I can't sleep. Every time I lie down, my mind refuses to stop racing and let me sleep. So I suffer in my insomnia and swear tomorrow will be the day I finally get some sleeping pills to make this stop. But, there is one thing I have figured out tends to relax me to sleep.

Data entry.

So I made a graph for you.



It's a graph of the data available from >WowJutsu's The Vault<. To the best of my knowledge, no-one's ever graphed this data. It just shows the raw % of people who raid who happen to be X class and have a piece of loot from Y instance.

-Some classes will be underrepresented because they have no need for any of the loot at that tier/at that boss. I'm looking at you, tailors.
-Lurker/Void Reaver kind of fucks up the SSC/TK data, particularly Lurker. Void Reaver drops a tier piece so pretty much everyone will get credit for a VR kill EVENTUALLY, but for some specs, Lurker drops NOTHING they'd ever want. I may redo this chart, but with a dataset that includes the average of every boss BUT Lurker, just to see if anything interesting falls out.
-The BT/MH data is going to change a LOT come 2.4, so any sort of analysis on it needs to be done before then.
-I put ZA at the end because I felt like it. Deal.
-Further production of charts depends on how severe my insomnia remains.

But with all that in mind, things that jump out at me:
-Shaman representation goes up a lot between T5 and T6, particularly in Hyjal.
-To the surprise of no-one, paladins are very popular in ZA
-Hunters are, vs their actual proportion of 70's, underrepresented among raiders at every level although their proportion really don't change much from tier to tier.

Quite interesting data.

Would be quite nice to see graphs on the overall boss and instance progression - time as well somehow.
Or per boss loot % for each class instead of the whole instance.

Keep it up, i'm sure you'll find sleep soon.

What's the reason for warrior's jump between MH and BT? Should single boss encounters such as Ros (Deaden) and Illidan (Shear) made such difference?

Also, it would have been a bit more intuitive if you used the standardized colors for each class. I kept looking at Paladin and thinking Warlock.

BT just has quite alot of plate? 3x Tier, all the additional tanking loot, various dps loot and trash drops etc...
Something like 13 plate pieces + weapons and shield, and tier items.

Probably because there's a shitton more warrior loot in BT than MH, or there are simply more healers taken to MH than BT because of dps checks? I'm not at the level of raiding to say definitively, but from how those %s pan out and the loot tables on the bosses, that seems pretty cut and dry to me.

Few Elemental Shaman will loot much in T5, particularly if their guild has a dkp system that carries over between tiers, as the Tier 5 set has terrible bonuses and is better for most other classes/roles. I believe I got three non-Resto pieces out of Tier 5, the ring off Vashj, the neck off Illidan, and the craftable boots (the Karathress trinket didn't drop much for my guild at the time). On the other hand, there was a piece I wanted off of almost every T6 boss. I'm not sure what proportion of raiding Shaman are Elemental but this may help to partially account for the jump.

Can't explain the Hyjal/BT thing though... that one's confusing to me. Why should our representation go down in BT? Drop rates on early gear really that low?

I appreciate the data but putting ZA at the end of the graph just ruins it for me, sorry. Why would you do that.

From the POV of an endgame raiding guild that was in T6 content when ZA came out, this represents the chronological progression of things; I don't mind where he has it. Another way to do the X axis is difficulty level, and at that point I would put it among the T5 instances, maybe before them and after the T4 25 mans.

Regardless, it's pretty clear why he did it, and I'd hardly say that the arrangement of the X axis could 'ruin' this graph.

I also don't fully understand what we're supposed to take away from this. Who looted how much of what doesn't seem to be a good indicator of anything other than X class is getting more loot from Y zone; or is that what the whole point of this is?

And order of the zones don't seem to make a difference to me (though I may be missing the point here), nor do the connecting lines between the data points.

The only thing this is telling us is the amount of usable gear that is placed in each zone. The graph would be much better served broken up by armor class or even by dps/caster/tank gear. Having it tied to classes is very much apples to oranges.

It just points out poor itemization. For example, it's clear that there is no useful rogue gear from Magtheradon.

Well, I tried to say that when I specified 'usable' gear in each zone. Even then it's skewed by people looting things to idle in but not use, or specifically to get Wowjutsu credit for a kill when it normally would have been sharded. I'd have to agree with you Malan, I can't see this providing data that is actually useful for anything.

It would be far more useful to compile a suggested upgrade progression with items for specific class+spec combos at each tier. That data could be graphed as points with "best item for X class with Y spec in Z slot". It would also require intimate knowledge of each class, as well as research on theorycraft (or pre-made ranked gear lists.)

Not everyone who plays WoW is into theorycraft, and oftentimes will take non-optimal items or even downgrades based on personal preferences (taking hit > crit, going for max damage instead of balancing for better sustained dps, etc.)

The chart does have some interesting data (such as the low hunter line, the low rogue plot on Mag, the high Paladin plot in ZA, and the mostly linear Shaman line till Hyjal.) It's a good starting point, hoping someone with more time than I runs with the idea .

Is the data you used comulative? I mean, if a character was spotted by WoWJutsu with T5 shoulders and later on with T6 shoulders, will that character be included on your data twice - once as a TK raider and once as a BT raider?

I believe the answer to this question is made fairly irrelevant by the fact that (as far as I can tell) one person wearing any combination of items from from (for example) SSC, TK, BT and MH would count towards the percentage in each instance. Meaning that it isn't until someone has T6 drops in every single slot that they are no longer counted as T5, etc.

You know, something was bothering me about the graph and I could not figure it out. That is it. Thanks.
I assumed most people would get at least one upgrade in each instance - Even tailors, eventually. Is that incorrect?
Edit:
Some numbers on that:

SSC: 9.17%
-Resto: 55.07%
-Enhance: 32.17%
-Ele: 12.75%

TK: 9.80%
-Resto: 61.02%
-Enhance: 25.20%
-Ele: 13.77%

Hyjal: 13.97%
-Resto: 67.85%
-Enhance: 20.47%
-Ele: 11.66%

BT: 10.62%
-Resto: 41.99%
-Enhance: 37.57%
-Ele: 20.52%

The hyjal bump is all Resto. In BT, enhance and ele shamans get a lot of growth at the expense of resto's. Whether this is some artifact of respecs or a real phenom is unknown to me.

Well; the tanking ring reward from Magtheridon's Head is quite nice for Feral Druids. But yeah, his loot is pretty awful. Gruul has multiple useful items on him, yet he is still the same ilvl (and much easier, to boot).

I think for Mag it's effort vs reward -- not so much that the reward is awful, but the effort is more than it's worth with the nice selection of heroic badge items. Something like [Soul-Eater's Handwraps] is very nice, but you have a badge item which is strictly superior: [Studious Wraps]

The upcoming nerfs to Magtheridon may make him more plausible for Gruul-level guilds to take more stabs at him.

Since the removal of the TK attunement, guilds have had no reason to work on Magtheridon. He's not a particularly hard fight if you can get people to learn it, but Void Reaver is far, far easier and with better rewards. I'd have wanted the gloves and wand off him if they didn't come put t5 level versions of those item slots on Geras. The chest I'd have wanted because Nightbane never dropped his even though t4 is a piece of crap for mages (spell pen lol), but I got the haste robe off Lynx. Thus, in general I'd say that ZA and t5 badge items have removed the need to kill Magtheridon for loot.

So while you might see some Magtheridon kills back in the 2.1-2.2 days (2.1 removed attunements?) just for loots, after 2.3 there has been little reason to bother learning the fight. That might change if there's no infernals since my guild's problem was a general lack of competent Warlocks; we always died because healers got eaten by infernals and we couldn't keep up the MT consistently once Mag arrived even though we normally had 4 channelers down at the time. Cube clicking coordination wasn't a problem as much as not having enough people alive (although being able to delegate it to 5 people who are at the top of their game/computer spec helps). While I don't think we'd get much in decent loot off him, my guild will likely make sure to get him down once.

--

I personally find this data very nearly useless. It tells you who looted items, wore them, and had them picke dup on the armory. So what? Wowjutsu never picked up me wearing Fury of the Ursine even though I played at least a month or more with it being my main DPS bracer. I assume every time it checked me I had logged out in my higher stam bracer (ones off Maiden +12stam). While that's an anecdotal issue, there are so many "mistakes" Wowjutsu can make that are skewed against certain items that such data collection is horribly biased towards items that are more likely to be worn when logging out.

There really are far too many variables for one chart to mean anything. For example, say mages score higher on a zone than rogues, does that mean:
a) That zone has better mage drops.
b) That zone is bad for rogues (or requires mages)
c) That zone's mage drops are unappealing to other caster classes
d) Rogues have heavy competition for that zone's rogue drops
e) The randomly distributed loot has fallen oddly
f) Wowjutsu isn't accurately scanning those zones.
etc...

The only way to know which choice is correct is to go to other sources, which means those other sources are where we should be looking for insight into these concerns.

Perhaps if this data was presented differently there would be more use to it.

One thing that I would be interested in knowing is, breaking down by guild progression, breakdown of class and spec who have looted something over the past 2 months. That is, not who has looted something from a t6 zone, but who, from a GUILD that is in a t6 zone, has looted something over the past 2 months.

Looking at overall loots for a 2 month period would be a good indicator of activity, and looking at progression from a guild-wide standpoint will give you a better idea of guild make up then simply where the specific loot came from.

So the report would look something like:
guilds that have cleared all content:
5% of all players who have been active over the past 2 months were enhancement shamans, etc

guilds who are still in t5 zones:
6% of all players who have been active were retribution paladins, etc.


This way, we would learn more about raid composition, without having to deal with so many random variables, like availability of loot and so on.

What exactly does the X-axis represent???

I am guessing bosses in that particular zone, then how do you sort the bosses? based on what?