My guild "Virtue" isn't listed on the Scilla progression. I also noticed that when I tried to search for "Virtue" on the armory, and click the guilds tab, it isn't listed on the first page. When I try to click page 2 or 3, it reverts to the list of players with the name of Virtue. I'm not sure if the problems are related, just thought I'd report it.

I sent an email to the web admin, but has anyone else come across the problem?

From the FAQ on the site, it says that to be listed you need at least 20 points, which equates to being about half way through Karazhan. Given Curator's placing, I guess this will be similar to the number of guilds with access to Tier 4 loot.

A small, but somewhat amusing (and embarrassing) problem arises with our guild. We've killed Kaz'rogal, but have yet to get any loot that anyone in the guild would actually wear. Specifically we've gotten the BM hunter shoulders and the Moonkin belt, neither of which anyone is using. Thus, the data mine misses it. This has been somewhat addressed previously as the sidegrade problem, but it's something of its own class.

I'll chime in with the choir here. Very well done, thank you for the time and effort you must have put into this Aspir.

Hmm... there have been quite a few suggestions on how to increase the accuracy and eliminate problems with /gquitting and realm transfers. If I'm not mistaken, the suggestions so far can be summed up as follows.

1. Add "must have at least x number of loots in guild (and perhaps not all on the same character)" variable.

2. Add "logical check", i.e. if a character has loot from Prince, but noone else in the guild has loot from the previous (linear) bosses, then that guild could not be said to have killed Prince.

3. Manual correction/reporting.

I couldn't see that anyone has suggested it yet, but how about also making an ingame WoWjutsu addon which collects and reports data much like the Thottbot/Wowhead/Allakazham-addons? I realize that an addon could be abused by false reports, but still, the more sources the better. This also helps with the paranoid "log out clothes"-wearers.

To eliminate abuse, the addon reports could be made subsidiary to the armory reports. I.e. if an addon reports loot on a character which the armory doesn't show yet, the update could take place when the armory has been updated or when a large enough number of addon reports have been submitted.

An addon would not help eliminate the problems with /gquitting and realm transfers in itself, but it would increase the available data (making cross checking possible) and perhaps be a little faster than the armory at times.

We got those loots as well. I wanted to shout angelista's grasp into the microphone for some reason.

Still, this system beats the one over at Bosskillers.
Submit a picture, submit a dkp page showing the date you kill it (bit ridiculous). Make sure you got that guild first kill shot of prince and curator.

Don't get me wrong though - bosskillers is doing a good job.. its just a bit tedious.

World of raids really missed the boat big time in keeping their website relevant as well. I mean I hate to be a critic and an ass, but they were the website to go to for rankings and even sometimes news. Now they're a convoluted forum or something (I can't even tell whats going on when I'm at that website).

My only question is, is that this system would appear to award guilds with healthy alt kara farming operations in full swing (potentially more than it should). I realize its not a mountain of points, and very few if any guilds run multiple 25man ID's, however it is not uncommon for 2-3 kara id's per week. My guild is a 4 night a week guild, we ran 3 kara's up until a couple months ago, now we maybe run one per week if people are bored enough on alts on a friday night. My alt Warrior is in probably 7 or 8 kara epics at least and he shows up in my guilds ranking, I was wondering if you didn't care how much continued kara clears per week would slowly skew guild rankings relative to eachother or not, but that is one thing that is mildly inaccurate.

Every clear of kara with no rot OR where people looted everything and logged out until it showed up on their loot history(like a plate DE'er!), would be worth a miniscule .135 points, however, 10 clears is a free point, the more alts the more points over time. Maybe that doesn't mean much in the grand scheme but when you are comparing a pair of guilds side by side at near same progress, what kind of farming of kara they have done will probably upset the order. I think Kara drops should probably be rated as nothing due to the nature of multiple ID's of 10man's vs 25's. IF you are going for the most accuracy. I guess Kara only clearing guilds would want this detail, but as a 25man raiding guild Kara doesn't matter at all to me.

Yes I realise this is trivial, I also found the idea of having a plate class enchanter loot all the rot and log out for points in bizzare gear.

This has been mentioned many times before and is in the FAQ.

" I don't currently have the European guilds indexed, but it is in the works and will be made available soon."

That'll teach me to jump on the post button (my brain "occulted" the 2nd and 3rd page of the thread).

So from now on you allways make sure one item from a new boss isn't sharded if the ranking page is important to you. Sure beats any other alternative.

One of the top two guilds on my server killed Kael first, I think due to missing out on the first vashj kill to the other leading guild, so it can happen.

How much work would it be to separate out the "Progress" chart into T4/T5/T6/Combined lists? Then you could look at what you're interested in. Frankly, someone looking for a group may *want* to focus on groups with a healthy alt/newbie Kara farming status, as it indicates a different type of group from a solely progression-focussed group.

Regarding the transfer issue, what if you saved guild roster histories? IE: the last four times 'BossSlayers' was crawled everything was normal. This latest time though, 'Bob' is now on the roster when he wasn't the last 4 times. 'Bob' also has Al'ar loot when the guild only seems to have downed Mag. Hey look, none of the Kara+ guilds on this server have a 'Bob' on their roster any time in recent history either. Obviously a transfer, Bob's loot is ignored. As with everything it isn't a perfect system, but I'd think it would be accurate enough.

I lack sleep... is there an obvious problem with this that I'm missing?

Anyways, great job! New community features are always great to have.

Very creative use of the armory! Nice one.
Now I only wish someone bored enough will take the time and write a more lightweight front-end for it (armory)!

I should have looked better at the main page.

I assume that 676642 players actually means 676642 characters.
So that means 50 lvl70 characters average per guild.
How many of those would be alts ?

My guess would be that at least a third of them would be alts.
I would also guess that the amount of alts is larger in smaller guilds.
In larger guild (who raid 25-man content) people have more to do, so they will be less inclined to level an alt.

My guesstimate: 500k players are playing a character in Karazhan.
This is assuming that all players with a lvl70 character in a guild that does Karazhan, are in fact invited to join a Karazhan raid once in a while.
Out of the 2 million US players, that is actually a pretty amazing high number, imho.
25%.


Next question:
Does the 100% in Karazhan mean:
a) 100% of the 13545 guilds that were investigated ?
b) 100% of the ranked guilds ?
c) somewhere in between ?

75% of the 100% have killed Prince Malchezaar.
That indicates that the 100% can not mean 100% of the 13.5k guilds investigated.
The 100% must refer to 10538 guilds who are ranked.

So let's calculate what percentage of the playerbase has killed which raidbosses ?
Assuming 2 million US players.
Assuming 40 active raiders per guild.


7970 guilds killed Prince Malchazeen. 319k players. 15.9%
5543 guilds killed Maulgar. 222k players. 11.1%
4317 guilds killed Gruul. 165k players. 8.3%
2012 guilds killed Magtheridon. 80k players. 4%

1003 guilds killed Kazzak. 40k players. 2%
770 guilds killed Doomwalker. 31k players. 1.5%

1228 guilds are in SSC. 49k players. 2.45%
254 guilds killed Lady Vashj. 10k players. 0.5%.

1398 guilds killed Void Reaver. 56k players. 2.79%
279 guilds killed more than VR in TK. 11k players. 0.56%
87 guilds killed Kael'Thas. 3500 players. 0.17%

A suggestion that might help to weed out guild transfers.

Look at the reputation of characters with drops from instances.
Reputation with The Violet Eye indicates time spend in Karazhan.
Reputation with Ashtongue Deathsworn indicates time spend in BT.

Look at the 10 players with most reputation in Karazhan.
If one player has substantially more reputation that the rest of those 10, then that is an indication that he has transfered from another guild. Do not count his gear towards boss kills. Even if a guild decided to gear a tank first, or the drops were all plate, the reputation will still indicate that they have played together.

The same thing can be done with reputation of the top 25 players in BT.

Unfortunately there is no such indicator for SSC or TK, I think.

Hope this helps.

Those numbers look way more legit, than those quickshot statistics on the frontpage.

They should divide those statistics from the ranking evaluation way more. The numbers get way more skewed if they just try to find trends in the ranking data. Tracking alts and shared characters will be quite impossible. There are huge amounts of data to be processed and with the armory's slow performance this will actually take some time. I think what we look for, is something completely different than what the site is trying to achieve right now.

From what I have read so far (and considering this a ranking system judged by content completed), I would assume that "100% Karazhan" means that 100% of those Ranked guilds have killed something in Karazhan. Unless the "Total guilds" count how many guilds were found to have a character with Karazhan loot but did not meet the requirements for ranking (ie: require 10 "items" in order to be eligible for a rank - hence avoiding pug Karazhan's, that have a random player in a solo guild, to attain a rank.)

I would also like to add that your original math "guestimates" seem correct. Since were mainly concerned with Ranked players you can calculate the following (Not sure if this is how you did it or it was really a good guess):

(10538RG / 13545TG) * 676642TP = (526427 Ranked Players / 10538RG) = 50 RP/RG

Finally, you can use the same idea to filter out alts as you would for transfers. Using the reputation pane, you can evaluate with the rest of the guilds ranked players which character is an alt or not. Obviously this is not 100% accurate as you will have players with very well established alts that have better reputation than some mains (Altough I don't think this small percentage would be a huge factor compared to how many you'll be able to filter out.)

I was wondering the same thing. My guild only has a handful of members, but most of them (myself included) raid in conjunction with a guild that usually somewhere in the top 5 on my server.

Definitely looking forward to seeing this working for the EU as well!

Not sure if anyone has posted this yet, but http://www.wowjutsu.com/us/detheroc/...html#_karazhan is a good example of a transfer inflating a ranking. One person has every single piece of loot outside of karazhan, so they are number three on the server.

This is just 1 of those things that this system will never be able to cater for.

On Earthen Ring (EU) of the 8 groups that have killed Magtheridon only 3 are actually guilds. The other 5 generally have members from at least 5 different guilds if not more. It's even worse if I look at the groups in Gruul's Lair (haven't worked out stats for that as there are a few groups that I'm not sure whether they are guilds or not). If you looked at my guilds stats for example you'd assume we'd killed Hydross and Void Reaver when in actual fact we've never actually done a guild raid anywhere. Even if you tried to some fancy checking to discredit us a viable raid guild it would be difficult as we actually have enough people to raid and most of the level 70s are geared up in Karazhan/Maulgar epics at least. They've just got that gear from different raid groups.

Hello, Euro ER brother! Same thing on the US ER; my group has generated most of the loot for the guilds ranked #15 and #22 at present, while a second raid coalition is a bit further ahead than us and their loot is spread out over probably five or six guilds. But that's OK; it's just not as useful a tool for us. It's still pretty interesting.

Hmm... well... I'd say that the system works anyway. If I've understood it correctly, what the ranking actually measures is if a certain group of players have killed a particular boss, and, based on the items used by those characters, roughly how many times they have done so.

So, if several guilds kill a boss in a joint team effort, then both guilds could be said to have killed the boss. However, the overall ranking of those two guilds will be lower than that of another guild who has also killed the same boss, but done so without mixing with another guild. This is simply because in the first case the two guilds share the items, while the in the second case all items stay within the same guild.

(Ok, this is a meta debate, but what is a guild anyway? If only one person from guild A tags along with 24 people from guild B to kill Gruul, does that mean that guild A "has killed Gruul"? Most people would say no, and in a sense I agree. But, consider then if 25 people from 25 different guilds join together and kill Gruul. Using the previous argument, none of those players' guilds have then killed Gruul. Hence, Gruul... ehrm... has not been killed? Eh no, wait, lets start over. This should mean that a guild has to make up more than a certain percentage of the raid group to be said to have killed a boss?

Yes, I actually argue for a living. But sadly not Monthy Python style. )