[MUD-Dev] Psychology & Player Motivation (was Geometric Content Generation)

Sasha Hart Sasha.Hart at directory.reed.edu
Sun Oct 14 02:56:38 New Zealand Daylight Time 2001


>> Lack of punishment isn't what makes things boring. But the
>> possibility of failure may be the spice which makes things like
>> leveling rewarding *to begin with*.

> Maybe, although I think social comparison and the lack of other
> valid activities are more important reasons.

I'm not sure we're fundamentally talking about different things. All
of the "punishments" spoken of certainly are not operating on the
exact same basis as starvation or electric shock.

They might *act* the same in some sense, e.g. if they follow "a
behavior" (whatever that is, right) the probability of "that
behavior" occurring again decreases - while being entirely different
in another sense (physiological processes involved, what person
would say about the whole thing, what other behaviors are affected,
etc etc.)  Likewise "reward" of a certain kind (getting a higher
score in pac-man) might actually not *be* reward by such definitions
without other factors (say, anyone to see or care that you got the
high score in pac-man.) There is no reason to exclude all rewards
but, say, food and levels from consideration as rewards when there's
good reason to think that other things serve as motivators and
pleasant things (just what the real meat of the term "reward"
is). There is no reason to be dogmatic that a certain game construct
like the level will always be motivating or pleasant. And so on.

But it's pretty trivial to say that the terms are vague enough.  The
terms aren't that useful by themselves, are not the point.
Arguably, the terms add nothing _by design_: understanding of how it
works ought to be filled in by going out and seeing what predicts
what it does.

I usually develop understandings about things I see naturally.
Sometimes, I want better support for what my common sense predicts,
or a check that my common sense is not wrong.  Other times I do not
naturally develop an understanding and may have no direction - in
which case the questions others ask and the things others see seem
to supplement my own common sense. That's the use I get out of
academic work.

My tastes extend to laboratory experiments, among other things - not
because lab experiments make good game designs, but because taken in
groups they make for models which give an understanding that gives
starting points for design which are more exploratory and more
productive than what I would do otherwise (produce an amalgam of all
the designs I ever enjoyed.)

For me, it is useful not only for giving me basis for intuitions and
data to supply new intuitions, but also to cover blind spots which I
don't think of.

> I think that being able to predict behavior to some extent, rather
> than clamping down on the designs one can think of, gives one a
> structure to improvise from -

What if I wanted to build a beautiful and unusual suspension bridge
without any engineering understanding? I'd have problems unless I
was some kind of natural genius - this sort of thing can help us
mortals, not only in the quality of what we produce but in its

But it's not necessarily necessary, especially when the field is
broadened to, e.g. "a beautiful bridge" rather than "a beautiful
suspension bridge," or when "engineering understanding" is extended
to include experience gained trying to build bridges with one's
hands, etc..

>This is just plain rational behaviour, isn't it? 

Just plain rational behavior would require complete information.
As you argue, we never had that - and so what we do bears
interesting relations to what information we have. This 
accounts for many differences between "the rational model"
and how people turn out to actually behave.

With some other differences  it's not so apparent that this is the
reason - but I agree that there are probably reasons behind
these "failures of rationality" that explain why they take place.

(e.g. the matching law does not work in the way
we conceive of as optimally, probably because it is tuned for
robustness and generality rather than simple optimality with
respect to some narrow "rational" criterion.)

> Most people are not going to work their asses off to have a great
> funeral (ok some might).

Oh, my - this constitutes one of my most frequent moves on roleplay
MUSHes. Dying in games is liberating, but maybe only for someone who
likes subverting the intent.

> Still, there is clearly a difference between having to kill 10
> orcs 10 times in a row, and first killing 10 orcs, then entering a
> cave, then picking a lock etc... The last version potentially
> offers a drama that progress, a story, complex rituals, rising
> expectation and awareness by dramaturgical means etc.

Of course these are different, and of course, the fact that there
are other relevant differences does not remove the similarity (10
orcs.) The problem would be with assuming that only the number of
orcs killed mattered for any reason.

I think there are probably many more important differences between
them even than what you mention. For whatever reason, the latter
does sound more fun.

> Games are simple, but not that simple. Especially the multi-user
> _worlds_.

Games are an academic gold mine, and not just silly "academic"
games, but these games we are discussing.

I wouldn't be surprised if there was a lot going on in this list
that outstripped current academic understanding - we are, after all,
dealing with tons of revealing data, however undisciplined the

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list