Game Development Reference
In-Depth Information
So I replace myself with other playtesters. Ideally, I'd get real play-
ers drawn randomly from a sample of the public representing the target
market of the game. But even if this isn't possible, there are other alterna-
tives. I usually use coworkers. I pull programmers, testers, artists, and
audio engineers who haven't seen the fight, sit them down at my com-
puter, and watch them play. I tell them nothing, stand far behind them,
out of their line of sight, and wait for the design to fail.
It always does. Some players break the fight by inventing strategies I
never thought of. They'll refuse to advance and snipe every enemy from a
distance. Or they'll charge past enemy lines without firing a shot. Others
fail in frustration because they don't know the fight the way I do, or they
don't notice a key element. They'll miss a hole in the floor, and fall through
to their death. They'll get shot in the back by the guy I sent to flank them.
They'll step on the blinking land mine I thought was obvious. To para-
phrase Bill Cosby, playtesters do the darndest things.
After one playtest, I've got a list of issues to solve. Some are simple
fixes (light that enemy better so that people can see him). Others are more
complex (restructure the left side flanking route so that both players and
enemies can use it). I get to work. Half a day later, the changes are in, and
I'm ready for the next playtest. I find someone who hasn't played the fight
yet, and watch him.
The loop goes around this way 10 to 20 more times. By the end, two
or three weeks have passed. The fight is well paced and well balanced, and
it handles players of many skill levels and play habits. I don't have to guess
how it will play out when it's handed to real players, because I already
know—the playtesters showed me.
But it still doesn't look like much—flat gray cubes suggesting a theme,
but not resembling a finished game. Now the artists really get involved.
Level artists do a first pass on the space, replacing gray shapes with real
art assets. We test again. Even if the mechanical shape of the fight doesn't
change, art changes affect how players perceive the fight, so we have to
watch how it affects playtests. As we see problems, we discuss them to find
solutions. Sometimes I might have to change a scripting detail, removing
or adding characters or tools. Other times, the artist might have to add
light to brighten a space, or simplify something to reduce art-driven noise.
The iteration loop is now several days long, since making art is slow work.
If we're lucky, the art causes no major issues. Since I playtested thor-
oughly using gray cubes, the base level should continue working as it did
Search Nedrilad ::




Custom Search