Friday, November 16, 2007

A Production Prespective

by Harvey Smith
Creative Director - Midway Austin

My experience with Q/A over the years has been interesting. I started out as a tester in 1993, knowing nothing about software development. I was an enthusiastic video game player, a writer, and open-and-paper RPG fanatic. This list of commandments would have instantly accelerated my learning by a year.

Now, as a creative director, I crave more polish in games relative to everything else, which usually stems from good development and tech practices. Best practices impact gameplay much more than people often acknowledge. Even in cases where development teams are trying something new and ambitious, a fast frame rate, a solid interface, stability and lots of feedback are the real key to enhancing the player's experience.

I especially like number five from this list, about inspiring testers to play creatively. There's always one person on the test team who thinks like that — someone who covers the basic route that 90 percent of mainstream players will take, then tries five or six creative alternatives. After a while, you develop this as an instinct. "Hmm, this front door triggers a scripted door-opening sequence... I wonder what happens if I break the side window and skip the door."


Q/A people know that most games can be broken (or will at least expose something that looks silly) if you try hard enough. Citing alt-path problems, then speculating on their likelihood is very useful; sometimes we take things like that and run with them creatively. We might say "If five percent of testers think to try this, what if we bullet-proof it, then make it more attractive or obvious, so that it occurs to maybe 20 percent of players." Working on the DEUS EX games, which were incredibly free-form, this happened a lot. Even working on an FPS like BlackSite, over the last year we've seen lots of cases where testers approach a combat scenario in a way we didn't expect—routes which might be under-supported and cause frustration. In all cases, having great dialogue with Q/A will make the game better.


The very concept of "testing" has evolved a lot to include a bunch of best practices for the genre, such as blind usability tests and post production (as a serious phase of development). The game never gets good as fast as it does in the final months, so it pays to invest in post production testing. Ideally you'll finish the game from beginning to end as soon as you can, then spend a ton of time pounding on it, observing new players stumble through it, and looking for opportunities to pay off dramatic moments. Bringing in round after round of people who have never seen BLACKSITE has taught us invaluable lessons about where players get stuck, run out of ammo, get lost, and have the most or least fun. I wish we had three to six more months of that sort of thing — it would make a tremendous difference in the final quality of the game. I think the difference between great and mediocre publishers is the wisdom to invest in the final period of testing and tuning; the discipline to avoid cheating into this phase, even if a game is late.

No comments: