I would like to revive this thread in order to present to you the klickers tracker:
Guild Ranking - klickers Tracker - World of Warcraft

It is a World of Warcraft tracking web site with separated 25-man and 10-man guild rankings, realm rankings, raid progression and boss encounter statistics as well as character tracking. I believe that the following properties are of particular interest:

• Best possible separation of 25-man and 10-man raid progression
• Extensive diagrams and statistics on raid progress and individual boss encounters. These also answer the question what encounters are relatively most difficult on what raid format (Thanks to songster for suggesting this functionality.)
• Reliable character tracking as characters change name, guild, faction or realm.
• Guild rank and guild status plugins for external web sites in a compact single-line format suitable for embedding in articles.
• The klickers tracker is free to use and completely free of advertising.

I am looking forward to your feedback.

Very nice, seems a lot better than the existing systems especially for separating and comparing 10-man and 25-man. There are two things I would like to comment on though. Firstly, I don't quite get what the difference is between "25-man guilds" and "25-man kills". The numbers seem to almost match on every boss. Perhaps you could elaborate a bit on this. Secondly, how does the geometric mean calculation handle bosses that no one has killed yet on a specific size? Geometric means are only possible for positive numbers so I would assume these bosses are either ignored or given a fictive very small number of kills for the sake of calculation.

I am not sure how you are picking which guilds to call 25-man guilds and which are 10-man guilds, but I do not think that guilds that only raid on 10 man should be included in the 25-man guild ranking. If you start seperating guilds based on the size of the raids they clear instead of the size of the guild itself, this would be useful. Right now, there are 10-man raiding guilds mixed into the 25-man rankings which makes it kind of pointless in my opinion.

Edit: Actually the guild I am questioning's stats say they killed heroic Halfus on 25-man but everything else on 10 man. I still don't think it makes sense to rank them with 25-man guilds because they did one boss on 25-man but the other 13 on 10-man.

"25-man guild" refers to a guild in the 25-man bracket (regardless of whether the kill is achieved with 10 players or 25 players.) "25-man kill" refers to a kill achieved with 25 players (regardless of whether the guild is in the 10-man or 25-man bracket.) As guilds are assigned to brackets based on the majority of their kills, the numbers are often very close. In fact, they have beomce that close that I am considering removing the "kill" columns from the encounters overview page (and just leave them on the encounter detail page.)

The geometric mean used in the relative difficulty calculation considers only bosses with at least one kill on both raid sizes. So yes, you are correct in that bosses without at least one kill on both raid sizes are ignored in this calculation.

The tracker assigns guilds to brackets based on the majority of their kills - with the heroic bracket overriding the normal bracket, which is what is happening for the guild you are questioning. As a guild moves into heroic progression, the normal modes of the past becomes less relevant which is why the heroic bracket overrides the normal bracket. However, the normal mode kills are still clearly documented including the raid size in which they were achieved. With many guilds switching from 10-man to 25-man for their heroic progression, it would be hard to justify to leave these guilds in the 10-man bracket while they are actually doing 25-man heroic progression. If the guild in question continues its heroic progression with 25 players, then the current bracket assignment is correct. If they continue their heroic progression with 10 players, they will be switched to the 10-man bracket as they achieve their next heroic kill.

The FAQ contains some more detailed information on the workings of the tracker.

A quick look at our realm's progression reveals some problems: US-Moonrunner Guild Ranking - klickers Tracker - World of Warcraft

- Dates for many of the early kills are improperly recognized as the date of the 2nd or 3rd time the boss was killed.
- Realm ranks are inconsistent and incorrect, with 2 guilds credited with 1st kills of Halfus / Magmaw / Conclave despite very different display dates.
- Infernal Affairs is incorrectly identified as killing Magmaw-25, around the time we had two 10-mans killing it at once after 3 previous kills.
- Equitas is identified as first killing Halfus-25 on Jan 4th, which they may have done but we know they killed it in 10-man on Dec 16th.

Ad (1). This is difficult to analyze without further details. Possible general reasons for this are that a kill requires a minimum number of guild membes to be recognized as a guild kill and that the Conclave of Wind achievement criterion was bugged when Cataclysm launched resulting in kill credit not being awarded to players that were dead when the encounter ended. Feel free to append ?analyze to any guild URL to view the aggregated kill counts the reasoning engine is working with when processing kills.

Ad (2). I cannot follow you on this. Realm rank 1 would appear to be uniquely assigned for Halfus and Magmaw, and shared by Equitas and Internal Affairs for Conclave of Wind which was defeated on December 13 by both guilds. Note that there is a (realm) rank 1 for both the 25-man and the 10-man first kill.

Ad (3). This has been fixed with todays push.

Ad (4). The tracker is currently giving unconditional preference to 25-man kills. I am considering altering this behaviour to give preference to kills machting the guild bracket.

The join dates may be due to characters below the level of 85 currently being ignored by the tracker.

The URL parameter should have read ?analyze if it is the first query argument.

Bethink - can you post here an example of calculation "relative encounter difficulty". May be on Heroic: Al'Akir. I don't clearly understand how to do it from this guide. But numbers seems totaly wrong.

Here is an example of the calculation for the normal modes. The following table shows the number of kills on each raid size as well as the ratio of these kills:

Encounter	25-Man Kills	10-Man Kills	Ratio
Magmaw	3796	22916	0.165648455
Omnotron Defense System	3663	20307	0.180381149
Maloriak	3563	12995	0.274182378
Atramedes	2828	10369	0.27273604
Chimaeron	2333	8561	0.272514893
Nefarian	884	1566	0.56449553
Halfus Wyrmbreaker	3438	20506	0.167658246
Valiona and Theralion	3076	16071	0.19140066
Ascendant Council	1637	8101	0.202073818
Cho'gall	1626	4378	0.371402467
Conclave of Wind	2624	14645	0.179173779
Al'Akir	690	4600	0.15

The geometric mean of the ratio is 0.2303.

Now let us calculate the relative encounter difficulty for Al'Akir on 25-man.

1. The number of 25-man guilds is 4144.
2. The number of 10-man Al'Akir kills is 4603.
3. The expected number of 25-man kills is the number of 10-man kills multplied by the ratio, i.e. 4603 * 0.2303 = 1060. (This is less than the 4144 25-man guilds, so there is no capping.)
4. The effective number of 25-man Al'Akir kills is 690.
5. The relative encounter difficulty is defined as (expected - effective) / expected, i.e. (1060 - 690) / 1060 = 0.35

Hopefully, I did a better job at explaining the calculation with this example than I did in the FAQ.

As for Heroic: Al'Akir, the calculation is likely to be more in the realm of divination than mathematics at this time as there are only two kills of this encounter so far (one on each raid size.) The relative difficulty calculation becomes more meaningful and stable as there are more kills. Also note that heroic encounters are calculated independent of normal mode encounters, so the current heroic ratio is different from the current normal ratio used in the example above.

I have added a personalized news feature to the klickers tracker:
My News - klickers Tracker - World of Warcraft

The personalized news page displays aggregated news about the guilds and characters that you have favorited. You are informed about their raid progress and learn when they change name, guild or realm. Use of this feature requires you to log in to the tracker web site using your Google account (which can be created free of charge if you do not already have one.)

I just wanted to let you know that Wowtrack now supports progress tracking based on the loot obtained by guild members - in addition to the existing achievement based method. This is of particular importance for the ranking of guilds that have formed during progress, or that have changed raid size while progressing. More details can be found in this blog post: Wowtrack: Loot Based Guild Progress Tracking

This is not supported. Given the current armory output, I see no way of doing this in a reliable way in the general case. Therefore, there are currently no plans of implementing this. Of course, plans may change as the armory evolves.

Wowtrack now provides push notifications on your iPhone, iPod Touch and iPad. You can receive notifications on raid progress and other events relating to your favorite guilds and characters. More information can be found in this blog post:
Wowtrack: Push Notifications on the iPhone