Re: PBM design and formulae - another fresh start

From: markiel@callisto.pas.rochester.edu (Andrew Markiel)

Date: Thu, 08 Jul 1993 11:09:00 +0000

In article <21e276$4ka@taloa.unice.fr> gilette@obsidian.unice.fr (Gilette Serge) writes:
>In article <1993Jul7.063106.15539@galileo.cc.rochester.edu>, markiel@callisto.pas.rochester.edu (Andrew Markiel) writes:
>
>I can tell you that fuel comsumption depends on mass, tech, distance,
>oldness of the ship. You'll figure out that the older the ship the
>more he's to consume and so on, but you`ll never get the exact formula.
>
>>(my comment deleted)
>
>Sorry but with my definition of fuel consumption, you have no idea of
>the maths underneath. And i bet it will take you lot's of game to figure
>out exactly what it is, if you ever manage to.
>
Well, suppose we were in real life, and I wanted to learn about the
fuel cost of ships in relation to certain parameters. What would I do?
I'd start by taking a ship, moving it around, and varying a single
parameter at a time while recording how the fuel cost varies as I
change that parameter. (This is, in fact, the scientific method for
determining formulas, which obviously works well since so many
scientists seem to use it). Thus, I would produce a formula that tells
me what I know about how the fuel cost depends on different parameters.

If you claim that I cannot produce such a formula via careful
experimentation, then the fuel cost must depend on these parameters
in an awfully complicated way. But then, how am I going to be able
to figure out enough information to use in the game? Basically, I
have to be able to make a reasonable estimate of the fuel cost in
any situation, or I'm just blundering in the dark wondering how
the game system works. To make a reasonable estimate, I have to have
a good idea of how the value I'm interested in depends on the
relevant parameters. If I have a good idea of this dependence, I
necessarily have a good approximation to the formula, since the
only purpose of the formula is to give exactly the dependence that
I need to understand.

To those who argue that this learning process is realistic: is it
really fun to go around doing tests to learn how the fuel costs
work? I don't think so. I'd rather take the approach that some
technician has already done this for me, and given me the results.
The "formulas' in the rules are just this knowledge that is provided
to me. Thus I think the rules should include fairly accurate
formulas, unless you specifically intend for the players to learn
how things work as they play (and recognize that many people
don't like this idea).

Now, there's the question of knowing the exact value versus having
a good approximate value. My simple answer is that it's a very strange
game if an exactly reproducible answer exists. If this is true, then
anyone who has seen a given ship before will have perfect knowledge
about it's fuel costs, and thus have a big advantage over people
who haven't. Thus, if you really like the idea of having uncertainty
in a value, then the proper thing to do is explicitely put this
into your equations (via either a random factor, or hidden information),
so that knowing the "exact" equations does not allow someone to
compute an exact value. It's a bad idea to try to conceal exact
values by hiding the formulas: "security through obscurity" is no
security at all.

Thus, my claim is that in most cases, the best thing to do is to
reveal the formulas, and if you want there to be some uncertainty
in the value, then use a random factor or hidden variable to prevent
people from doing perfect calculations. If you do things right, then
people who don't like math will be able to play by intuition, without
math people having perfect knowledge to beat them with. The only hidden
formulas scenario I'll agree to is where you make the formulas really
complicated, depending on many parameters, but still intuitive, so that
people generally know how things work but can't use past knowledge to
perfectly predict the future. My only comment on this is, do it first
and then you can hide your formulas if you want to.

>Serge Gilette   | email  : gilette@obsidian.unice.fr

-Andrew
markiel@callisto.pas.rochester.edu

References

Re: PBM design and formulae - a fresh start
gym@dcs.ed.ac.uk (Graham Matthews)
Tue, 06 Jul 1993 11:55:13 +0000
Re: PBM design and formulae - another fresh start
gym@dcs.ed.ac.uk (Graham Matthews)
Tue, 06 Jul 1993 16:23:42 +0000
Re: PBM design and formulae - another fresh start
markiel@callisto.pas.rochester.edu (Andrew Markiel)
Wed, 07 Jul 1993 06:31:06 +0000
Re: PBM design and formulae - another fresh start
gilette@obsidian.unice.fr (Gilette Serge)
Wed, 07 Jul 1993 08:41:10 +0000