Friday, June 17, 2011

Smoke and Mirrors

They say that people will listen to you and remember what you have to say more, if you put it into a nice story. That's what I did in the last post. While I am certain that most of this is at least roughly correct, I am also certain that some of it is wrong. Anyway, time for a dry quintessence.

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.

The old mechanic allowed players to explore (mostly) static raids at their own pace. Raid groups could take their time to complete the content. Since 'better' raid groups progressed faster to more difficult content and without as much equipment, they faced more difficult challenges than 'not so good' raid groups.

The new mechanic forced all players into exactly one raid tier. They had a specific amount of months to complete it, before it became obsolete from a character progression point of view. The challenge level was the same for all players. Heroic raids were insignificant, as, for various reasons, they were visited by only a handful of players.

The old mechanic did not encourage players to visit dungeons once a day. And although forming a group on their own was often tiresome, when players did visit a dungeon, it usually was a pleasant social experience. Moreover, dungeons mostly weren't boring, because players didn't visit them all that often. Dungeons had different challenge levels and required different amounts of time to be completed.

The new mechanic strongly encouraged players to visit a randomly chosen dungeon once a day. They completed this dungeon within an anonymous group of other players they would never meet again. Players ran dungeons so often that they very soon found them very boring. All dungeons were adjusted to have the same challenge level and to require the same amount of time to be completed.

The old mechanic used 'random loot' that made badly-geared players gain gear very fast, while well-geared players gained gear very slowly. Most raid dungeons never became obsolete, because there was always something someone would still need. Generally, loot was given out more slowly and often became infamous.

The new mechanic used 'badge loot' that made poorly and well-geared players gain gear at roughly the same speed. All, but the current raid dungeon, were very soon obsolete from a character progression point of view. Generally, loot was given out much faster than before and was usually forgotten fast.

The classic-TBC mechanics offered reasonably tuned content that players could tackle in a pleasant social environment and at their own pace. Sure, players moaned and complained about those who had more: The unemployed, the students, the born-rich. But at the end of the day they had fun when they played in the evening. And thus they stayed subscribed.

One complain about classic-TBC was that most players never experienced the end of the story. I think this is a very legitimate complain. I think it encourages us to think about separating the overarching story from the raiding. Alternatively, raids could be transformed into leveling dungeons when an expansion hits, or the balance could be tuned carefully to ensure that eventually every raid group overgears the next raid in a way that anybody, who invests enough time, sees the end of the story.

In the absence of such a solution, the question is what is more important for players? To experience the entire story or to always have reasonably challenging content that they can do in a pleasant social environment at their own pace?
It's a hard decision, especially for someone like me. But if I ran a business, I'd chose the content over the story. Players just care more about whether they can have fun this evening, than whether they will experience the end of the story in two years.

In classic-TBC, one complain of advanced raid groups was that the system made it often hard to replace players. I think this was a concern of a minority. Recruitment-runs in older raid dungeons and advertisement of advanced raid groups in cities actually enriched the game and gave new players a feeling of 'everything is possible'.

I do think that classic as well as TBC should have had more introductory dungeons and raids.

When the new mechanic was released, the players were busy fighting shadows and didn't understand that WoW had neither become too hardcore nor too casual. It had become a one-size-fits-all that couldn't satisfy the broad spectrum of 12 million players.

Most ironically, the new system was in direct contradiction to the intend of the developers to cater to all players. Maybe the developers were fighting shadows, too?

Edit: Seanas just sent me a really good link. It's from Summer 2009.


  1. 1) The 4.1 T11 nerf isn't something new, that happened even back in vanilla. What's new is that vanilla didn't mention such changes in the patch notes.

    3) Badge loot was invented for 5 mans.

    The problem with 5 mans is that they have fewer bosses then a raid and the bosses have smaller loot tables then raid bosses. That makes it much to easy to obtain a specific item.

    Unfortunately, they have to have smaller loot tables. Otherwise you would never get your item with only one item per kill.

    The same problem also exists for 10 man raids and the problem was already visible in Kharazan (and reported and many blogs). I got 7 tank items for my tank paladin in one run. Something like that wasn't even imaginable back in the 40 man raid days.

    The loot system which worked great for 40 man raids didn't work anymore with 5 and 10 mans. Badges are one solution to fix it. The other would probably be to remove the armor class and make all stats appealing for all classes. (eg: Chest with 100 Sta and +47 Awesomeness).

    5) That could be fixed with story modes for dungeons and raids. Story modes would be very easy and could be completed very fast but you don't get loot in story mode. Maybe story mode would get unlocked with the next tier.

  2. 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.

    The more ice cream sales occur, the higher the incidence of drowning deaths. Clearly, ice cream causes drowning.

    I would go point by point about how ridiculously the rest of your argument is - points 4 & 6 are especially galling - but without the fallacious premise this entire debate falls apart.

  3. I would say Blizzard has eliminated long-term goals in the game. With nothing to look forward to besides more of the same, current players will get bored sooner. Without the mystique of possibly getting an item most players don't have, potential players are less attracted. No more grand quest chains for legendary weapons like Thunderfury, Atiesh, or Lohn'goron. No more extensive reputation grinds for real benefits like Hydraxian Waterlords, Thorium Brotherhood, or Ogri'la. Now it's just raid the current dungeon, hit the LFG tool for some extra emblems, and maybe some daily quests for your faction-specific mount/pet and extra gold. Sure there are some long-term achievements in the game now, but at most they give you a special mount.

  4. "pleasant social environment": is it some kind of new buzzword we should expect to see in all future posts? :)

    BTW I actually went to have a look at the data you link every time. At first sight, it does not look like two straight lines as you claim, but more like a growth curve (the slope at the beginning is higher than that around 2008 for example), flattening out at 2009, i.e. the time when some other "big-numbers MMO" (in the same theme as WoW) were released.

    I'd like to understand why you dismiss completely the fact that the MMO environment changed a lot from 2006 to 2010, with more and more titles being proposed (and thus eating the WoW market share).

  5. Ok, I had missed one graph: the "Total active subscriptions", which also flattens out just like WoW, seeming to indicate that the plateau is not WoW-dependent, but more about the global population.
    Comparing the graph, WoW flattens a bit earlier, more or less at the time when Aion (WoW-clonish) appears.

    So I stand by my interpretation: the growth curve has nothing to do with WoW game design and a lot with the population total and distribution across games.

  6. Helistar, please look at the datapoints for the years 2006, 2007, 2008 and 2009. It is 6 mio, 8 mio, 10 mio, 11.5 mio and 3 months into 2009 it is 12 mio.

    You are right that there are other factors than the ones I describe. Many internal to WoW, and many external. You are free to blame any of them to the sudden drop in growth rate, including old age or reaching some mystical market cap.

    My argumentation relies on the assumption that appealing to players who like just one challenge level in the game - no matter what this challenge level is - appeals to less players. If this doesn't convince you, then more power to you.

  7. Azuriel, in the future,

    1) please understand that statistical correlation may not be mathematical evidence, but it does increase the likelihood of a hypothesis.

    2) don't comment that something doesn't make sense without going into it.

  8. There's still the problem of China completely dropping out of the picture in that very year.

    You always reference global numbers but leave this out. Do you have numbers that only reference "the West"? (Maybe including Korea and Taiwan.)

  9. Thanks for the comment, Srosh.
    If you look at the graph at, you realize that it does not account for any problems in China at this time. It does do so for Summer 2009, however.

    The graph does have curves for east and west until 2010. But I really have no idea where these come from, so I usually don't refer to them.

  10. Ok, what I do see is that from start to end of WotLK, based on the informations provided by Blizzard themselve, global numbers rose from 11.5 to 12 million "subcriptions". That's definitely a lot less than in both Vanilla and TBC.

    However, WotLK was never released until a few months before the new peak.

    I share your doubts about the curves for west and east because they move exactly the same. Something I can't believe is that subcriptions in China remain stable while they're without any new content for months and then without a game at all (but maybe that's because of the way subscriptions are counted in China?). Besides the offline period that looks like an interpolation based on a west:east ratio.

    Something that is expecially strange is the supposedly global subcription numbers fell from ca. 12 million at the beginning of 2009 to 11.5 million one quarter later. But both the western and the eastern graph stay stable (even slightly rising) from the same date point to the next. How can that be?

    I think what my main point is that the whole period between the announcements of Blizzard before and at the end of WotLK is a black space. Basing an argument on the development of subcription numbers between these two points is basing it on a very fragile ground.

  11. *However, WotLK was never released in China until a few months before the new peak.

  12. Note that WoW started the year 2009 with 12 mio subscribers.
    Link. That was very soon after WotLK had been released.

    You can bet that the sub numbers drop between points with lots of space in between, by the way ;). WoW oszillates a bit, because there are many old players who occasionally come back.

  13. By the way, the easiest way to confirm the numbers on your own, is to google for terms like "warcraft subscribers millions" and then specify search ranges, like 1/1/2008 - 1/1/2009.

  14. "Note that WoW started the year 2009 with 12 mio subscribers."

    That's what I said. Still it doesn't answer why the global would drop to the next quarter when both eastern and western supposedly both slightly grew.

  15. Because those East/West numbers are guesses at best. I can verify each number of the global numbers. I cannot verify the East/West numbers.

    I repeated that WotLK started with 12mio, because you commented earlier:
    Ok, what I do see is that from start to end of WotLK, based on the informations provided by Blizzard themselve, global numbers rose from 11.5 to 12 million "subcriptions".

    In fact, numbers didn't rise during WotLK at all, or, only during the first 2 month.

  16. That's not entirely right.

    According to Blizzard's own press releases, WoW hit 11.5 million subcribers in December 2008 a month after the WotLK release:

    And, two years later at the end of WotLK, in October 2010, 12 million subcribers:

    Is that growth comparable to the one during Vanilla or TBC? No definitely not.
    Is it still a growth? The numbers say yes, even despite the problems in China.

    You don't help your arguments (which I can relate to, even if I don't agree on all of them) if you base them on numbers that are not so clear cut as you make them out to be.

  17. Scrosh - didn't you just agree that WoW started the year 2009 (that's Jaunary 2009, 2 months after WotLK release) with 12 mio?

    This number is the same number they had at the end of WotLK, October 2010.

    So, as I said before, they did grow during the first two months of WotLK. These were the months players explored Naxxramas and experienced the new raid character progression mechanic.

    During the rest of WotLK WoW didn't grow; at least not sustainably.

  18. If raid difficulty was correlated to subscription numbers, then why didn't subscription numbers go up again when Cataclysm completely reversed WotLK?

    You really should do your research and read Raph Koster's brilliant piece on how all subscription curves for all MMORPGs have exactly the same shape.

  19. Thanks for the comment, Tobold.

    I never wrote that sub numbers are correlated with raid difficulty. And it troubles me a bit that you might have gotten that impression.

    I wrote that a good game needs to appeal to as many players as possible and catering to just a few of them - no matter whether the 'hardcore' or the 'casuals'- will make it bleed sub numbers.
    (Read the next post if you think that appealing to all players is identical with catering the lowest common denominator..)

    I read R.K. piece. It's empirically correct. But that's really not all that hard. Of course, sub number rise for some time and fall for some time. To find some coefficients to make the curves fit an exponential isn't all that hard.

    And I disagree with the notion that you can do nothing about these curves at all. As does Blizzard disagree: otherwise they hadn't revamped the entire old world. You know, if sub numbers behaved according to R.K. anyway ...

  20. I wrote that a good game needs to appeal to as many players as possible and catering to just a few of them - no matter whether the 'hardcore' or the 'casuals'- will make it bleed sub numbers.

    I think WoW's age may be it hurting here. Players aren't exactly burned out, but their comfort zones have shrunk. They have a better idea of what kind of gameplay will entertain them, and what kind won't.

    Whereas previously, the zones were broad and overlapping, now they're narrower and separate. It's harder to make a game that, as you say, appeals to a broad range of players.

    What will be interesting to see is if this narrowing of interest carriers over to other MMOs. The broad appeal WoW had in early days may only be a product of a naive playerbase. That may be impossible to reproduce except in new markets.