Did you know that ice cream makes it more likely you will drown? It’s true. When ice cream sales increase, so do the number of drowning deaths. Clearly linked! Speaking of spurious correlations…

I fully expect Rift to now follow in the footsteps of WoW, in that it will decline. Vanilla and BC days had challenging content, and it’s not a surprise that sub numbers grew. WotLK made things ‘accessible’, and surprise surprise, the response was pretty meh (sub numbers dropped in the US/EU, but were offset globally by WoW launching in new regions, hence the overall stagnation). Cata tried to play both sides of the fence, but a combo of too little too late, a gimmick of progression (hard mode rehashes rather than straight-up new content), and a one-track, insult difficulty 1-85 game did it in. With no new regions to offset things, subs are dropping.

(SynCaine in the post “Accessibility killed Rift“)

World of Warcraft’s growth rate went from a perfectly stable 2 million subscribers per year during 2006 to 2009, to zero during WotLK. This was exactly the time when Blizzard changed the character progression mechanic.

(Nils in the post “Smoke and Mirrors“)

“If developers design a game which requires too much effort from the average player for too little gain, the average players will start leaving the game. “

This is the part I strongly disagree with, and WoW’s sub history does as well. Vanilla/BC, which had a MUCH harder end-game that fewer players saw to completion, saw massive growth. WotLK/Cata, with raids being cleared by all who stepped inside, have brought decline.

(SynCaine in a comment on Tobold’s post “Syncaine on Accessibility“)

The reason I bring these examples up is because this type of thinking (or lack thereof) is what I consider one of the most pernicious, asinine fallacies in any discussion of World of Warcraft. It is intellectual laziness at best, intellectual dishonesty at worse. Before I begin in earnest however, here is a slightly augmented graph from MMOData that most people refer to when they talk about WoW subs:

1) Correlation does not mean causation.

Standard preface to any claim that X means Y. Ice cream and drowning are only “linked” because there is a third factor involved.

2) Even if correlation did mean causation, why this particular correlation?

This specific point is the reason the argument is intellectually lazy. When you look at the graph, it is true what Nils and SynCaine said about there being a relatively rapid period of growth during vanilla and TBC that was not apparent after the release of Wrath. However, tying that solely (or even partially) to accessibility/character progression/difficulty/etc is a completely unsupported leap of logic.

There is zero evidence given by either author as to why it was “existence of more challenging content” and not, I dunno, the introduction of the PvP Honor System and BGs in the summer of 2005, which coincides with a 500k sub spike in WoW-West on graph. Or the release of ZG in September of that year, also suspiciously near another 500k sub bump. Or if I looked at WoW’s overall numbers like Nils does with his “2 million per year growth” argument, perhaps I could argue Patch 1.12 with it’s wildly successful:

The stage is set for intense, objective-based land battles as Horde and Alliance vie for control over important strategic positions and resources around Azeroth. Head out for Silithus and Eastern Plaguelands to engage the enemy on the field!

…was responsible for the corresponding bump of 1 million (!) subscribers. Clearly, clearly, more things like Silithus and the old Eastern Plagueland towers is just what WoW needs.

3) What does endgame accessibility/difficulty have to do with anything?

This is another intellectually lazy part of the argument that the authors never bother to address. What percentage of the playerbase ever actually makes it to the endgame, and is this percentage big enough to even impact subscription growth? That is an open question.

The best metric that I can come up with is to look at the number of guilds who killed Beasts of Northrend in 10m ToC after two years of it being out (86,187 guilds), multiply that by something charitable like 30 players, and then divide by the approximate population in the graph above while only taking into account the regions in which WoWProgress collects data (~6.5 million). The result is 39.77% of players killing the easiest boss in the easiest tier of which we have data (something like Noth the Lootbringer from Naxx 2.0 would have been better, but alas…). That actually sounds like a lot of people, and 19.88% assuming only 15 raiders per guild is not too shabby either when referring to raid content.

That said, there is no evidence whatsoever from those two that difficulty-related gyrations amongst the top 1/3rd of players doing raiding content has a meaningful impact in comparison to whatever the remaining 2/3rd non-raiders are doing. Between 2005 and 2009 the subscriber base was growing at ~25% per year. Is it even remotely likely that the top 40% had anything to do with a meaningful drop in growth rate?

4) Growth, or lack thereof, does not really mean anything other than what it is.

What I mean by this is that you cannot simply look at growth as anything other than what it is: growth. It does not mean anything else without further information. For all the talk about growth rate percentages and “the design decisions that caused them,” look at the pink line for a moment. That represents subscriptions in NA alone. Unfortunately MMOData stopped tracking that information individually (or perhaps Blizzard stopped giving it out), but the whole of TBC resulted in ~650k more subscriptions in NA over a two-year period.

Is 325k sub growth per year more than the apparent zero sub growth in the year of Wrath? Sure… but we have no real way of knowing why that growth was occurring. Was player churn less of a factor in vanilla and TBC? Was the growth simply due to the release of WoW in additional regions? Does market saturation have any impact? Do we simply ignore, I dunno, one of the worst global recessions in world history?

Oh, wait a minute… early 2009 was when the markets were at their worst? And yet WoW subs were relatively stable in most regions during that entire year? Clearly Wrath’s accessibility and stress-free raiding were the only things stopping WoW’s overall decline in a tough market, as evidenced by Cata’s increased difficulty leading to subscription loss once markets improved. QED, amirite?

The bottom line here is that you cannot use WoW subscription numbers as evidence of a claim without first proving said numbers have anything to do with said claim. Did World of Warcraft gain six million subscriptions worldwide in its first year? Yes. Was that because of the strength of its class balance? Its risk versus reward structure? Its accessibility? No one can really say; all of it would be conjecture.

Personally, I believe the initial rush was due to the strength of the IP – I know I certainly gave WoW a shot because of how much I enjoyed Warcraft 3 – and also due to the strength of the Blizzard brand. The designers also got a lot of things down perfectly that I feel other MMO designers stumble across to this day, such as letting characters jump, making solo-play possible, having quests with interesting plots, getting the reward faucet just right while questing, and so on. The tone and tenor of game balance has certainly shifted quite a bit from when I began in TBC, but where I disagree with Nils and SynCaine is that I feel that Wrath was actually a step in a better direction in most (not all) ways. Unfortunately, until the duo, and others who believe as they do, let go of the absurd notion that “the numbers” support their conclusions, it is impossible to have any rational discussion about it.

There is a separate argument as to linear raid progression vs episodic progression, but that is an OT for another time.