Document Sample
Collusion Powered By Docstoc
					Jaren Peterson Presents

   Can We Prevent Collusion in Multiplayer
              Online Games?

      Jouni Smed, Timo Knuutila, Harri Hakonen
Key Points

    What is collusion?
        Is it different from cooperation?
    Collusion Detection is a Hard Problem
        Its simpler with a smaller state space
    Types and Roles in Collusion
        Difference between players and participants
What is Collusion?

    What is collusion?

    How is it different from cooperation?

    Examples of Collusion in gaming not
     specifically mentioned in the paper any of
     the class may or may not have participated
Difficulty of Detection

    Why is detecting collusion such a difficult

    What legitimate behaviors may give false
     positives as collusive behaviors?

    What ways does the paper suggest for
     detecting collusion?
Difficulty of Detection

    Legitimate Behaviors
        Learning Curve
        Familiar Company
        Sensible Play
        Conflicting Interest
Difficulty of Detection

    Ideas for collusion detection
        Data Mining
        Player voting / reporting mechanisms
             I liked the idea at the end of the paper of a
              centralized player profile service across games.
Types and Roles

    Levels of Agreement
        What are they?

        Are they all really collusion?
Types and Roles

    Levels of Agreement
        Express Collusion -planned, purposeful
        Tacit Collusion – understood, unspoken
        Semicollusion – fair weather cheaters
Types and Roles

    Content of Agreement
        Concealed Stance
        Knowledge Sharing
        Information Sharing
        Resource Sharing
Types and Roles

    What is the difference between a player
     and a participant in this paper?

    Why does it matter and how does this affect
     collusion detection and prevention?

    Any exciting personal experiences of these
     role differences?
More on Roles

    Multiple users sharing a single account
        What a great idea! (kidding of course)

        How can this be stopped/detected?

        What about bots (like in Curtis' paper?)‫‏‬

    Single User, many accounts

    Spectator collusion

    Assistive Collusion
        sting operations?!!

    Association Collusion (finally some meat)‫‏‬
        Ingenious ideas, anyone remember?

    Ingenius ideas for association collusion
        Make world and item names unique
        Hide player identity from other players
        Player rating/reporting

    Also mentioned (I found less interesting)‫‏‬
        Multigame Collusion
        Insider collusion
My Top 5 Points List

    Detection is mostly done by data mining
    Surveillance can be managed by feeding old
     data to surveyors
    Unique item and world names
    Anonymous players (no names or skins)‫‏‬
    Rating and Reporting Service

Shared By: