August 2003
- A potential best-seller? Richard A. Bartle
- Identifying Players Scion Altera
- Identifying Players Crosbie Fitch
- Metrics for assessing game design David Kennerly
- ADMIN: Crunch thread J C Lawrence
- Mapping real money into MUD money Alex Chacha
- Mapping real money into MUD money Katie Lukas
- Mapping real money into MUD money David Kennerly
- Mapping real money into MUD money Kent Peterson
- Mapping real money into MUD money Peter Tyson
- Mapping real money into MUD money Matt Mihaly
- Mapping real money into MUD money Paul Canniff
- Research in the Gaming Industry Damion Schubert
- Research in the Gaming Industry Kerry Fraser-Robinson
- Research in the Gaming Industry Richard A. Bartle
- Research in the Gaming Industry Matthew S. Ayres
- Mapping real money into MUD-Money Henrik Johansson
- Java or LPC (DGD)? Ben Chambers
- Java or LPC (DGD)? Ammon Lauritzen
- Java or LPC (DGD)? T. Alexander Popiel
- Java or LPC (DGD)? ceo
- Java or LPC (DGD)? Lars Duening
- Java or LPC (DGD)? Torgny Bjers
- Java or LPC (DGD)? Ryan Underwood
- Reputation systems: a possible path for investigation J C Lawrence
- Reputation systems: a possible path for investigation david.l.smith@mail-x-change.com
- Reputation systems: a possible path for investigation Brian 'Psychochild' Green
- Reputation systems: a possible path for investigation Andrew L. Tepper
Something related to reputation systems are ranking systems. In a
reputation system, you say "I trust JC, and JC trusts Ed, therefore
I can trust Ed. A good property of a ranking system is that if JC
tends to beat Ed, and I tend to beat JC, then I should tend to beat
Ed.
Both systems are things that players will want to game
(manipulate). We use an algorithm in A Tale in the Desert which has
this property, and is game-resistant. It's described here:
http://www.egenesis.com/erank.rtf
Andy Tepper
<EdNote: Below, minorly reformatted for clarity>
--<cut>--
The eGenesis Ranking SystemA highly cheat-resistant, zero-sum
ranking system for multi player games
Andrew L. Tepper, BS
Josh M. Yelon, PhD
February 15, 2002
Problem:
Ranking systems such as ELO are often used to rate chess
players. These systems attempt to encourage players to compete
against others of a similar skill level by making matches between
an advanced player A, and a beginner player B be risky for A, and
rewarding for B.
In the case that A wins:
A gains a small number of points
B loses a small number of points
In the case that B wins:
A loses a large number of points
B gains a large number of points
These systems are subject to cheating. If a player A wishes to
boost his rank, he conspires with friends B, C, and D as follows:
- B Plays C several times, intentionally losing to boost C's
rank.
- C Plays A several times, intentionally losing to greatly boost
A's rank.
- Another friend D is brought in. The three play each other in
the same pattern to boost D.
- D Plays A several times, intentionally losing to further boost
A's rank.
The pattern is recursive: Players P1 thru PN-1 play each other to
boost PN-1, who then loses to PN.
Solution:
Each player in the system starts with a vector of 256 bits. Half
of these bits, randomly selected, are set. A players rank is the
number of bits set.
We number each bit 0-255. When a match occurs between two players,
A and B, a series of 32 hash values in the range 0 to 255 are
computed based on the players names such that the same two players
always yield the same 32 values. For each value, we inspect that
bit in both players vectors. When a bit is set in the losers
vector, and clear in the winners vector, we transfer that bit:
clear it in the losers vector, and set it in the winners. In all
other cases we do nothing. This system forces players to play a
wide variety of others to boost their rank.
Still, its possible to boost your rank by only playing
beginners. To combat this, we modify the algorithm as follows:
Each player starts with their bit vector clear. They also have 128
reserve bits which are stored simply as a count. A players rank is
equal to the number of bits set in his vector, plus the number of
bits in reserve. When a match occurs, bits are transferred as
above, with the following additional rule: When one of the 32
values bits are clear in both the winners and losers vectors, a
bit is attempted to be transferred from the winners reserve to a
specially selected clear bit in his own vector. This specially
selected bit is based on a different hash function that again
incorporates both players names. The losers vector and reserve are
untouched. In this way, someone who only plays beginners doesnt
affect his own rank, but may boost the beginners rank.
Practical Applications:
We designed this ranking system to be used in massively
multiplayer online games such as A Tale in the Desert
(http://www.ataleinthedesert.com). There are some additional
considerations to take into account.
First, beginners should have the sense that they are making
progress when they first start playing. To resolve this, we report
everyones rank as simply the number of bits set in their vector,
ignoring their reserve bit count.
Second, friends like to play each other repeatedly, and should see
progress on more than just the first game. To resolve this,
instead of attempting transfers for all 32 bits, we randomly
select 8 of the bits to transfer. In subsequent matches won by the
same person, there are fewer bits to transfer, and eventually this
number dwindles to none, but it allows for friends to play each
other repeatedly and still see some progress.
--<cut>-- - Reputation systems: a possible path for investigation Matt Mihaly
- Reputation systems: a possible path for investigation Vincent Archer
- Reputation systems Castronova, Edward
- Reputation systems J C Lawrence
- Mapping real money into MUD-Money Ren Reynolds
- MudDev Faq - part 2 Marian Griffith
- PHP muds Peter Harkins
- PHP muds Torgny Bjers
- Slashdot story about review of Bartle's new book Christer Enfors XW {TN/PAC}
- Slashdot story about review of Bartle's new book Dave Rickey
- Slashdot story about review of Bartle's new book Evan Harper
- Slashdot story about review of Bartle's new book Richard A. Bartle
- Slashdot story about review of Bartle's new book Tamzen Cannoy
- Slashdot story about review of Bartle's new book Kerry Fraser-Robinson
- Slashdot story about review of Bartle's new book Richard A. Bartle
- Slashdot story about review of Bartle's new book Dave Rickey
- Slashdot story about review of Bartle's new book Marc Bowden
- The lack of Creativity and Beauty a game user james_nesfield@nesfieldcapital.com
- Artists and Copyrights Derek Licciardi
- Artists and Copyrights Paolo Piselli
- Artists and Copyrights Marian Griffith
- Artists and Copyrights Paul Dahlke
- Using Windows Scripting Host Owen Matt
- Using Windows Scripting Host F. Randall Farmer
- Using Windows Scripting Host Karl Dyson
- Using Windows Scripting Host Tess Lowe
- Better Game Design through Data Mining David Kennerly
- Better Game Design through Data Mining Chris "Diamonds" Stewart
- When Will Player-Avatar Integrity Be a Feature of Persistent Worlds? vladimir cole
- When Will Player-Avatar Integrity Be a Feature of Persistent Worlds? Martin Bassie
- When Will Player-Avatar Integrity Be a Feature of Persistent Worlds? Craig H Fry
- When Will Player-Avatar Integrity Be a Feature of Persistent Worlds? Matt Mihaly
- When Will Player-Avatar Integrity Be a Feature ofPersistent Worlds? Michael Tresca
- When Will Player-Avatar Integrity Be a Feature ofPersistent Worlds? Baar - Lord of the Seven Suns
- [Fwd: Metrics for assessing game design] ceo
- Examine/Look Elia Morling
- Examine/Look Ammon Lauritzen
- Examine/Look Marc Bowden
- Examine/Look Lars Duening
- Examine/Look Eamonn O'Brien
- [BUS] Account-management systems ceo
- [BUS] Account-management systems Rayzam
- [BUS] Account-management systems Christopher Allen
- Job opportunity on Star Wars Galaxies Koster, Raph
- NCSoft yearly report Mathieu Castelli
- MUD using the .net framework Norman Beresford
- MUD using the .net framework John Buehler
- MUD using the .net framework James F. Bellinger
- MUD using the .net framework Linder Support Team
- Virtual property lawsuit in China Koster, Raph
- Virtual property lawsuit in China Nicolai Hansen
- Virtual property lawsuit in China Daniel Anderson
- Virtual property lawsuit in China Kerry Fraser-Robinson
- Virtual property lawsuit in China Vladimir Cole
- Virtual property lawsuit in China Ren Reynolds
- Virtual property lawsuit in China Nicolai Hansen
- Virtual property lawsuit in China ren@aldermangroup.com
- Expected value and standard deviation. Jeff Cole
- Expected value and standard deviation. Scion Altera
- Expected value and standard deviation. Jeremy Hill
- Expected value and standard deviation. katie@stickydata.com
- Expected value and standard deviation. Ben Chambers
- Expected value and standard deviation. Zach Collins {Siege}
- Expected value and standard deviation. Ben Chambers
- Expected value and standard deviation. Robert Zubek
- Expected value and standard deviation. Kwon J. Ekstrom
- Expected value and standard deviation. Eamonn O'Brien
- Expected value and standard deviation. Kwon J. Ekstrom
- Expected value and standard deviation. Freeman, Jeff
- Expected value and standard deviation. Zach Collins {Siege}
- Expected value and standard deviation. Bernard Graham
- Expected value and standard deviation. Freeman, Jeff
- Expected value and standard deviation. Jeff Cole
- Expected value and standard deviation. Koster, Raph
- Expected value and standard deviation. Katie Lukas
- Expected value and standard deviation. Fidelio Gwaihir
- Expected value and standard deviation. Katie Lukas
- Expected value and standard deviation. Matt Mihaly
- Expected value and standard deviation. Martin Bassie
- Expected value and standard deviation. Katie Lukas
- Expected value and standard deviation. Matt Mihaly
- Expected value and standard deviation. Paul Schwanz
- Expected value and standard deviation. Matt Mihaly
- Expected value and standard deviation. Koster, Raph
- Expected value and standard deviation. Paul Schwanz
- Expected value and standard deviation. Amanda Walker
- Expected value and standard deviation. John Buehler
- Expected value and standard deviation. Kwon J. Ekstrom
- Expected value and standard deviation. Jeff Cole
- Expected value and standard deviation. Paul Schwanz
- Expected value and standard deviation. Dr. Cat
- Expected value and standard deviation. David Loving
- Expected value and standard deviation. Pat Ditterline
- Expected value and standard deviation. Michael Chui
- Expected value and standard deviation. Matt Mihaly
- Expected value and standard deviation. Kwon J. Ekstrom
- Expected value and standard deviation. Chanur Silvarian
- Expected value and standard deviation. Katie Lukas
- Expected value and standard deviation. Daniel.Harman@barclayscapital.com
- Expected value and standard deviation. Oliver Smith
- Expected value and standard deviation. Daniel Anderson
- Expected value and standard deviation. Koster, Raph
- Expected value and standard deviation. Nicolai Hansen
- Expected value and standard deviation. Dark Lamenth
- Expected value and standard deviation. Fidelio Gwaihir
- Expected value and standard deviation. Ola Fosheim Grøstad
- Expected value and standard deviation. gbtmud
- Expected value and standard deviation. Tom "cro" Gordon
- Expected value and standard deviation. Sheela Caur'Lir
- Expected value and standard deviation. Roger Hicks
- Expected value and standard deviation. Ola Fosheim Grøstad
- Expected value and standard deviation. Freeman, Jeff
- Expected value and standard deviation. Matt Mihaly
- Expected value and standard deviation. Tom "cro" Gordon