[MUD-Dev] The Best Guy on the Mud Thing

[Travis S. Casey]

This is mostly various random thoughts on the topic... haven't had time to
respond individually to interesting messages, unfortunately.  I chose
JCL's post to reply to because he's made some points I'd originally wanted
to make.

On Wed, 15 Sep 1999, J C Lawrence wrote:
> Dundee  <SkeptAck at antisocial.com> wrote:

> > Even over a given time period, it would mean that if the average
> > person uses the skill 10 times a day, then anyone who uses it less
> > than 10 times a day is effectively getting worse at it - since
> > whether you are "good" or "average" or "terrible" at a skill would
> > just depend on where everyone else was.

That's not strictly true, since it's a comparison of ratios rather than
absolute uses.  For example, if the average character does 100 skill uses
a day, and uses skill X 10 times a day, its ratio for skill X will be
1/10.  If Fred uses skill X 9 times a day and does 100 skill uses per day,
his ratio will be 9/100.  

Now, while Fred's absolute number of uses of skill X will get farther and
farther behind the average over time, the ratio of his ratio to the
"average" ratio isn't affected by that -- it remains constant.  Thus, Joe
wouldn't get worse at the skill -- he'd simply stay below average.

> There's a base player perception principle underlieing this:
> 
>   Players expect, even demand, that the game will always reward them
> for their time investments, and doesn't remove those awards once
> gained.

The majority of players, yes -- probably even the vast majority of
players.  However, this isn't true of *all* players.  In the paper RPG
world, there are players who willingly choose to play systems in which
skill decay exists, and even those who play systems where there is *no*
possibility of improving character skills.

> > You'll just be better at the skills you use more often, and worse
> > at the ones you don't.
> 
> Predictability comes into play here.
> 
> I play for a week or so and get to the point that I know that
> roughly one fifth of my fireball spells will succeed.  On that basis 
> I also know that I can go whup GooGoo without too much bother.  
> 
> A few days later others have been beating up on their Fireball
> skill, unbeknownst to me, and the ratios, and my position within the 
> ratios, have changed.  Now my fireball spell succeeds one tenth of
> the time.  GooGoo eats my arse.

This is a possibility; however, I'd expect it to become less likely over
time.  Consider:

Mud X has just opened.  To date, there have been 20,000 skill uses, 50 of
which have been fireball.  Thus, fireball's ratio is 1/400.  People start
using fireball more heavily, and an hour or two later, its been used 150
times, while total skill uses have only climbed to 25,000.  Fireball's
ratio is now 3/500 -- more than twice what it was before.  People who
haven't stepped up their use of fireball have effectively lost skill.

It's a year later.  To date, there have been 4,000,000 skill uses; of
these, 24,000 have been of fireball.  Someone starts ramping up their use
of fireball, using it 200 times in ten minutes.  There have now been
4,000,400 skill uses, of which 24,200 have been fireball.  Fireball's
ratio changes from 0.006 to 0.00605.  

Note that in the second case, the actual increase in rate of fireball use
was higher than in the first, but the effect was smaller because of the
greater "inertia" of the well-established ratios.


Now, this is still a loss of predictability -- my point is simply that as
the system gets older, the potential for sudden variations in established
skills becomes smaller.  Sudden variations are most likely to happen with
new skills or skills that have been used only very rarely.  This problem
can be mollified by having broader skills -- i.e., instead of having a
skill in "fireball", have a skill in "fire magic" which is used for all
fire spells (possibly with modifiers for different spells).  This skill
will be used more often than any individual fire spell would be, and thus
would be more insulated from sudden ratio changes.

It might be possible to make the gathering of inertia faster by initially
having skill uses count for more -- e.g., during the beta-test period,
count each skill use as 10 uses.  Once the mud opens, count each skill use
as just 1 use.  This would make the skill ratios become more stable than
they were during the beta test period.  The admins might also want to
"initialize" the ratios with guesses -- e.g., take 10,000 skill uses,
decide how they think they'd likely be distributed, and set starting
totals for each skill according to that.  This would give an initial
"mass" of skill uses to give some inertia to the system and some basis for
calculating values, but allow the mud to self-adjust over time.

Probably the best option would be to combine these -- initialize with
reasonale guesses, run an alpha test with accelerated variation, crank the
variation down, run an open beta with normal variation, and then keep
separate track of the skill uses in the last month or two of the open
beta, and use those ratios to reinitialize the database with a large
initial mass for the real opening.  That would give you something that
ought to reasonably approximate what the real ratios should look like, and
you could choose an arbitrarily large initial mass -- maybe 5 or 10 times
the actual mass of that month or two.


A bigger problem that I see is how this would affect new characters, if
implemented directly.  Consider a new player who gets on the mud and
creates a mage.  The first skill her character uses is "fire magic".  Now,
this character has a fire magic ratio of 100%, and thus, by the strict
definition, has just become the best possible fire mage -- until he/she
uses some other skill.  If she then uses "sword" skill, she now has ratios
of 0.50 with both of those skills, and 0 with all other skills.  

There's also a question of how initial skill would be determined -- using
the system strictly, someone who's never used any skills has a ratio of
0/0 with all skills -- which is undefined mathematically.

IMHO, a good simple solution is to simply give each starting character a
starting divisor -- e.g., add 100 to their actual number of total skill
uses.  Thus, a character who has never used any skills has a ratio of
0/100 in all skills.  A character who has only done one skill use has a
ratio of 1/100 in that skill.  

This would serve to do two things:  first, it damps the initial wild skill
swings down.  Second, it makes characters improve over time, as their
number of actual total skill uses rises.  (Without the added "filler"
total uses, a character's ratios all *always* add up to 1.  With it added,
a new character's ratios will add up to something small -- starting at 0,
going up next to .01, then to .02, and so on.  This would be an asymptotic
increase -- the total of ratios would never reach 1, but could come
arbitrarily close to it.)

(When doing this, I wouldn't include those "phantom" 100 uses for each
character in the mudwide statistics -- doing so could create problems,
since a sudden rash of people creating characters and then abandoning the
mud could impact the mudwide ratios significantly.)

--
       |\      _,,,---,,_        Travis S. Casey  <efindel at io.com>
 ZZzz  /,`.-'`'    -.  ;-;;,_   No one agrees with me.  Not even me.
      |,4-  ) )-,_..;\ (  `'-'
     '---''(_/--'  `-'\_) 




_______________________________________________
MUD-Dev maillist  -  MUD-Dev at kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev