One of my favorite
WoW information sites,
Armory Data Mining and
its related blog, has now been updated for patch 3.3.3. We've
profiled the site here before, but if you're not familiar with it, it's run by a fellow named Zardoz who trawls the armory assembling statistics on race, class and spec popularity. He also gathers information on class battleground performance and professions.
From what I've seen, there haven't been any giant changes between the
patch 3.3 and
patch 3.3.3 data (you can find the former at the profile link above). Paladins are still the most popular class, followed by death knights, druids, priests and warriors. It's well worth a look if you're interested in seeing what
WoW's statistics look like right now.
Tags: armory-data-mining, class-popularity, class-population, class-statistics, data-mining, datamining, demographics, patch-3.3.3, patch-3.3.3-wow, race-popularity, wow-class-demographics, wow-class-popularity, wow-class-population, wow-class-statistics, wow-demographics, zardoz
Filed under: Analysis / Opinion, News items
May 19th 2010 at 2:00PM 
Reader Comments (Page 1 of 3)
Jamie May 19th 2010 2:03PM
If only I could create a graph to express my love for these graphs... :(
Zandalis - BS May 20th 2010 3:18AM
Someone should be sacked for showing this graph to anyone. Now even more people will think that battleground effectiveness is measured in killing blows... :(
Baba May 19th 2010 2:06PM
What we want to see (if everything is balanced) is a straight, proportional graph. People can choose a high-damage, low survivability class, or the opposite. The only real outliers here are the pallies, and the warriors need a bit of love, perhaps in a "oh shit" button, which they lack quite badly.
loop_not_defined May 19th 2010 4:31PM
Representation != power. At least, not absolutely. If a lot of people just don't like playing Druids, it'll take a *large* power difference to bring their representation above others....and it's not just a large power difference, it's a large *perceived* power difference. Popularity begets popularity, in a sense.
I feel comfortable saying "You'll never see a straight line. Ever." It would be foolish to balance classes around representation, as it would necessitate unbalancing strength to do so. Historically unpopular classes would have to be overpowered to be equal in representation, while historically popular classes would have to be underpowered. You can't tell me that's "fair".
Hollow Leviathan May 19th 2010 6:22PM
Loop, these graphs are per capita. There's nothing about the comment to which you replied or the data itself that relates to class representation. The point of the comment is that were all classes balanced, they would have proportional ratios of killing blows to deaths, with less of both meaning higher survivability traded for less offensive ability.
This is not the case, most obviously for paladins and warriors. Paladins die too seldom, warriors die somewhat too often.
Keep in mind that the graph's X and Y are 7 and 1.2, respectively. It's been designed to highlight disparities, which is the positive way of saying blowing them out of proportion. I think a graph that started at 0,0 would show these numbers are closer than they appear.
Hollow Leviathan May 19th 2010 6:23PM
OK, I lied, there is some unrelated class popularity mentioned in the second paragraph. But that still wasn't the point of the parent comment.
mpeton May 19th 2010 7:21PM
Correct, ideally you should be able to draw a strait line from the lower left hand corner to the upper right corner, and the further off the line the class, the less imbalanced they become. How far high or low, or left or right a class is doesn't really matter as much, and in fact no dot should be too close to each other on that line as that's when homogenization becomes a problem. According to the data presented currently, there is still some balancing to do, however I'd be really interested in seeing the data Blizzard has and uses to help make these determinations, and it's really a total shame things like this aren't published by them.
mpeton May 19th 2010 2:16PM
Seems to me that by reading this Pallies are a bit too hard to kill, which I think sums up my hatred for Bubble quite nicely.
Carra May 19th 2010 2:17PM
"Paladins are still the most popular class, followed by death knights, druids, priests and warriors."
Interesting, every single one of them is a hybrid. Where are the times when pure DSP like hunters were on top?
clundgren May 19th 2010 2:34PM
That would be in arenas, where "pures" have always tended to dominate.
Hybrids tend to do better in bgs where there are a lot of inexperienced players, and not a lot of target coordination. Since most hybrids have some self-healing capability, they can stay up longer.
Also, so-called "pure" classes tend to be targeted quickly since they do a lot of damage and get everyone's attention.
Once rated bgs come in, these stats will probably start to resemble arena stats more.
loop_not_defined May 19th 2010 4:43PM
Honestly, I think hybrids will normally be more popular simply because there's more roles they can perform. Pures are three flavors of the same thing.
Remember that popularity != strength.
clundgren May 19th 2010 2:19PM
It should be noted that the paladins are out of line not in power -- they are average in killing blows -- but in survivability.
However, this is not necessarily good news for paladins. They are a very survivable class, but they are also a low priority class. Ret paladins, in particular, are near the bottom of most experienced players' kill list (this is particularly obvious in arenas, where ret paladins, when found, are almost always the last class targeted). In other words, one of the reasons paladins are surviving longer is because the other side is killing the trees, then the clothies, then the rogues, then other melee, first.
Camo May 19th 2010 2:38PM
This and in BGs the abilities to outlast the enemy until help arrives or turtle and back up to your team mates. I don't like chasing a druid or pally around the BG just to be jumped by the resurrected.
Luci May 19th 2010 2:53PM
Yeah I kill all the others first because you know the paladin is just going to bubble when he has 2% hp left, heal all the way back up, and then proceed to kill you along with that rogue and hunter you should have gone after to begin with.
Veloxe May 19th 2010 2:21PM
That's some pretty neat info for sure. Even beyond just the PvP stuff. Not overly useful really, but pretty cool to just flip through it and look at the averages.
Thundrcrackr May 19th 2010 2:22PM
They got the colors all rong.
Deathknighty May 19th 2010 2:27PM
You got "wrong" all wrong.
Thundrcrackr May 19th 2010 2:30PM
wrofl.
KJP May 19th 2010 2:32PM
Am I missing something or is something on those charts mislabeled? Seems to me a death would have at most one killing blow associated with it. So how are there roughly six times as many killing blows as deaths? Is the x-axis actually HKs?
Lemons May 19th 2010 2:38PM
The X axis is probably mislabeled, instead of Killing Blows it should be Honor Kills. As an avid BGer I can tell you that 25 killing blows is probably the maximum in a BG, not the average, even if you are a paladin. Most people can barely get 4 or 5.
An honor kill is different than a killing blow, you simply have to help kill someone. Sometimes I don't even help, but I still get them from simply being in proximity of a kill.