slashdotted!

Apologies to everybody trying to get to my Google App Engine site. That wow.com shout-out has run me right out of bandwidth. I’ve been linked to from wow.com before without being overwhelmed by it, so this has taken me by surprise a bit.

Normal traffic for the site is well within the bandwidth limits that Google sets, so I’m sure I’ll be baaack soon. Just bookmark the link and try again when the storm has passed…

statistics for dummies

Thanks to all my visitors from wow.com who left comments on various posts. One thing that came up in a  couple of places is the suggestion to use more powerful statistical techniques on the data. Great idea! Unfortunately I took my last stats course at Queensland University in the first semester of 1978… and no, that’s not a typo… and I’m sure things have moved on a bit since then.

So I’d like to issue an open invitation to anybody who does have the time and the statistical skills to help play around with the data and see what we can come up with. I’m happy to have guest posts here, with the proper attribution for you, or you can post on your own blog.

We’ll have to see how to handle access to the data, but I’m sure there’s somewhere in cyberspace that’ll help with that problem.

Just leave a comment here. Your email address stays unpublished but I can see it and I’ll reply as soon as I can.

the first rule of fight club

Thanks to Epicsnakehips who left a comment on my last post (about class popularity) to the effect that the less popular classes like rogues and hunters don’t seem so unpopular in PvP. No, indeed they don’t; your basic rogue may be passé in PvE, but spend just a bit of time in a level 19 battleground and you’ll have rogues up to your armpits.

That comment has inspired me to improve my reports on class distribution. You can find two new tables over at my Google App Engine site, which show the class breakdown across PvE and PvP realms. As well, I’m working on a table that shows the same data for x9 BG twinks (where you’ll find that rogues, warlocks and hunters are all popular choices at many of the x9 levels).

Here’s the breakdown on PvP and RPPvP realms:

Class	Popularity	Tree One	Tree Two	Tree Three
Paladin	15.7%	Holy: 4.4%	Protection: 4.8%	Retribution: 6.5%
Death Knight	13.6%	Blood: 5.8%	Frost: 3.6%	Unholy: 4.2%
Druid	11.5%	Balance: 2.7%	Feral Combat: 4.1%	Restoration: 4.6%
Priest	9.7%	Discipline: 3.0%	Holy: 3.2%	Shadow: 3.6%
Warrior	9.2%	Arms: 3.0%	Fury: 2.5%	Protection: 3.7%
Mage	8.6%	Arcane: 4.0%	Fire: 1.9%	Frost: 2.7%
Shaman	8.5%	Elemental: 2.9%	Enhancement: 2.0%	Restoration: 3.5%
Rogue	8.5%	Assassination: 5.2%	Combat: 2.8%	Subtlety: 0.5%
Hunter	7.4%	Beast Mastery: 1.5%	Marksmanship: 2.2%	Survival: 3.6%
Warlock	7.4%	Affliction: 2.2%	Demonology: 1.1%	Destruction: 4.1%

It is interesting that the basic popularity of the classes doesn’t change all that much between PvE and PvP realms. I was expecting rogues and hunters especially to move way up in ranking.

The basic issue seems clear enough – classes like rogues and warlocks are near the bottom because they have only one tree that PvPers want to use. The destruction warlock at 4.1% and the assassination rogue at 5.2% seem competitive with DK and paladin builds, but DKs and paladins have all three trees as competitive in world PvP.

What surprises me is that the hunter is not more popular on PvP realms. They certainly are popular and effective in battleground PvP.

appetite for destruction

… seems to be running at about 4%. But more on that in a second. I’ve updated my Google App Engine site for patch 3.3. Some of the twink reports are a bit iffy because of the small sample size so I’ll refresh them again next week when I’ve got more data. Not a lot of people playing level 59 twink Shaman, it would seem.

Meanwhile, I’ve added a couple of new general reports, including the ever-popular class distribution table. This time, I’ve added a view of the overall popularity of the 3 talent trees within each class. Note that the percentages for each tree are based on the total level 80 population and not on the count of the class. Destruction warlocks are 3.9%, which means 3.9% of all level 80 toons are destruction warlocks. And the query is based on the active spec, where the toon has dual specs. I’m hoping that, statistically speaking, the sample is playing with their preferred spec.

Class	Popularity	Tree 1	Tree 2	Tree 3
Paladin	15.4%	Holy: 4.2%	Protection: 4.8%	Retribution: 6.3%
Death Knight	14.0%	Blood: 6.1%	Frost: 3.9%	Unholy: 4.1%
Druid	11.5%	Balance: 2.9%	Feral Combat: 4.1%	Restoration: 4.5%
Priest	9.8%	Discipline: 2.7%	Holy: 3.4%	Shadow: 3.7%
Warrior	9.4%	Arms: 2.9%	Fury: 2.6%	Protection: 3.9%
Mage	8.7%	Arcane: 4.0%	Fire: 2.2%	Frost: 2.5%
Shaman	8.4%	Elemental: 2.9%	Enhancement: 2.0%	Restoration: 3.5%
Hunter	8.1%	Beast Mastery: 2.1%	Marksmanship: 2.2%	Survival: 3.7%
Rogue	7.4%	Assassination: 4.1%	Combat: 2.8%	Subtlety: 0.5%
Warlock	7.4%	Affliction: 2.1%	Demonology: 1.4%	Destruction: 3.9%

So we have the retribution pally as the most popular at 6.3% down to the subtlety rogue the least popular at 0.5%.

Comparing this to what we had last time, the most notable thing is the decline of the hunter. Paladins and DKs continue to rule. And not a lot of love for the rogue or the warlock.

dances with patches, part deux

This sounds kinda ominous… Meanwhile I’m just cranking out new reports for patch 3.3. Should be up by early next week.

blog status

I trust everybody had a happy Christmas; I certainly did. Unfortunately, the post-Christmas weather down here has moved into what I call the “traditional Christmas heatwave” – a few mad days that seem to arrive pretty much on time each year now. I sure hope there is no such thing as global warming; I’m not going to like it one bit, if this is a taste of the future.

Its 38° outside (no, that’s Celsius; 100° F) so I’m hiding indoors until it blows over. A perfect excuse to update the site so I’m scanning the armoury now for patch 3.3 data. Looks like there have been no significant changes to the XML (not the bits I use, anyway). I should have enough data to start updating the reports in a few more days.

dances with patches

I’ll be updating the data for patch 3.3 in due course. There’s a lot of class changes, so we’ll give it a couple of weeks to percolate through the playersphere. But there’s nothing much new in the XML that I can see, unfortunately.

There are also hints of interesting new armoury features to come. I’m not sure that very much can be datamined from them, but character activity notifications via RSS may hold real possibilities.

Meanwhile I’m working slowly through a to-do list based on your requests and suggestions. I’ve got an economics page up over at my Google site. I’m adding consolidated reports – ie across the total population – on the most popular gems, glyphs, enchants and crafted gear. I’m also looking at a similar set of reports based on my twink samples – that’s where some of the big spenders are so it would be interesting to see an overall picture of where the cash goes. They’re not spending up big on Haris Pilton bags, that’s for sure.

smarter than the average bear

Well only just… I’ve got some results from my attempt to divide up the feral druid population into cats and bears. We started from the fact that there is no “form” tag in the armoury XML – no direct way to count the thing we want to count. The only way to get an insight into this is to find a proxy for each of the forms – something that is in the data which can be used to separate the sheep from the goats, if you’ll pardon the mixed metaphor.

Talents seem to be the obvious choice, so long as there is one talent that bears will take and cats not and another talent that is vice versa. Glyphs are the other possibility. Whatever we choose just has to be i) something that players are highly likely to take  and ii) something that is orthogonal; something that definitely points in one direction for bears and another for cats.

But the basic problem is that there are a lot of um… how to put this politely… there are a lot of left-of-centre specs out there. Talents and glyphs are both less orthogonal than I was hoping for – many specs look a bit bearish and a bit cattish at the same time. And there is a big group that takes none of the talents or glyphs that we want to use.

That’s why I decided not to make the queries very complex – adding more talents or glyphs into the selection criteria  just increases the number of toons that fall into the grey area. Also I’ve counted specs and not toons since the original question was related to the number of druids specced for tanking.

Thanks to the commenters who made suggestions on possible talents and glyphs that might fit these criteria. I’ve run two queries against the data.

The first query counts feral druids who have Natural Reaction versus those who have Predatory Instincts. A druid with some points in Natural Reaction and none at all in Predatory Instincts might be a bear; t’other way round for cats. Those with points in neither are marked as “unknown”; those with some points in both are the “could be either” group.

The second query counts druids who have a Glyph of Maul versus those who have either a Glyph of Shred and/or a Glyph of Rip. Equipping Maul but not Shred or Rip indicates bear; Shred or Rip but no Maul indicates cat. Again we have groups with a mix of these glyphs, and, unfortunately, a huge group with none of them.

Anyway this is what we’ve got:

(Patch 3.2.2 data; sample size 16327 level 80 feral druids with 28970 specs).

Talent-based spec count:

• Bear: 30%
• Cat : 33%
• Could be either: 5%
• Unknown: 31%

Glyph-based spec count:

• Bear: 18%
• Cat: 9%
• Could be either: 15%
• Unknown: 58%

Frankly I’m still not sure how valid these numbers are, but I hope they provide a bit of insight. The talent-based count may at least provide a low-water-mark indication of the number of bearish specs in there.

earth to druids… come in please…

There’s a post over at wow.com that has set me a bit of a challenge. The post is about bear tanks, but makes the valid point that we don’t have any clear data on the popularity of the various druid forms. There’s a pretty simple reason for that – the armoury data doesn’t provide any direct way of getting such a count.

Still, we don’t let little obstacles like that get in our way. What we need are some data items that can be used as proxies for what we want to count. Unfortunately I’m far from being a druid expert, so I’m looking for suggestions on what items to use.

What we basically want is a talent, or a glyph (or maybe a gem) that bears will want to equip and cats not. And then something that’s vice versa – something that cats will have and bears not. One talent or glyph, or several… whatever makes the most sense. All suggestions on this are most welcome.

(Thanks to the commenters who have already made suggestions on other threads; I’ll be taking those comments on board.)

If I can get suggestions for both talents and glyphs then I can run more that one query and see how well the numbers match up.

I’d imagine that, with dual specs, players who liked both forms would have a spec for each. In any case, going from specs to forms and getting a count against the total druid population should tell us something interesting.

ebony and ivory

Thanks to reader Armagon who asked for a consolidated report on race distribution. This subject is covered by a variety of other sites, but there is some doubt about whether those sites are maintaining a representative sample.

It’s easy enough to produce a couple of simple tables that give us the information we need. (The data is from the patch 3.2 scan.)

Race	Popularity
Human	20 %
Blood Elf	17 %
Night Elf	16 %
Undead	10 %
Draenei	10 %
Tauren	9 %
Orc	6 %
Gnome	5 %
Dwarf	4 %
Troll	4 %

That seems to be a reasonable match to the Warcraft Realms data. There will always be some degree of sampling error in this work so everybody’s numbers have to be treated with a bit of caution.

Tables like that always make me shake my head a bit but… the fantasy RPG where everybody wants to roleplay the cute kid next door… Hey, trolls are people too you know!

Anyway, if we want the distribution of race and class then we get this. (Percentages here are based on the total population so the popularity column adds up to 100%.)

Race	Class	Popularity (%)
Blood Elf	Paladin	6.1
Blood Elf	Death Knight	3.2
Blood Elf	Priest	1.6
Blood Elf	Mage	1.6
Blood Elf	Hunter	1.4
Blood Elf	Warlock	1.4
Blood Elf	Rogue	1.3
Draenei	Shaman	3.9
Draenei	Death Knight	1.6
Draenei	Paladin	1.2
Draenei	Priest	0.8
Draenei	Hunter	0.7
Draenei	Mage	0.7
Draenei	Warrior	0.5
Dwarf	Hunter	1.3
Dwarf	Paladin	1.2
Dwarf	Priest	0.7
Dwarf	Warrior	0.7
Dwarf	Death Knight	0.4
Dwarf	Rogue	0.2
Gnome	Mage	1.7
Gnome	Warlock	1.4
Gnome	Rogue	0.9
Gnome	Death Knight	0.8
Gnome	Warrior	0.5
Human	Paladin	5.3
Human	Death Knight	3.3
Human	Mage	2.7
Human	Warlock	2.5
Human	Warrior	2.4
Human	Priest	2.1
Human	Rogue	1.7
Night Elf	Druid	6.2
Night Elf	Hunter	3.3
Night Elf	Death Knight	2
Night Elf	Rogue	2
Night Elf	Priest	1.5
Night Elf	Warrior	1.3
Orc	Death Knight	1.5
Orc	Shaman	1.3
Orc	Warrior	1.3
Orc	Hunter	1.1
Orc	Warlock	0.5
Orc	Rogue	0.3
Tauren	Druid	4.1
Tauren	Shaman	1.4
Tauren	Warrior	1.4
Tauren	Death Knight	1
Tauren	Hunter	0.6
Troll	Shaman	0.9
Troll	Hunter	0.8
Troll	Mage	0.5
Troll	Priest	0.5
Troll	Rogue	0.4
Troll	Death Knight	0.3
Troll	Warrior	0.2
Undead	Rogue	2
Undead	Warlock	1.9
Undead	Priest	1.9
Undead	Mage	1.7
Undead	Death Knight	1.3
Undead	Warrior	0.9

I’ll add this to my set of reports over at my Google site, so that it stays updated with each new scan. But it does seem that the other sites that cover population are doing a reasonable job of reporting what is really going on.