Wednesday, September 12, 2007

Guild Wars: Test Related Error

Still not done picking over the bones of that GC speech. I could seriously make about half a dozen more posts on the points raised within. I am also, in case it's not readily apparent by now, completely and utterly insane. I'm not joking. I need help.


But before the people from Bellevue arrive to dress me in my old familiar straight jacket, let me just flag this passage:


An MMO must deliver content at three distinct stages: the early game, which is the first twenty hours, the mid game, which is the first few hundred hours, and the late game, which is at a thousand hours and beyond. Each of these stages represents a chance for your game to continue to grow, or to decline and ultimately fail. The traditional QA model is just not equipped to deal with this – there is simply no way to effectively test 1,000 hours of content in the final months of the project, particularly when you are not focused only on bugs, but instead on that hard-to-define feeling of satisfaction that a good MMO provides hour after hour. (Emphasis added.)


This, by the way, is something I've heard before. And that I'd personally agree with. There's just no way to effectively test the kind of multiplayer online game we're talking about here - whether it's Guild Wars or Tabula Rasa or something more traditional like the almighty WoW - before release. No way to predict just what the teaming hordes you hope to attract will do once they get their filthy hands all over the pristine beauty a development team and their testers have been laboring over for so long. It's simply a numbers game. The more people you have, the more likely they are to find that flaw, that bug, that exploit, that you could never predict. And there's no way to adequately recreate that sense of community that exists only tangentially to the gameplay which is what really keeps people playing these games.


It's a failing, I think, not of the willpower of the game designers, not a problem with the people who make the game not being familiar enough with it, but a problem in current testing models. I'm minded of an article I read a while back about the developers of Halo and how they've created a vastly improved testing bed for their games. I'm not sure that what works for the developers of a console FPS also works for something like an MMO. They're two different beasts, after all. But the lesson to be drawn from that piece, I think, is that it's possible to get much better results if you're willing to invest and innovate not just in your game's code but in how you get feedback about it. Somewhere, somehow, there has to be a better way to test these online games. Ways that will make them better for everyone involved.

No comments: