Your games always seem to have this strong educational component. I was
wondering, how do you balance that with making the game entertaining?
I was never concerned with education until the game was fun. Any educational value a
program might have is totally wasted if people won’t play it. Probably the one game
which I learned that the most from wasSimEarth.SimEarthwas potentially the most
educational game I ever made, but yet it wasn’t fun. A surprising number of people
bought it; I’m still surprised by the sales figures. I think most of them played it for two
hours and then put it away. So I really think the fun has to come first. And the educa-
tional side, it’s not something that you tack on, it’s got to be fundamental to the design.
InThe Sims, it was all about learning to extrapolate design from behavior. That’s a fairly
deep lesson, it’s not just a fact that I’m going to teach you. It’s more like a way of looking
at things. If the entire design is true to that, it might be educational at some deep level
even though you might play the game for hours and not think of it as educational even
once. One of the main things thatSimCityteaches — it’s not explicit but it’s there — is
the shape of chaos. The fact that the best-laid plans can always go wrong, and that the
system is more complex than you think it is. Building a road to solve traffic doesn’t
always solve traffic; it frequently breeds traffic. Those types of lessons are hard to
explain in other media. But when you’ve experienced them through a process like
SimCity, you really get the lesson much deeper. It’s experience rather than exposition.
Do you ever have to compromise realism to make the game fun?
Oh, all the time. There’s also a frequent thing that we did in our games where we would
decide to match expectation and not reality. In fact, nuclear power plants don’t blow up.
They just don’t. But when everybody saw it, they said, “Oh, a nuclear power plant, can I
make it blow up?” It’s just what they thought of. So there are a lot of things we do just
because people expect them to happen that way for fun, even though it’s not realistic.
With the open-ended nature of your games, do you have to spend a lot of time
in playtesting them?
We do, but it’s invaluable time. You spend that time, or else you go spend months build-
ing the wrong thing and solving the wrong problems. We just had what we call
“kleenex” testing on one little component ofThe Simsmulti-player that we’re working
on. We have this one data display that’s convoluted and twisted. And the programmer
just got it implemented a few days ago, so we scheduled five people to come in today.
We call them kleenex playtesters because we use them once and then they never come
back, just because we want people who have never seen it before, with totally no pre-
conceptions about it. We don’t even tell them what it is, we just say, “Look at that, play
with it,” and have them describe to us what they’re seeing and what that represents. We
got some very consistent feedback from all five people today where we understood that
three of the variables we were communicating they all understood, the other three they
had no clue about. So for the last tester, we turned off the last three variables that
everybody was having trouble with and it was perfect. We do this at every stage of the
project now. It’s not just at the end when we have the whole thing working, we do this
with little components, even the art prototypes. And this was a lesson that was really
driven home to me by the late Dani Berry. She’s the one who didM.U.L.E.and all those
446 Chapter 22: Interview: Will Wright