Re: PBM design and formulae - a summary

From: bruno@cerberus.csd.uwm.edu (Bruno Wolff III)

Date: Thu, 01 Jul 1993 22:53:36 +0000

From article <Jul.1.16.51.55.1993.15294@gnumath.rutgers.edu>, by bonar@gnumath.rutgers.edu (Doug Bonar):
] In article <20v7ds$755@harder.ccr-p.ida.org> desj@ccr-p.ida.org (David desJardins) writes:
]
]> Graham Matthews <gym@dcs.ed.ac.uk> writes:
]> > You are not going to be able to back engineer *exact* formulas from
]> > output if either the formulas have some randomness in them, or
]> > if the result of evaluating some formula effects the value of
]> > one of the parameters to that formula (eg: think about combat - the
]> > result of a combat depends on the experience of the combattants,
]> > but the experience of the combattants will change as a result of
]> > the combat).
]>
]> That's right.  However I am going to be able to apply a lot of
]> sophisticated mathematical and statistical analysis to the problem.  I
]> do mathematical and statistical research for a living; I know what I can
]> do in this kind of situation.  I can certainly learn a _lot_ more than
]> someone who doesn't either write any programs or apply any sort of
]> mathematics.
]
]
] 	I agree that you will probably get some relationships out of
] whatever statistical tools you apply to the data you get.  Over time,
] I'm willing to grant that you'll get a set of formulas that start to
] approximate the game's formulas.  Assuming that the game's
] documentation tells people what effects (at least generally) effect
] combat (for example) you will even have formulas that include the
] various factors in the game (as well as being just generally good at
] prediction).  I assume that even for small effects, with enough data
] you (or any statastician) would be able to detect and include the
] random effects into the formulas you generate.
]
] 	I wonder about the actual effect of all this on the game
] though.  It will take a large amount of data (if the game's formulas
] are complex).  So, your statistics will take some time to come into
] play.  During that same time, someone who plays the game has time to
] get a feel for how the world works (you will too as well, but your
] effort is going into analyzing the data statistically).  In a well
] designed game, I'm not sure that your quasi-formulas will give any
] real advantage over the other gamer's experiential knowledge.
] (Perhaps  that is even a measure of a well designed game.)
]
] 	In essense, I suppose what I'm saying is that I think that
] some people gain the ability to deal with the world by heavy analysis
] of what it send them.  Others buy 'rules of thumb' they gain by
] working inside the world.  The first type of people will analyze any
] game you give them, the second will experiment to see what works,
] think about it and act on what they have 'experienced'.  I agree that
] the second route is probably not as successful in the long run.  But
] the analysis method still has to have the data to analyze generated
] somehow.
]
] 	So, in playing a game w/o the formulas given, you will
] experiment with things and analyze the results statistically.  To get
] a good model (of the model the game uses), you will need to do a lot
] of different things, to get the data for the analysis.
] 	A non-analyst would play around, but wouldn't get treat the
] resulting data the same way.  He probably wouldn't treat the formulas
] (if given) in the same way either.  So, he is probably better off (at
] least in the short run) with the formulas hidden.

Yes in the short run, nobody knows the rules so nobody can play well. After
a while people who carefully analyse the game will have a much better idea
of the rules than those who don't. At this point, if skill is needed to win
the game the people who have analysed the game will a have an advantage
over the people who don't. This advantage will be bigger than if everybody
knew the formulae since the players who have done carefull analysis of the
game know the rules better than the other players they will be better able
to predict the results of their actions. If the game does not require any
skill to play (i.e. what you do has little or no affect on the outcome) than
what is the point of playing?
The main affect of hiding the rules is to penalize new players to the game
and to give incompetent players a chance against new competent players.
]
] Doug
] bonar@math.rutgers.edu

References

Re: PBM design and formulae - a summary
bonar@gnumath.rutgers.edu (Doug Bonar)
Thu, 01 Jul 1993 20:51:56 +0000