When you should write a plotless game.
Carl de Marcken
> devoid of meaning. I know what I want the player to end up doing in the
> end.
Although I accept that this writer's premise is valid for many situations,
let me relate my experience in writing a game with an unusual purpose. The
experience may be relevent and important for many other adventure game
authors.
I, with help from Dave Baggett (of Adventions) and others, just finished a
game designed to be played by members of the MIT Artificial Intelligence
Laboratory (once home of Zork) as one of many events in a lab-wide
Olympics, with 4 teams of maybe 40 people each competing to solve the game
first. Of course, not all 40 were interested in playing the game.
Our goal was to make the game fun. People should enjoy playing it. But
there are a lot of things that set this situation appart from those that
most authors face. First of all, we knew a lot about our audience. They
are by and large well-educated computer scientists, and we could write many
"in" jokes, and include puzzles that relied on very specialized knowledge.
The game was also going to be played by a team (some of whom would be
experienced gamers), not an individual. We didn't know what effect that
would have, but we made the game very difficult. Eminently fair, but very
difficult. Experienced game players that tested got quickly overwhelmed
and never came near solving the game. Their evaluation was that the game
presented too many possibilities at once- they couldn't decide what to do
with 50 different weird objects in a world where there seemed to be 50
different strange puzzles. Unfortunately, we couldn't get any game testers
that were insiders, that could evaluate the game from the perspective of
the intended audience, but we did get computer scientists from elsewhere
that were experienced gamers.
Think about how things would change if you were writing for such an
audience instead of a "generic" individual. First of all, you want lots of
people to be able to play at once and cooperate. That means you can't
organize a game in a linear plot as so many of the current games are. We
have many, many puzzles presented to players at once. There are very few
"doors" to confine people's view, to focus their attention. Different
people could concentrate on different puzzles, or ideas. In fact, we
didn't even make an effort to tell the players what the puzzles were.
Secondly, in the arena of amusement, we had all sorts of ideas for objects
and puzzles that would be hilarious for people in the AI lab to encounter-
the game is a conglomeration of computer science, linguistics, physics and
mathematics. It would have been nearly impossible to organize these
puzzles and objects into a literary "plot", and unnecessary.
So instead, we picked the objects and the puzzles first, created an
alice-in-wonderland university setting that allowed us to put in anything
without it seeming out of place. Unlike most games, the locale is small.
No more than about 15 rooms. But the rooms are more heavily populated and
interconnected than probably any other game ever written. It doesn't seem
overly full, but when the players start realizing just how many connections
there are between different objects and how many different uses different
things can have, they can quickly generate thousands of good ideas to try
out. There are no rooms in the game there just to connect two others.
The game is 15,000 lines of TADS code for a small locale- implementing the
complex puzzles that come from thinking about AI and Physics and Math and
Linguistics resulted in objects and events that were much more interesting
than those most games have.
We put in a large number of red herrings. A very large number, though most
had some subtle hints that they were red herrings. Several times we
thought about removing them, in case they made the game to hard, but in the
end most stayed in. Some puzzles, when solved, presented information
suggesting that a certain implied task was indeed a red herring. The
herrings turned out to be a good idea. When a team seemed to be stumped,
they would make a list of what many puzzles seemed to need to be solved
(the list would include many herrings) and they would each pick items that
seemed promising to them and attack individualy. Again, this helped
maintain interest.
Many of the puzzles required a realization, such as the realization that a
horse with a bow on its neck and a tag saying "happy birthday" was a gift
horse, and that players should look in its mouth. This turned out to be a
good thing, because often when the "core" players were stuck they turned to
the remainder of their team, people who had not even been playing the game,
and found solutions from people with open minds. Or people who were not
playing the game saw the e-mail transactions and said "but didn't you just
say that ... isn't it obvious that you should ...". That helped keep the
game from being confined to a small few.
The result was gratifying. About 10 people on each team played it over an
80 hour period. It took them that long before they met the criterion for
winning, the earning of a degree from the university. The cooperation
between teammates (via e-mail mainly) was amazing- literally hundreds of
messages a day running around. It was quite obvious that a very difficult
game was reduced to one of a perfect level of difficulty because of the
cooperation. This cooperation would have rendered a game that only
presented a few puzzles at a time either trivial or uninteresting and
unable to hold people's attention. As it was, the players enjoyed the game
immensely.
Most of this experience is relevent to anyone that plans to make a game
that will be played by several people rather than one, or who plans to make
an "inside" game.
Carl de Marcken
BTW- if any CS/AI people in university environments would like to get the
game and use it for a similar contest, contact me. It has a restricted
audience, but not one confined to MIT, and we'd be happy to distribute the
source. The reaction here has only been positive.
Keywords: