[WoW] How has class popularity changed since Jan 2010?

Zardoz recently refreshed his datamining reports, and I thought this was a good opportunity to look at how class distribution is changing in WoW.

In January 2010, I wrote a post using his data to look at class popularity. But how have things changed since then?

Class Rank
Aug 2010
Class as % of level 80 chars
Aug 2010
Class Rank Jan 2010 Class as % of level 80 chars
Jan 2010
% change in pop since Jan
Paladin (1) 15.1 1 15.4 -0.3
Death Knight (2) 12.8 2 13.8 -1.0
Druid (3) 11.6 3 11.4 0.2
Priest (4) 10.1 4 9.9 0.2
Mage (5) 9.1 6 8.8 0.3
Warrior (6) 9.0 5 9.8 -0.8
Shaman (7) 8.9 7 8.4 0.5
Hunter (8) 8.3 8 8.2 0.1
Warlock (9) 7.6 10 7.4 0.2
Rogue (10) 7.5 9 7.4 0.1

You can see clearly here that balancing has nudged towards the desired effect. The two most popular classes as of January have both dropped in relative popularity, and the two least popular classes have both gained in popularity over the last 8 months.

The biggest changes are the drop in numbers of Death Knights and Warriors relative to the rest. People just aren’t rolling new Death Knights and Warriors any more, or at least not as much as other classes. Maybe the DK restriction of only one per server has started to kick in, as people roll multiples of their favourite toons?

It’s possible, but my gut feeling is that these numbers represent people’s initial reaction to announced Cataclysm class changes. They are rushing to level prepare alts who can be ready when the expansion hits. Death Knights face one of the biggest changes because instead of having three different and viable play styles for both tanking and dps, they will have one tank spec and two dps spec lines. Suddenly the golden child class doesn’t seem so very different from the rest. I’m at a loss to explain the warrior numbers too, the class has been in pretty good shape recently from my perspective.

Maybe it’s a tanking issue. Anyone who wants a tank already has one, new players are put off by the aggression of the player base. (Wishing I’d recorded the spec popularity back in January also to compare.)

Classes which saw the biggest bump in popularity as alts were shamans and mages. (Note that hunters are very close to mages in popularity on PvE servers, I mention that in more detail below.) Maybe people also really like the changes they’re getting in Cataclysm (most notably, a version of heroism/ bloodlust). I’m sure this is also a factor with shamans, who look to be getting some nice boosts in the new expansion.

The other reason that people level alts towards the end of an expansion is to fill existing holes in their raid group. It gets harder to recruit because of players burning out, so rushing an alt up as a replacement makes more sense. That could explain why the tank classes saw a relative drop in numbers (hard to gauge with druids because of the numerous specs), they’re less likely to be the spots that needed to be filled.

Zardoz also splits out the percentage distribution on PvP realms as well as PvE realms. The numbers in the table above are combined, looking at the separate numbers does give some more insight.

Hunters are massively more popular on PvE realms (8.9%) than on PvP ones (7.5%) which has skewed the combined  numbers here, bringing them much closer to mages on PvE servers.

Conversely, rogues are more popular on PvP realms (8.4%) than PvE (6.7%).

Warlocks however are identical on both. They’re just less popular in general :)

Class Distribution Redux – Who Raids?

A couple of weeks ago, I took a lot at level 80 class distribution via armoury datamining.

On the official boards, Venom has another take on this. He’s looked at the first 100 25-man kills on Professor Putricide (one of the tougher ICC bosses) and broken down the numbers of classes/ specs who were taken along in those raids.

Note: Yes, the fight may not be typical and top 100 guilds may not be typical, but it’s close enough to draw a few conclusions.

% first 100 PP Kills Class
14 Paladin
13 Druid
12 Priest
11 Shaman
9 Rogue
9 Warrior
9 Mage
8 Death Knight
8 Hunter
8 Warlock

Let’s start with some elementary maths. In a 25 man raid, each raider represents 4% of the total raid force. There are 10 classes in World of Warcraft. So in a perfectly class balanced 25 man raid, you’d take 2 of each class, and an extra member for five of the classes.

Looking at the numbers above, Ghostcrawler must be thrilled. Because that is exactly the split we’re seeing here. Even the lowest represented classes still account for 8% of the raid force.

In fact, the top four most represented classes are also hybrids, which is exactly what you would expect in a perfectly balanced position. If a class has three roles which are all viable for raiding, then a balanced raid is more likely to take one of each than to stack several members of the same class/ spec. Not only that, but paladins are also the most played characters at level 80 (see previous post as linked above) so if they are also more represented in raids that just reflects the number of available characters.

Perfect Balance? Shamans, Priests, Warriors

Shamans have the distinction of being the class where all three specs are almost equally represented. At first sight that looks astounding as a feat of balance. We’re a long way from TBC days when raids used to stack 5 resto shamans.

But actually resto shamans are currently the least represented of all the healers. When you think that resto is also the most popular of the shaman specs among the player base, that implies that a lot of people aren’t able to raid as their favourite spec. Fortunately, the other specs are also popular with raid leaders. (Note: Based on these numbers, I’m starting to be tempted to roll a Goblin Shaman myself next expansion, I never was good at picking a spec and sticking to it.)

Priest representation in these 100 raids is impressive because it almost exactly matches the spec distribution among the player base. Really priests in this expansion have been very well balanced, and this is the proof of it.

Warrior representation between roles is also close to the general population; at least if you assume most DPS warriors switch to Fury for raiding. Devs stopped citing the number of warrior tanks in  more recent comments on tanking as a reason to nerf, instead saying that warriors presented a larger sample size to compare. I think they realise that it just happens to be most players’ favourite spec, rather than anything to do with class balance.

Once a Death Knight is Enough

Ah, Death Knights. Second most popular class at level 80, but way down at 8% in raid representation.

What happened? Two things: their tanking could use a boost, and also most people who have level 80 Death Knights also have another alt which they may currently prefer. The class hasn’t dropped off the scale, and is still filling it’s share of slots. But as a hybrid, you’d expect to see them showing higher than the 8% seen here.

One Spec to Rule them All

For some classes, one spec is so superior for high end raiding that there is almost no comparison.

Arcane Mages are the single most represented spec in the whole survey, for example. Closely followed by Marksmanship Hunters, Unholy Death Knights, Assassination Rogues, and Holy Paladins.

Other specs just aren’t played at all: Frost Mage, Beastmastery Hunter, and one lone Subtlety Rogue flying the flag.

Still, in these days of Dual Spec, there’s no reason for a dps class not to go with their highest dps option unless a fight specifically requires more utility.

Are paladins popular enough yet?

I’m always fascinated when Zardoz publishes one of his regular Armoury Datamining updates – this is about as accurate a census as anyone outside Blizzard can hope to compile. It’s based on current armoury data, from which he can assemble tables of most popular classes, specs, races, and even most popular items of gear.

It’s a terrifically underused resource, but if you believe in the wisdom of crowds and want to know which are the most popular builds (for example) or which race has the most even gender split (blood elf, possibly because no-one can  tell the difference?), there’s a lot of current information to be had there.

So from Zardoz’ site, here’s the current state of the level 80 WoW population as of 21st Jan. There are ten classes, so a totally even split would give 10% of the population playing each one.

There is no information here about which characters are mains as opposed to alts, but that’s not such a big distinction as it once was.

% of level 80 characters Class
15.4 Paladin
13.8 Death Knight
11.4 Druid
9.9 Priest
9.8 Warrior
8.8 Mage
8.4 Shaman
8.2 Hunter
7.4 Rogue
7.4 Warlock

So, a few things that jump out.

  • Four out of the five most popular classes are (or can be) tanks
  • Three out of the four most popular classes are (or can be) healers
  • Four out of the five least popular classes can only dps.
  • The most popular class is over twice as popular as the least popular class.
  • The least popular class/spec combination is Subtlety specced Rogues which make up a mere 0.5% of the level 80 population.
  • Female Dwarf Rogue is still the way to go if you want to stand out, they are the least popular class/race/gender combination.

Zardoz also tabulates the most popular talent trees and specs for each class. So what role are those hybrids playing? It’s difficult for me to interpret Death Knight data since any talent tree could be a tank, so laying those aside.

Paladins: The majority are retribution, but both holy and protection are also popular secs. Paladins are relatively easy to play and have three strong trees at the moment, all of which are highly played. To put this in perspective, there are more people playing the second most popular paladin tree (Protection) than are playing the most popular druid tree (Resto).

Druid: The majority are resto, although feral isn’t far behind. I can’t tell how many of those feral druids are tanks, except to assume that it won’t be 100% of them. Balance lags behind – perhaps there are just plenty of options for people who want to play healer hybrids and not everyone wants to look like a fat owlbear while doing it.

Priest: Shadow beats out Holy for popularity by 0.4%, Discipline lags behind. So the majority of  priests are healers but a lot of people like the dps tree also. I think Blizzard has done a decent job on priest class design – fun dps, fun heals, and can use similar gear for both. I suspect that this is why they’re the more popular of the non tanking classes.

Warrior: Protection has always been the most popular Warrior spec and that’s still true. Arms and Fury are close in terms of popularity, I think well geared raiders are shifting back to Fury at the moment, but Arms is viable and still the preferred PvP build. Although Warriors are a less popular class than the other tanking classes, there are probably still more warrior tanks than druids or death knights. Clearly Paladins are by far the most popular tanking class at the moment, though.

Shaman: It has always been strange to me to see Shaman lagging behind priests and druids in the tables. They’re all healer/caster hybrids who can use similar gear for both roles. Shamans (like druids) also have the option for a melee dps spec, which is usually popular with players. Maybe people just don’t like totems, or shamans aren’t viewed as interesting to play?

So what has changed?

Here’s an older set of data from July 2009 (just after patch 3.1). Main changes are:

  1. Paladins overtake Death Knights. This is a large leap, so lots of people have levelled Paladin alts since then.
  2. Priests overtook Warriors. Not such a big percentage change, but they’ve clearly been popular alts too.
  3. Shaman overtook Hunters. Again, people looking to the hybrid classes as popular alts.

Solving the Tanking Problem?

One thing is very clear. The tanking problem isn’t that the classes are not being played, it’s that either the barriers to tanking are too high, people are enjoying the other specs more, or people just don’t want to do it.

So I’d expect to see Blizzard making tanking (even) easier, and exploring ways in Cataclysm to let tanks use melee dps gear (we know they are talking about this). People are evidently flocking to their paladins so that design has to be seen as a success.

But I do wonder how far ahead one class will be allowed to get in popularity. My guess is that paladins are a lock-in for the rest of Wrath and that Blizzard will be aiming to make other classes more appealing when they revise them for Cataclysm – we know that’s going to be a substantial amount of work.

Having said that, does it really matter if one class happens to be the most popular? Maybe people just like their knights in shining armour more than their demon-summoning warlocks?