[MUD-Dev] Reputation systems: a possible path for investigation

[J C Lawrence]

Reputation systems are usually thought of as interesting ways for the
game system to define and provide predetermined and service-qualified
hinting and decision making keys to players about other players (PK,
trade, trust etc).  To date this has been an uncomfortable and halting
process which is behind the needs curve and chasing the definition of
both the actual requirements and possible implementation models which
don't carry excessive collateral costs.  Additionally, little has been
done to extend these player-derived reputation systems to represent
player views of in-game artifacts and systems.

Part of the problem is that the hinting data requested is contextual and
subjective, both at the time of the sequence points that form reputation
atoms and at the time the reputation is reported and interpreted by a
particular player at a particular place and time for a particular
purpose.  More explicitly the needs definition at both ends is not
uniform across an entire game or for all players of that game.

  Sometimes the problem for a reputation system is as trivial as, "Who
  is online right now that I'd like to play with?"  Other times its as
  subtle as, "Can I trust him to proxy my in-game interests when I'm not
  here?" say for minding a co-owned store, a merchant cooperation, money
  trading, or even simple employee/employer or client/patron
  arrangements.  Then there are the simple questions like, "Is this
  rotting rope bridge safe to cross?"

Arbitrary graining and subjective contextual per-player interpretation
seem to be called for, but are not an attractive implementation or
management prospect, especially for a casual gamer audience, and more
especially for a centrally defined, managed, and reviewed set of
metrics.  A possible approach instead is to treat the players as a
homogeneous mesh and allow them to define the nouns and verbs in the
reputation system, and possibly some of the high level grammars.

Guilds and other player group formations have partially coopted this
reputation space, using membership and internal group management as a
way of establishing and maintaining reputation at a group level ("He's a
member of guild XXX.  They're all PKers!").  Problems of this approach
are a lack of transparency and high barriers to entry both in social
networking and learning curve before that reputation data can be
effectively used.  Additionally such group-derived structures select
against casual gamers who don't invest the time or effort required to
breach those entry barriers and yet need and want an effective
reputation system for themselves as represented to others, and for
others as represented to them.

  Of course if these needs and wants were fully satisfied by the base
  system it would remove one of the significant functions which drive
  group formation and interaction on our services.  Depending on the
  exact definition of your service and audience this could be a problem.
  In the general case however I expect that the other forces which drive
  group formations like guilds would still dominate.  However it would
  change some of the basic definition and functions of guilds with
  games.

Content creation costs are a significant expense, making community
authored content attractive -- if only the risks of "bad content" could
be managed (where bad can refer to legal branding, service, product
definition or other implications).  Reputation systems are a possible
approach to helping to automate the detection and handling of bad
content so that human supervision and vetting costs for instantiating
and official sanctification of community derived "good content" (if your
model calls for that) can be constrained.

First Monday (http://www.firstmonday.org/) has several papers which
intersect these areas.  In particular:

  Open content and value creation by Magnus Cedergren
  http://www.firstmonday.org/issues/current_issue/cedergren/index.html

  The Augmented Social Network: Building identity and trust into the
  next-generation Internet by Ken Jordan, Jan Hauser, and Steven Foster
  http://www.firstmonday.org/issues/current_issue/jordan/index.html

  TOOL: The Open Opinion Layer by Hassan Masum
  http://www.firstmonday.org/issues/issue7_7/masum/

Cat Okita presented a paper on a peer-to-peer authentication system at
DefCon called "Aura" which also looks interesting (he promises sample
source RSN):

  Aura -- A peer-to-peer reputation system
  http://www.geekness.net/tools/aura/aura.pdf

  Aura DefCon presentation:
  http://www.geekness.net/tools/aura/aura_presentation.pdf

These aren't a compleat set; just bits I've been reading and musing on.

--
J C Lawrence
---------(*)                Satan, oscillate my metallic sonatas.
claw at kanga.nu               He lived as a devil, eh?
http://www.kanga.nu/~claw/  Evil is a name of a foeman, as I live.
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev