8/05
to

Rene a Ă©crit :

> Heck for one second, just take a look at the drops from Razorgore:
> http://www.thottbot.com/?n=685299
>
> Is there ANY sense in those drop rates?

Yep.

Thottbot requires an add-on, and if several people have this add-on
when the boss dies, the drop will be accounted for multiple times.

Look at the amount of kills: 181 loots. It's low. Not even
statistically relevent.
All classes are at a theoretical drop rate of 1/8 I'm pretty sure of
it.

These global stats have nothing to do with randomness or "streaks", but
I'm convinced streaks exist, it's just hard to believe that majordomo's
chest will loot eye of divinity 90% of the time for a guild only due to
pure luck.

Christian Stauffer

unread,
Dec 8, 2005, 3:24:25 AM12/8/05
to
"Adam Russell" <adamr...@sbcglobal.net.invalid> wrote:

> "Christian Stauffer" <wildc...@bluewin.ch> wrote:
>
>> Hrm.... So, what you say is:
>> A connection between the raid leader and the drop rates is bullshit,
>> because there is no bullet proof evidence.
>
> No, that is not what I said.

If that's not what you said, why do you repeat it then in this post?

>> Do I assume right you have proofs that there is no connection? If so,
>> please feel free to explain them here. If not, I'd be more careful
>> before labelling something as bullshit, it put's you in a very bad
>> position.
>
> In a game where stats are thoroughly tested to the nth degree, no one yet
> has proposed a theory to the effect of "If you have a priest raidleader then
> mostly lock stuff drops but if you have a warrior raidleader then mostly
> priest stuff drops".

Why do you think there had to be a relation between the raidleaders class
and the drops? Noone has been talking about a relation between the
raidleaders class and the drop tables, why do you?
As someone said, random number generators use seeds. Assumptions are this
seeds are based on a character. This could be his charname, his account
name, the date he joined the guild/the game, anything.

> The lack of a testable theory in this case *is* evidence of a lack of
> relation.

Now that is bullshit. A theory is automatically wrong, if there's no way
(yet!) to prove it? So, the existence of intelligent life outside this
planet is impossible?

adamc...@fastmail.us

unread,
Dec 8, 2005, 3:44:05 AM12/8/05
to

Rene wrote:
> adamc...@fastmail.us wrote:

> > If you run MC a dozen times, you will see a streak of _something_. It
> > may be something awesome like the Talisman of Aggro. It may be
> > something terrible like Shardstrike. More likely it'll be something in
> > between. But it's a statistical near-certainty that you will see a
> > streak. That's just the way random numbers are.
>
> Indeed and I would not have said a thing, except for the simple fact that
> some groups experience THE VERY SAME STREAK WEEK AFTER WEEK AFTER WEEK.
> sorry for shouting but I believe you (not just you, Adam) just don't
> realize the extent of the problem. If you get Bindings of the Windseeker 5
> weeks in a row (droprate according to thottbot ~ 3%, legendary quality) AND
> other streaks concurrently, would you still cite that seeing patterns is
> human and "normal" ?

Another factor here is that the people who don't see streaks don't
complain about it. Nobody ever starts a conversation with "Wow, we've
been doing MC for a few months and the loot distribution's been pretty
normal!" If a thousand guild are doing MC and 990 of them are seeing
no unusual streaks at all, people will point to the ten who are
complaining about imbalanced drops as "proof".

> Btw. when everyone says normal I expect people mean to say random uniform
> distribution (or at least distribution according to the probability of
> thottbot's recording - both is simply not happening).
>
> Heck for one second, just take a look at the drops from Razorgore:
> http://www.thottbot.com/?n=685299
>
> Is there ANY sense in those drop rates? Why shouldn't all class gear be at
> 1/8 drop chance ? (except paladin and shaman of course, because they are
> faction dependend). As a normal, non statistical-casual WoW player you
> would expect the drop rate to be near 1/8 for each of those, no? What if
> they are but the loot-drop system is so broken this is the result thott
> sees?

Because that sample size is nowhere near large enough. You're looking
a total of 300ish Tier 2 bracers there divided between eight classes
(you can lump Paladin and Shaman into one group for this purpose.) No
way is that a large enough sample.

> Look, drops ARE random. The distribution is just very non-uniform. And you
> can influence the distribution. That's what I'm saying and that's my
> experience. I cannot go and test it, because my Raidgroup would kill me at
> this point if we would go back to the same (L) opening the instance every
> week that already caused all our warlocks to be fully equipped and caused 0
> paladin and 0 rogue drops during that time (and 5 earthshakers). Please if
> you're sooo inclined in your statistical data proof YOUR claim first.

Why should I? I'm simply pointing out that the streaks everyone points
to are nothing surprising from a statistical perspective. You're the
one making the claim of unusual behavior; the burden of proof rests on
you.

But if you do want proof, the work has been done by Blizzard, who
examined all their drop data over several months and found no
significant deviation from the prescribed drop rates. This was
reported at Blizzcon.

> I find it very odd to be answered with a "you can't proof it, therefor your
> claim is invalid" and at the same time not willing to proof the other claim
> "it works as intended" without any proof and without even the
> acknowledgment that this, too, would need a proof of the same scope...

I'm not saying "you can't prove it, your claim is invalid". What I'm
saying is that there's no reason to assume some kind of non-randomness
at work when the streaks we see in loot can be explained simply as the
normal behavior of a randomly determined distribution.

> It's actually a prime example of hypocrysis. The "seeing patterns or
> streaks is normal" is just as anecdotical as all the stuff I have said for
> this topic.

No it's not anecdotal at all. It's well-known in statistics that
streaks happen.

> And at the end of the day, playing advocatus diabolii: why is there a token
> system in Zul'gurub and not a pure drops system and why is the same coming
> for Ahn'Quiraj? Does Blizzard not trust its own drop system anymore?

A far more reasonable explanation is that they realized it's a pain in
the ass to keep farming an instance over and over just for the 1/8
chance that your last member will finally get his Tier 1 chest.

[snip]

> That's exactly my point, as I already wrote. I'd discount it if it were
> only ONE streak. It's many concurrently, for weeks (as long as we had the
> same person opening the instance). It sucks.

There's another reason I don't buy this. For the random-number-seed
theory to hold water, you'd have to get _exactly_ the same drops from
week to week -- every single one would have to be identical. That's
the way random number generators work; seeding them with the same
number produces exactly the same result every single time. Seeding
them with a similar number does _not_ produce a similar result -- it
produces a completely different result. Unless you were getting
exactly the same loot from week to week -- and by your own admission
you were not -- then you weren't running with the same random number
seed.

You have my sympathies if you've gotten a string of loot you don't
like, but I don't believe for a second that it's anything other than
your own luck.

Adam

Rene

unread,
Dec 8, 2005, 3:58:26 AM12/8/05
to
"Babe Bridou" <babeb...@hotmail.com> wrote:
> Rene a =E9crit :

>
> > Heck for one second, just take a look at the drops from Razorgore:
> > http://www.thottbot.com/?n=3D685299

> >
> > Is there ANY sense in those drop rates?
>
> Yep.
>
> Thottbot requires an add-on, and if several people have this add-on
> when the boss dies, the drop will be accounted for multiple times.

Yes, and allakhazam has much more "expected" drop probabilities. Yet for
some items, the values are almost the same between the two sites. Of
course, the same above applies to allakhazam. It's a bit more towards what
I'd expect though.

> Look at the amount of kills: 181 loots. It's low. Not even
> statistically relevent.
> All classes are at a theoretical drop rate of 1/8 I'm pretty sure of
> it.

Yes it's low. There's not too many people with the recording addons doing
these bosses. Heck' it's all endgame stuff. But as for the 1/8, you have
the same hard time giving proof that this is so as I have in showing that
the raid loot system has irregularities.

By the way, did I tell what Razorgore dropped on our first kill last week?
Two times bracelets of wrath... if the chance is truely 1/8 then we already
had our first 1/64 = 1.5625% chance of this happening. Damn are we the
(un?)lucky bastards - _yet_ again.

> These global stats have nothing to do with randomness or "streaks", but
> I'm convinced streaks exist, it's just hard to believe that majordomo's
> chest will loot eye of divinity 90% of the time for a guild only due to
> pure luck.

There's a guild on our realm that had the eye drop 8 times in a row. Their
hunters are really pissed.

Again, single occurences of such streaks are fully ok. But not if that's
all you get. Rotating leaders can make this go away immediately. But the
new person could also get the same loottable, so this is by no means a
guarantee to get other drops if you are on a (un-)lucky streak.

And now I stop participating in this discussion. I don't care if anyone
else believes it. I do. We rotate. We get better (other) loot now. I cannot
give proof and I won't ever be able to, since I cannot go back to my raid
and let the same person cause same drops again to proof a point to some
complete outsiders. Hopefully this all will be unneeded for 1.9 anyway.

Babe Bridou

unread,
Dec 8, 2005, 4:27:57 AM12/8/05
to

Rene a Ă©crit :

> CU
>
> René

You know, I do believe in these not-so-random streaks. Or at least
biased.

In the end, I'm pretty sure the bullshit theory I came up with in my
previous post is the right one. Once you have a boss on farm status
you'll have more or less the same amount of rolls every time you go for
it, giving you more chance to fall inside the same "heavily patterned
randomness interval", which would give you something such as [eye,
leaf, eye, eye, eye, leaf, eye, eye, eye, leaf, eye, leaf, eye, eye,
eye, eye, eye, leaf, leaf, eye]*100, until the chaotic sensible
parameter which is for example the computer reason for the first "eye"
in the pattern reaches a certain value and forks into a leaf, breaking
it and leading to another pattern... In the end it's still configured
with 50% leaf, 50% eye, it's just that between random number roll
#1000000 and random number roll #2000000, your instance seed tells you
that you will roll under 50% 14 times out of 20.

Just change the way you raid, and get into another rolling interval
(slow weapons, big spells, big heals, big crits vs fast weapons,
instant spells, flash heals, no crits). Try killing gehennas before
lucifron, for example, if you want Maggie to drop something else. Or
switch raid leader, it could mean another seed. But you won't be
escaping another pattern I'm afraid.

Rene

unread,
Dec 8, 2005, 4:48:56 AM12/8/05
to
adamc...@fastmail.us wrote:
> Rene wrote:
> > adamc...@fastmail.us wrote:
>
> > > If you run MC a dozen times, you will see a streak of _something_.
> > > It may be something awesome like the Talisman of Aggro. It may be
> > > something terrible like Shardstrike. More likely it'll be something
> > > in between. But it's a statistical near-certainty that you will see
> > > a streak. That's just the way random numbers are.
> >
> > Indeed and I would not have said a thing, except for the simple fact
> > that some groups experience THE VERY SAME STREAK WEEK AFTER WEEK AFTER
> > WEEK. sorry for shouting but I believe you (not just you, Adam) just
> > don't realize the extent of the problem. If you get Bindings of the
> > Windseeker 5 weeks in a row (droprate according to thottbot ~ 3%,
> > legendary quality) AND other streaks concurrently, would you still cite
> > that seeing patterns is human and "normal" ?
>
> Another factor here is that the people who don't see streaks don't
> complain about it. Nobody ever starts a conversation with "Wow, we've
> been doing MC for a few months and the loot distribution's been pretty
> normal!" If a thousand guild are doing MC and 990 of them are seeing
> no unusual streaks at all, people will point to the ten who are
> complaining about imbalanced drops as "proof".

Had I never seen a thread about this in a forum and hadn't we tried it out
ourselves I'd never said a thing about it myself. I'd just accepted it. You
can influence it. But nobody says you have to.

> > Btw. when everyone says normal I expect people mean to say random
> > uniform distribution (or at least distribution according to the
> > probability of thottbot's recording - both is simply not happening).
> >
> > Heck for one second, just take a look at the drops from Razorgore:
> > http://www.thottbot.com/?n=685299
> >
> > Is there ANY sense in those drop rates? Why shouldn't all class gear be
> > at 1/8 drop chance ? (except paladin and shaman of course, because they
> > are faction dependend). As a normal, non statistical-casual WoW player
> > you would expect the drop rate to be near 1/8 for each of those, no?
> > What if they are but the loot-drop system is so broken this is the
> > result thott sees?
>
> Because that sample size is nowhere near large enough. You're looking
> a total of 300ish Tier 2 bracers there divided between eight classes
> (you can lump Paladin and Shaman into one group for this purpose.) No
> way is that a large enough sample.

Consider for one second that thottbot is actually a sample sequence drawn
from everything that drops. It could now either be a bad sample sequence or
it could actually just show that the distribution is absolutely not
conforming to random uniform.

> > Look, drops ARE random. The distribution is just very non-uniform. And
> > you can influence the distribution. That's what I'm saying and that's
> > my experience. I cannot go and test it, because my Raidgroup would kill
> > me at this point if we would go back to the same (L) opening the
> > instance every week that already caused all our warlocks to be fully
> > equipped and caused 0 paladin and 0 rogue drops during that time (and 5
> > earthshakers). Please if you're sooo inclined in your statistical data
> > proof YOUR claim first.
>
> Why should I? I'm simply pointing out that the streaks everyone points
> to are nothing surprising from a statistical perspective. You're the
> one making the claim of unusual behavior; the burden of proof rests on
> you.

Well if you see it that way then please only say you don't believe in my
theory and never say something along the lines of "it works fine", because
that is just a claim of the same magnitude as is mine.

> But if you do want proof, the work has been done by Blizzard, who
> examined all their drop data over several months and found no
> significant deviation from the prescribed drop rates. This was
> reported at Blizzcon.

Yes. Blizzard employees also said that a Priest needs Prayer of Fortitude
Rank 1 before he can obtain and use PoF Rank 2. They also said that for
Mind View to work, you need to be within LoS AND distance, mob - priest.
Look I'm sure what I see is not intended and there is a bug somewhere. They
are people, they make mistakes, the latest one being the 10minute rule that
shouldn't have been included. But the fact remains that there are some that
influence this effect to cause certain drops to happen. I'm pretty sure we
both (and blizzard) don't want it to be in player's hands to force certain
specific items to have a very high chance to drop.

> > I find it very odd to be answered with a "you can't proof it, therefor
> > your claim is invalid" and at the same time not willing to proof the
> > other claim "it works as intended" without any proof and without even
> > the acknowledgment that this, too, would need a proof of the same
> > scope...
>
> I'm not saying "you can't prove it, your claim is invalid". What I'm
> saying is that there's no reason to assume some kind of non-randomness
> at work when the streaks we see in loot can be explained simply as the
> normal behavior of a randomly determined distribution.

I'm claiming the mechanism of chosing the distribution (loottable) that in
effect "creates" the drops is broken. I've detailed how I believe it works
at the current time and why this allows some to have "balanced" loot and
others to have completely unbalanced loot at the same time.

> > It's actually a prime example of hypocrysis. The "seeing patterns or
> > streaks is normal" is just as anecdotical as all the stuff I have said
> > for this topic.
>
> No it's not anecdotal at all. It's well-known in statistics that
> streaks happen.

Yes, and if it's too many streak, especially when they appear all at once,
that means that the base hypothesis is wrong and the distribution is not
what was hypotized. If I had enough samples I'd give you the chi-squared
test in a second to show you what I mean. I don't have them and I won't get
them. But the values are sooooo off limits for us and some other groups
that I lean that much out of the window to say it out loud for all to see:
You don't even need statistics to see from afar that something's broken
_for some_ groups.

> > And at the end of the day, playing advocatus diabolii: why is there a
> > token system in Zul'gurub and not a pure drops system and why is the
> > same coming for Ahn'Quiraj? Does Blizzard not trust its own drop system
> > anymore?
>
> A far more reasonable explanation is that they realized it's a pain in
> the ass to keep farming an instance over and over just for the 1/8
> chance that your last member will finally get his Tier 1 chest.

Especially if the 1/8 turns to 1/100 because the distribution is not random
uniform _for some_ due to a wrong/bugged mechanism that chooses the
loottables (which all in all viewed together DO have a random uniform
distribution)

> > That's exactly my point, as I already wrote. I'd discount it if it were
> > only ONE streak. It's many concurrently, for weeks (as long as we had
> > the same person opening the instance). It sucks.
>
> There's another reason I don't buy this. For the random-number-seed
> theory to hold water, you'd have to get _exactly_ the same drops from
> week to week -- every single one would have to be identical. That's
> the way random number generators work; seeding them with the same
> number produces exactly the same result every single time. Seeding
> them with a similar number does _not_ produce a similar result -- it
> produces a completely different result. Unless you were getting
> exactly the same loot from week to week -- and by your own admission
> you were not -- then you weren't running with the same random number
> seed.

See the message where I described how I see it works. This theory of mine
does in fact explain it why some are affected and others aren't and also
why it is NOT 100% guaranteed to get drop X.

> You have my sympathies if you've gotten a string of loot you don't
> like, but I don't believe for a second that it's anything other than
> your own luck.

Not just mine.. but oh well. This really is my last answer to this topic.
It eats too much time for too few gains. I need to get some strictly
deterministic work get done now. Thanks for listening and I wish you all
better luck than what happend to us. If not, well I explained how you can
safe the day once the problem is identified.

Jack D

unread,
Dec 8, 2005, 5:05:35 AM12/8/05
to
On 2005-12-08 09:24:25 +0100, "Christian Stauffer"
<wildc...@bluewin.ch> said:

> Why do you think there had to be a relation between the raidleaders class
> and the drops? Noone has been talking about a relation between the
> raidleaders class and the drop tables, why do you?
> As someone said, random number generators use seeds. Assumptions are this
> seeds are based on a character. This could be his charname, his account
> name, the date he joined the guild/the game, anything.

I don't believe for a second that Blizzard will use a static value as
seed. As someone else mentioned already, the most common seed value is
based on the time of the seed taking place. With the top winner here
being the number of seconds elapsed since january 1st 1970 00:10 as
this is the stored time on which the most popular unix time library is
based (and the return value of the time() call). There used to be a
post on the forum somewhere where a Bliz dev tells the pseudorandom
generator they are using and that it can be swapped out for another
algorithm should it prove to be not random enough. I'm also quite sure
Blizzard keeps statistics of about everything that happens in the game,
and if their loot distribution data (which you can expect to be a
couple of 1000 times the size of the datapool allakhazam uses) would
seem off, they'd notice.

I also wouldn't be surprised if they didn't seed a new generator for
each instance. A new generator means at least one number has to be
stored and allocating objects takes resources. So it would make more
sense if they would use a fixed number of generators, making the
generators more widely used and polled a lot, keeping it rotated enough
to increase the randomness factor. When you use recap to track your
crit% or DPS, you will see that after a long enough time, the values
will stabilize around the expected value. Which in itself is proof
enough that their generator has a good distribution.

One question that remains though is how the loot is selected based on
the random value they acquire. Would they roll for each item? Roll once
for each mob? Or itemize an entire instance on a couple of random
numbers?

Jack D

unread,
Dec 8, 2005, 5:20:06 AM12/8/05
to
On 2005-12-08 10:48:56 +0100, Rene <inv...@email.addr> said:

> Consider for one second that thottbot is actually a sample sequence drawn
> from everything that drops. It could now either be a bad sample sequence or
> it could actually just show that the distribution is absolutely not
> conforming to random uniform.

As you spend too much time in this thread, I'll just give my opinion on
one topic ;)

For starters as long as your sample sequence is not big enough, you
can't know if it is a bad sequence or the system is broken, basicly
your data is useless. Add to that the fact that if more than one person
on the same raid submits the drops to thottbot (or allakhazam), the
results are already skewed, this could be witnessed when Nefarian was
killed for the first time and 3 different people in the raid submitted
the loot to Allakhazam, giving a 100% drop chance for 3 specific items
and ignoring all the rest.

BombayMix

unread,
Dec 8, 2005, 5:20:43 AM12/8/05
to
>> Why do you think there had to be a relation between the raidleaders class
>> and the drops? Noone has been talking about a relation between the
>> raidleaders class and the drop tables, why do you?
>> As someone said, random number generators use seeds. Assumptions are this
>> seeds are based on a character. This could be his charname, his account
>> name, the date he joined the guild/the game, anything.

>I don't believe for a second that Blizzard will use a static value as
>seed. As someone else mentioned already, the most common seed value is
>based on the time of the seed taking place. With the top winner here
>being the number of seconds elapsed since january 1st 1970 00:10 as
>this is the stored time on which the most popular unix time library is
>based (and the return value of the time() call). There used to be a
>post on the forum somewhere where a Bliz dev tells the pseudorandom
>generator they are using and that it can be swapped out for another
>algorithm should it prove to be not random enough. I'm also quite sure
>Blizzard keeps statistics of about everything that happens in the game,
>and if their loot distribution data (which you can expect to be a
>couple of 1000 times the size of the datapool allakhazam uses) would
>seem off, they'd notice.

Actually, from a statically point of view that's not a good way to
generate a seed. Most instances will be started in the same time frame
so the majority of the key will be the same, ie only the last few
digits will change. It's easy to code but a very bad way to generate
a seed, so I doubt Blizzard would of done it that way.

Java generates a seed for its random algorithm in a far more complex
and random way, ie "The (random) class uses a 48-bit seed, which is
modified using a linear congruential formula."

Rene

unread,
Dec 8, 2005, 5:55:15 AM12/8/05
to
Jack D <jack_2...@Mhotmail.com> wrote:
> On 2005-12-08 10:48:56 +0100, Rene <inv...@email.addr> said:
>
> > Consider for one second that thottbot is actually a sample sequence
> > drawn from everything that drops. It could now either be a bad sample
> > sequence or it could actually just show that the distribution is
> > absolutely not conforming to random uniform.
>
> As you spend too much time in this thread, I'll just give my opinion on
> one topic ;)

Ok I'll bite, it's meal break anyhow... (another one last post, sue me :) )

> For starters as long as your sample sequence is not big enough, you
> can't know if it is a bad sequence or the system is broken, basicly
> your data is useless. Add to that the fact that if more than one person
> on the same raid submits the drops to thottbot (or allakhazam), the
> results are already skewed, this could be witnessed when Nefarian was
> killed for the first time and 3 different people in the raid submitted
> the loot to Allakhazam, giving a 100% drop chance for 3 specific items
> and ignoring all the rest.

And you're right. But as more people contribute, so does the sample space
grow. Molten core's being farmed since around april so it's got more people
every week contributing to the database since around then. You'd expect it
would have converged and stabilized by now but it hasn't (or it has
and...). Oh and from an information-theoretical point of view: No
information is ever completely useless. You can at least derive that there
is no convergence yet. Also I'd need to know if thottbot-client reacts on
the act of fetching loot by onesself (which wouldn't record those with
lootmaster on, but would prevent duplicate reports) or any other means.
Since thott-client is obfuscated (I had a look at it some months ago) I'll
leave that as an exercise to the reader :-)

Another little tidbit: It's not as if Blizzard never goofed up randomness.
Remember (or read up on) Diablo 2 bonewall farming. In certain spots (x/y-
coordinates) almost always the same items "dropped". Almost always. Just
like here.

Now I'm off, eating.. :)

Jack D

unread,
Dec 8, 2005, 6:06:00 AM12/8/05
to
On 2005-12-08 11:20:43 +0100, "BombayMix" <bomb...@altavista.co.uk> said:

> Actually, from a statically point of view that's not a good way to
> generate a seed. Most instances will be started in the same time frame
> so the majority of the key will be the same, ie only the last few
> digits will change. It's easy to code but a very bad way to generate
> a seed, so I doubt Blizzard would of done it that way.
>
> Java generates a seed for its random algorithm in a far more complex
> and random way, ie "The (random) class uses a 48-bit seed, which is
> modified using a linear congruential formula."

The offset of the seed has very little to do with the following random
number that gets generated. Random algorithms often include large
powers, making small offset result in big differences. Besides, I
really doubt each new instance will receive its own seed, on the
contrary. Random algorithms are made to be uniform distributed,
reseeding resets it. Only when the generator is initialised it needs a
seed to prevent deterministic behavior.

adamc...@fastmail.us

unread,
Dec 8, 2005, 6:19:43 AM12/8/05
to

Rene wrote:
> adamc...@fastmail.us wrote:

> > Because that sample size is nowhere near large enough. You're looking
> > a total of 300ish Tier 2 bracers there divided between eight classes
> > (you can lump Paladin and Shaman into one group for this purpose.) No
> > way is that a large enough sample.
>
> Consider for one second that thottbot is actually a sample sequence drawn
> from everything that drops. It could now either be a bad sample sequence or
> it could actually just show that the distribution is absolutely not
> conforming to random uniform.

Or it could be too small a sample size to draw conclusions from. In
fact, it IS too small a sample size to draw conclusions from. That's a
fact.

> > > Look, drops ARE random. The distribution is just very non-uniform. And
> > > you can influence the distribution. That's what I'm saying and that's
> > > my experience. I cannot go and test it, because my Raidgroup would kill
> > > me at this point if we would go back to the same (L) opening the
> > > instance every week that already caused all our warlocks to be fully
> > > equipped and caused 0 paladin and 0 rogue drops during that time (and 5
> > > earthshakers). Please if you're sooo inclined in your statistical data
> > > proof YOUR claim first.
> >
> > Why should I? I'm simply pointing out that the streaks everyone points
> > to are nothing surprising from a statistical perspective. You're the
> > one making the claim of unusual behavior; the burden of proof rests on
> > you.
>
> Well if you see it that way then please only say you don't believe in my
> theory and never say something along the lines of "it works fine", because
> that is just a claim of the same magnitude as is mine.

Okay, fine: I can't prove that it's working fin. However, there's no
real evidence whatsoever that it isn't working fine, aside from
anecdotal.

> > But if you do want proof, the work has been done by Blizzard, who
> > examined all their drop data over several months and found no
> > significant deviation from the prescribed drop rates. This was
> > reported at Blizzcon.
>
> Yes. Blizzard employees also said that a Priest needs Prayer of Fortitude
> Rank 1 before he can obtain and use PoF Rank 2. They also said that for
> Mind View to work, you need to be within LoS AND distance, mob - priest.
> Look I'm sure what I see is not intended and there is a bug somewhere. They
> are people, they make mistakes, the latest one being the 10minute rule that
> shouldn't have been included. But the fact remains that there are some that
> influence this effect to cause certain drops to happen. I'm pretty sure we
> both (and blizzard) don't want it to be in player's hands to force certain
> specific items to have a very high chance to drop.

It's not a fact that the drop rate can be influenced. It's a
supposition you're making based on your own subjective interpretation
of your own experience.

Yes, Blizzard has made mistakes in the past. What you seem to be
saying here is that they did a study of months worth of drop data
representing tens, possibly hundreds of thousands of instance raids,
and _mistakenly_ came to the conclusion that the drop rates did not
differ noticeably from what they were supposed to be. Am I correct
that this is what you're suggesting.

> > > I find it very odd to be answered with a "you can't proof it, therefor
> > > your claim is invalid" and at the same time not willing to proof the
> > > other claim "it works as intended" without any proof and without even
> > > the acknowledgment that this, too, would need a proof of the same
> > > scope...
> >
> > I'm not saying "you can't prove it, your claim is invalid". What I'm
> > saying is that there's no reason to assume some kind of non-randomness
> > at work when the streaks we see in loot can be explained simply as the
> > normal behavior of a randomly determined distribution.
>
> I'm claiming the mechanism of chosing the distribution (loottable) that in
> effect "creates" the drops is broken. I've detailed how I believe it works
> at the current time and why this allows some to have "balanced" loot and
> others to have completely unbalanced loot at the same time.

And I've shown that there's no evidence of loot bias that can't be
explained by the natural streakiness in truly random sequences.

> > > It's actually a prime example of hypocrysis. The "seeing patterns or
> > > streaks is normal" is just as anecdotical as all the stuff I have said
> > > for this topic.
> >
> > No it's not anecdotal at all. It's well-known in statistics that
> > streaks happen.
>
> Yes, and if it's too many streak, especially when they appear all at once,
> that means that the base hypothesis is wrong and the distribution is not
> what was hypotized. If I had enough samples I'd give you the chi-squared
> test in a second to show you what I mean. I don't have them and I won't get
> them.

See what you're saying here? You admit you don't have enough samples
to draw this conclusion, but you say that if you did it would support
your theory. How much more subjective and biased can you be? It's
like you've got blind faith in this. Absolutely amazing.

> But the values are sooooo off limits for us and some other groups
> that I lean that much out of the window to say it out loud for all to see:
> You don't even need statistics to see from afar that something's broken
> _for some_ groups.

Yes, you do need statistics. Otherwise it's just so much anecdotal
evidence and hand-waving.

> > > And at the end of the day, playing advocatus diabolii: why is there a
> > > token system in Zul'gurub and not a pure drops system and why is the
> > > same coming for Ahn'Quiraj? Does Blizzard not trust its own drop system
> > > anymore?
> >
> > A far more reasonable explanation is that they realized it's a pain in
> > the ass to keep farming an instance over and over just for the 1/8
> > chance that your last member will finally get his Tier 1 chest.
>
> Especially if the 1/8 turns to 1/100 because the distribution is not random
> uniform _for some_ due to a wrong/bugged mechanism that chooses the
> loottables (which all in all viewed together DO have a random uniform
> distribution)

So your theory now is that Blizzard knows their loot table is bugged,
and rather than fixing it they've implemented the token system to get
around it?

> > > That's exactly my point, as I already wrote. I'd discount it if it were
> > > only ONE streak. It's many concurrently, for weeks (as long as we had
> > > the same person opening the instance). It sucks.
> >
> > There's another reason I don't buy this. For the random-number-seed
> > theory to hold water, you'd have to get _exactly_ the same drops from
> > week to week -- every single one would have to be identical. That's
> > the way random number generators work; seeding them with the same
> > number produces exactly the same result every single time. Seeding
> > them with a similar number does _not_ produce a similar result -- it
> > produces a completely different result. Unless you were getting
> > exactly the same loot from week to week -- and by your own admission
> > you were not -- then you weren't running with the same random number
> > seed.
>
> See the message where I described how I see it works. This theory of mine
> does in fact explain it why some are affected and others aren't and also
> why it is NOT 100% guaranteed to get drop X.

You'll have to restate it then. I saw no such explanation.

> > You have my sympathies if you've gotten a string of loot you don't
> > like, but I don't believe for a second that it's anything other than
> > your own luck.
>
> Not just mine.. but oh well. This really is my last answer to this topic.
> It eats too much time for too few gains. I need to get some strictly
> deterministic work get done now. Thanks for listening and I wish you all
> better luck than what happend to us. If not, well I explained how you can
> safe the day once the problem is identified.

I'm not the superstitious type, thanks.

Adam

Adam Russell

unread,
Dec 8, 2005, 12:50:41 PM12/8/05
to

"Rene" <inv...@email.addr> wrote in message
news:20051208021928.986$r...@newsreader.com...

> adamc...@fastmail.us wrote:
>> BombayMix wrote:
>>
>> > It's human nature to see patterns in things.
>>
>> The rest of your post was well-written, informative, and insightful,
>> but this one sentence really hits the nail on the head.
>>
>> If you run MC a dozen times, you will see a streak of _something_. It
>> may be something awesome like the Talisman of Aggro. It may be
>> something terrible like Shardstrike. More likely it'll be something in
>> between. But it's a statistical near-certainty that you will see a
>> streak. That's just the way random numbers are.
>
> Indeed and I would not have said a thing, except for the simple fact that
> some groups experience THE VERY SAME STREAK WEEK AFTER WEEK AFTER WEEK.
> sorry for shouting but I believe you (not just you, Adam) just don't
> realize the extent of the problem. If you get Bindings of the Windseeker 5
> weeks in a row (droprate according to thottbot ~ 3%, legendary quality)
> AND
> other streaks concurrently, would you still cite that seeing patterns is
> human and "normal" ?

My experience was that people thought we were having too many of the same
drops till we went back over the drop logs we kept and realized the patterns
we thought we saw simply werent there. Going by your gut to determine
statistical patterns is always the wrong way. Did you keep records of the
drops?


Adam Russell

unread,
Dec 8, 2005, 1:03:12 PM12/8/05
to

"Christian Stauffer" <wildc...@bluewin.ch> wrote in message
news:4397edb9$0$1164$5402...@news.sunrise.ch...

> "Adam Russell" <adamr...@sbcglobal.net.invalid> wrote:
>
>> "Christian Stauffer" <wildc...@bluewin.ch> wrote:
>>
>>> Hrm.... So, what you say is:
>>> A connection between the raid leader and the drop rates is bullshit,
>>> because there is no bullet proof evidence.
>>
>> No, that is not what I said.
>
> If that's not what you said, why do you repeat it then in this post?

Again, I didnt say the reason was "because there is no bulletproof
evidence". I said it because there is "no testable theory offered". I also
have a hard time believing that you dont get that point. It was stated
quite simply, twice.

Again, this is my *opinion* that theories about statistics that use gut
feelings and dont propose actual testable statements are almost assuradely
bullshit.

Brian

unread,
Dec 8, 2005, 4:50:16 PM12/8/05
to
"Bother!", said Pooh, as he read Adam Russell's latest post to
alt.games.warcraft.

Okay, testable theory:

"Given a relatively stable raid group capable of fully clearing Molten
Core, the raid group that always forms under the same leader will
eventually see statistical anomalies in the drop rates of certain items,
whereas the raid group that reforms under a new raid leader for each raid
will not see significant statistical anomallies in drop percentages over
time."

There ya go.

I don't have the *means* to test this theory, since I don't have access to
the drop records and leadership of any such raid guild. But it's testable.

Brian
--
ICQ#: 68214833 | AIM: LineNoise54
.
Cahn's Axiom: When all else fails, read the instructions.

David Carson

unread,
Dec 8, 2005, 6:33:32 PM12/8/05
to
Adam Russell wrote:
> My experience was that people thought we were having too many of the same
> drops till we went back over the drop logs we kept and realized the patterns
> we thought we saw simply werent there. Going by your gut to determine
> statistical patterns is always the wrong way. Did you keep records of the
> drops?

My guild has full records of our MC drops, since our DKP system is all
handled on our web page. For the curious, here's the totals of our tier
1 set drops (this is from weekly raids going from our first Lucifron
kills back in July to our Luci-Domo clears of the last month or so -
haven't killed Ragnaros yet)

Arcanist - 21 pieces
Cenarion - 22 pieces
Felheart - 27 pieces
Giantstalker - 27 pieces (no Breastplate yet)
Lawbringer - 34 pieces
Might - 24 pieces
Nightslayer - 21 pieces
Prophecy - 17 pieces (no Robes yet)

Is that unbalanced? You be the judge. It's mostly pretty close but we've
got twice as much Lawbringer as Prophecy..

Cheers!
David...

Brian

unread,
Dec 8, 2005, 11:35:32 PM12/8/05
to
"Bother!", said Pooh, as he read David Carson's latest post to
alt.games.warcraft.

>Adam Russell wrote:

We changed our raid leader about halfway through our runs, and actually
started rotating every week towards the end (our guild imploded a bit, and
the remnant that is still KoD can't field enough people for MC anymore).
But here's the total for us.

17 Arcanist
17 Might
19 Prophecy
23 Cenarion
24 Felheart
21 Giantstalker
21 Lawbringer
24 Nightslayer

If I narrowed the time span, I'm sure I could develop significantly more
skewed results, as our first raid leader had Cenarion and Felheart dropping
like candy, and it was three or four weeks before we saw a *single*
Prophecy drop.

Brian
--
ICQ#: 68214833 | AIM: LineNoise54
.

A waist is a terrible thing to mind.

0 new messages