[MUD-Dev] Balancing a Mud
Martin C Sweitzer
msew+ at andrew.cmu.edu
Fri Mar 12 03:09:00 New Zealand Daylight Time 1999
When you have a mud you can have the best world, ideas, classes, skills,
EQ, etc etc but unless it is balanced the mud is garbage.
Now of course, (well let us hope to god) that you have designed your mud
to allow easy customization and tweaking of the variables you have. How
strong a class of sword is or how strong a specific artifact is. Spell
effects. Tables are nice for this.
If you don't have this then I would argue that your mud is not going to
scale and going to just bust at the seams when you begin to add lots of
things to it. (ie add that new Spanky Class and watch your carefully
balanced mud collapse under the refactoring you get to do now).
So how do you balance a mud?
I have N sets of V variables and then I instances of those variables to
balance which boggles my mind.
You have this wonderful:
Player vs Player
Player vs Mobs
PLayer vs Environment
PLayer vs Self
GroupsOfPlayers vs GroupsOfPlayers
GroupsOfPlayers vs Mobs
GroupsOfPlayers vs Environment
GroupsOfPlayers vs Self
(I don't believe that groups are the basic unit. If you insist on that
then how do you deal with a Group of One? It is the same thing as a
I can imagine trying to simulate the entire world, but then I stop and
think while that would be a totally kick ass project in itself, you will
end up spending more time coding the simulation and tweaking it than
working on making a place for players to enjoy. Way too complex also.
Oh just set up a beta test server. Bah now you have two muds in
reality. Seems sketchy to use a beta mud to test balancing issues and
tweaking. The thing here is that while a beta mud is really awesome for
allowing a wide range of testing and balance issues you will end up
making a small tweak and then having to run the beta mud for a while and
monitor before you place them in to the world. Seems sketchy to me at
SOME TYPE OF EVALUATION FUNCTION
Being able to to group the variables into groups of some sort. This
reduces the complexity issues but still is going to suck I think. I can
imagine something like having the set of variables that represent a
maxed out foo class, an average foo class, and the wussy foo class. And
then having those variables vs everything else in a spreadsheet.
wussyFoo vs lvl 1 mobs
wussyFoo vs lvl 2 mobs
wussyFoo vs lvl 3 mobs
wussyFoo vs lvl 4 mobs
And same for all of the the other vs players, self, environment. groups.
And then when you change something in the long sword hardness rating
everything is updated and you get a nice rating from 0-1 where .5 is
balancedd or something like that.
Problems with this is that the evaluation function for comparing the
groups of variables is going to be total hell to create. I think. I
think you might end up simulating the interaction between the groups
(fighting, exploring, working, etc etc).
These are more bottom up approaches. I think mob true balancing might
be is that you are able to set something like: a level 10 warrior class
should be able to get 5k exp per hour of RL. Should get 2k gp per hour.
And then this top down approach calculates out the variable values for you.
Does anyone have any advice or even better some proven techniques or
even a framwork that can help balance things en general?
MUD-Dev maillist - MUD-Dev at kanga.nu
More information about the MUD-Dev