[MUD-Dev] Re: Levels versus Skills

[J C Lawrence]

On Fri, 15 Jan 1999 16:14:48 -0800 
Caliban Tiresias Darklock<caliban at darklock.com> wrote:

> My game *resets*. You specifically set up your configuration, and
> nothing happens until you apply it. Each zone can be reset
> individually, or the entire game can be. When a specific zone is
> reset, all players in that zone get shoved into another
> one. Anything they own in that zone is lost.

> But... now there's virgin territory. Zones can get "played out"
> over time, if players spend a lot of time there; enemy
> regeneration is limited by existing power level, so if the enemy
> is sent packing often enough he will eventually dwindle into
> obscurity. Likewise, if players *ignore* the enemy, he will
> eventually snowball into an unstoppable force that will steamroll
> through the universe like a vast engine of doom, leaving nothing
> survive in his wake. Both situations are Bad and require a
> reset. This is a legacy problem, and one which I am still deciding
> whether to maintain. There are good points.

First thoughts:

  Change is good.  Extremes are bad.  The essential problem you have
is that the system, via positive feedback, will tend to extremes and
will then tend to stay there.  *THAT* is a problem.  

You need a system that will accellerate toward to the extremes, but
will, upon approaching them, create a repuslive force that will
accellerate the pendulumn back.

Aaack.  Even that won't work.  That just gives you wild and
accellerating bounces between the extremes.  Not good.  You need a
force graph that looks something like:


                             Game impetus to move to signed state
                           /
    . -ve extreme         |          +ve extreme .
    .                     |                      .
    .x                    |                      .
    . x                   |                      .
    .  x                  |                      .
    .   x                 |                      .
    .    x                |x                     .
    .     x               |  x                   .
    .      x              |    x                 .
    .       x             |      x               .
    .        x            |        x             .
    .         x           |          x           .
    .         x           |          x           .
    .         x           |          x           .
  --.---------------------+----------------------.--
    .          x          |           x          .  \
    .          x          |           x          .    Game state.
    .          x          |           x          .
    .            x        |            x         .
    .               x     |             x        .
    .                 x   |              x       .
    .                    x|               x      .
    .                     |                x     .
    .                     |                 x    .
    .                     |                  x   .
    .                     |                   x  .
    .                     |                    x .
    .                     |                     x.


Translation: The x's are the graphs.  The X-axis shows how how far
the game stat is along the NPC-in-control (-ve) or
players-in-control (+ve) scale.  The Y axis shows the in-game
force/tendency to move the game-state in the direction of the sign
of the Y-value.  ie A large positive Y-value will be a big force to
make the game more player controlled and a -ve Y-value will force
the game back towards NPC controlled.

The points to notice are that there are two inflection points, both
roughly in the middle of -ve and +ve.  These are the two states that
the game will tend toward: either medium-level player mastery, or
medium level NPC's in control.  Both are fairly enjoyable states
(you either have a respectably dangerous enemy to fight who can kick
you arse or you are mostly in control, but the enemy is still plenty
dangerous).

Deviation from the inflection points to the extremes responds with a
force massive countersigned that will rapidly (hope) force the game
towards the other inflection point.

Deviation from the inflection point towards zero state will respond
with a force that will tend to throw the game across the inflection
point and then back towards zero-state.  The hope is that this
mechanism will create a pattern where the state tends to vibrate,
with increasing amplitude, about an inflection point, until it
suddenly breaks out and whooshes across the zero-state towards the
other inflection point where it will then tend to settle to make the
same sort of ascending amplitude pattern before wooshing back again
etc.  

Occassionally (there is where care in the graph shape comes in (you
might need to make the graph shape different depending on the
direction of motion unless you can build entropic losses to make
sure that crossing the zero-state doesn't have any effect on the
state-change momentum)) the game state will ping-pong between the
big ramps at the extremes.  The hope however is that the centrat
ramps (near zero-state) will damp this effect such that end-to-end
vibrations will rapidly damp into vibrations about an inflection
point.

--
J C Lawrence                              Internet: claw at kanga.nu
(Contractor)                             Internet: coder at kanga.nu
---------(*)                    Internet: claw at under.engr.sgi.com
...Honorary Member of Clan McFud -- Teamer's Avenging Monolith...