[MUD-Dev] Re: Levels versus Skills, who uses them and when.

[Koster]

> -----Original Message-----
> From: J C Lawrence [mailto:claw at under.engr.sgi.com]
> Sent: Wednesday, January 13, 1999 5:04 PM
> To: mud-dev at kanga.nu
> Subject: [MUD-Dev] Re: Levels versus Skills, who uses them and when. 
> 
> 
> On Wed, 13 Jan 1999 09:58:36 -0600 
> Koster, Raph<rkoster at origin.ea.com> wrote:
> 
> > Quite aside from whether it is, I have become convinced that the
> > secrets to a really long-lived, widely appealing online game *in
> > the gaming sense* are multiple paths to advancement, ease of
> > switching between paths of advancement, clear milestones in said
> > advancement, and ideally, no sense of running out of significant
> > milestones (eg the ladder does not feel finite). 
> 
> <nod>  <ponder>
> 
> I quibble with that slightly.  Clear, or at least apparent
> advancement paths to known goals are important.  However I also see
> significant value in sideline growth opoortunities that are
> "advancement" per se (the player/character can now do things he
> could not do before), but which don't necessarily relate to known
> longer term goals (eg "learn to whittle chess pieces from
> driftwood").

"Side ladders," sure. Though my experience has been that players
promptly ask that said ladders be full ladders of their own. :)

> The reason for this quibble is that I see these philips as the key
> to expressive fertility,

Hmm, wasn't it you with whom I traded emails about "baited breath"? I'll
have to let Phil know you said that. ;) (The pedant in me mentally
corrects to "fillips.")

>  and, by the simple means of watching your
> players use of them, of finding new areas/goals/activities to
> expand/build_on as you increase the depth of the game (eg notice
> that people are playing chess with the whittled pieces, so add a
> chess AI and other explicit game supports).

Definitely. We have followed this path often in UO's ongoing
development. Players advertise " exceptional bows crafted by Iolo" for
high prices? OK, let's make it so that exceptional bows automatically
get a maker's mark, and watch the player economy improve.

>   Without such side-items
> there is a reduced stock of raw bits for players to start building
> with to *show* you want they really want (cf the old rule about
> satisfying player wants destroying the game).  Yes, flexible game
> mechanics can do a lot of this (cf the monster farms for sparring in
> UO), but adding abilities-to-do seems the real key to providing the
> seeds of later outgrowths.

Yes; I didn't mean to denigrate the importance of plain old rich feature
set, in stating my premises. The richer the feature set, the more
players will make use of it and they longer they will be retained, no
question. The points were really about advancement models.

> Perfect if minor example: The fishermsn in one of the LPC libs.
> Reese has commented several times that they were a throw-away
> feature that got ad hocly dropped in, and yet significant numbers of
> players like playing fishermen and swapping stories about the
> virtual big one that got away.

Something we experienced in UO as well. However, consider how much more
longevity this could have had for more people, if at first you could
only catch minnows and tires, and could eventually get to be good enough
to catch Moby Dick. We took the same route as Reese, and simply
randomized the fish received. In retrospect, there was a prime chance
for a ladder there. Not a ladder that will necessarily appeal to tons of
people, and not a ladder that has a lot of depth to it, but a ladder
nonetheless that can serve as a replacement for the advancement model of
the guy who got bored fighting.

> I'm not certain on this, but I suspect that actively ensuring that
> there are no uber-characters or otherwise stereotypic characters
> (even if only as a presumed logical ideal to attempt to approach
> with a real character) is key to this. 

Unfortunately, that's a holy grail of balance that classless systems
rarely find. :( Class systems by definition are about striving for the
ubercharacter, of course.

>  As soon as you get
> characters being gradated variations of the same definition you've
> lost, as you then have a battle for nearest approximate to a single,
> and thus known, goal, and that instantly reduces the player-apparent
> complexity of the game down to a selection of which particular
> uber-target to go for next. 

In UO terms, the "tank-mage." A common problem to most classless
systems, however. If you'll take a leap of faith with me and visualize
classless systems as a rubber sheet... :) Time for a physics
experiment...

Usually there's a single intersection point at which the rubber sheet is
distorted (tankmage, for example). Characterballs rolled on the sheet
will tend to fall into that depression. Secondary depressions will also
be formed, but one of the depressions will probably be more powerful.
Even if you strengthen the secondary depressions, you're never going to
get a flat sheet. A chessmaster/fisherman is simply never going to be
regarded as being as valuable as a warrior/mage, any more than a
grandmaster buggy whip manufacturer is regarded as being as valuable as
a computer programmer today. Certain skills and skill sets are not in
common everyday demand, and thus are not as well-regarded.

>  If instead all characters are *really*
> unique (largely guaranteed significant functional difference from
> all others), then even if you have a limited set of presumed
> uber-target characters, the path that any given character has to
> take to get there will be different and *none* of them will arrive
> at the same or similar locations but will instead create wildly
> variant approximations of the uber-target.  I suspect that once you
> reach that point that the deviance from the uber-target will then be
> more interesting than the uber-target, thus actively working to
> devolve the perceived value of the uber-target.

Now this is interesting. The question arises, how large a rubber sheet
do you need to accomplish this? In other words, how many truly
interesting and useful "secondary" skills can you provide?

In addition, I'd note that on the face of it, it seems somewhat
incompatible with the atrophy-and-learn-something-new model that Marian
described in her post. Thoughts on merging the two?

> <<Does this make any sense?>>

It did to me, at any rate. :) The list can warn us, hopefully, if we get
to overly rarefied discussion. :)

BTW, one factor that is overlooked here is the "mule" phenomenon--it's
rampant on UO, perhaps because of this very design concept. I had never
heard the term before Meridian 59, but I am sure it must have existed
before. This is where players simply make alternate characters to make
use of the lesser-used skills. Your "main" character fits a
stereotypical ubercharacter ideal, but you make alternates for those
occasions when you want the secondary skills. In UO, that is generally
for the crafting skills.

> Without digging up the archives, I'm reminded of a post I wrote to
> Marian on controlling/influencing player actions where I drew the
> analogy that controlling player actions is much like attempting to
> control a torrential river rushing (and cutting a channel) across a
> soft plain (King Knut?).  In that model putting up dams is doomed to
> failure as the torrent will perceive them as targets and
> aggressively undermine them.  In the above character development
> model, the goal is to create an evenly distributed mesh of minor
> tributaries (one stream per player), which wander, cross, and
> fragment the plain.  The problem being to prevent larger stream
> formation.

I assert that some streams will always be better than others. You'll get
interesting watersheds with multiple major streams and great variance in
the small creeks. But (to push the analogy into absurdity) some fishin'
holes are always going to be better than others. Some images of
ubercharacters will always serve as greater attractors.

-Raph