# Wiki Game Theory

{\displaystyle \sum _{\begin{aligned}p_{i}\in Human\\m_{i}\in Memes\end{aligned}}\lVert Know_{p_{i}}(m_{i})\rVert }

This early Wikipedia slogan "Sum of human knowledge" is a lens I'll use to look at what we do. Looking back at the early days of the project and seek to reinvigorate the current incarnation to a similar degree. But also looking forward at the future and what it can yet become if the project is able to adapt to changes in the environment.

My little investigating is based on deconstructing the problems through the paradoxes and dilemmas being faced by the project in the hope of finding the synthesis between the different approaches. It develops by modelling some of the phenomena described by researchers of the Wiki and occasionally their refutation by experienced Wikipedians.

## Sub pages

As this page grown in size I'm moving the material to sub pages and transculuding them below to allow printing and a complete bibliography.

disclaimer: formula and calculation presented here are far too onerous to update and I am aware of many improvements that I've made elsewhere.

## Preliminaries

Without definition them we assume that the article is a container of memes.

Subjective knowledge of person ${\displaystyle i}$  is the ability to map truth value to a memes.

${\displaystyle Know_{i}:m\in M\mapsto \{-1,0,1\}}$

where -1 is false, 0 is unknown and 1 is true.

### Objective knowledge

The NPOV doctrine may be modeled by a assuming the existence of an objective knowledge function which can accurately assign all memes their objective truth value. This is the Wikipedian's axiom of choice.

The objective knowledge axiom simplifies the analysis and can be relaxed or dropped. I'wll return to this when considering political disputes and foreign influence and edit wars. Since only god would be completely objective we can state this as

${\displaystyle Know_{objective}:m\in M\mapsto \{-1,0,1\}}$

• 1 is considered true (by people in the know) e.g. a scientific doctrine.
• -1 is considered false (by people in the know) e.g. an urban myth.
• 0 represents memes on whose validity no consensus could be reached
• in the objective function the domain of zero 0 is minimal.

## Good faith - a non conflict axiom

For now we will assume that all people's subjective knowledge is a subset of the objective function (all people will agree on what they know) later we may consider the case of people with conflicting subjective belief.

i.e. where the knowledge deviates from the objective function it is set to zero.

${\displaystyle Know_{i}=\delta _{i}\times Know_{god}}$

where ${\displaystyle \delta _{i}}$  is the subjective characteristic function.

### Domain expertise & Common sense

The domain expertise of a person is defined as a number between one and zero representing the normalized knowledge assignment on memes unique to articles spanning the SD (subject domain).

${\displaystyle DE_{SD}(p_{i})={\frac {1}{N}}\sum _{m_{i}}^{N}\lVert Know_{p_{i}}(m_{i})\rVert }$  where ${\displaystyle m_{i}\in SD}$

Using this function we can rank people using their knowledge in different subject domains. We consider that the domain expertise allows experts to self select. We call such domains as self selecting.

Self selction would be a useful concept to specify mathematically.

e.g. a group of people is self selected if acquiring the selection criteria is has more expense then benefits for for non members.

Which means that there is no premium on common sense knowledge which everyone has and little value on knowledge that can be squired easily. So the presence of common sense memes in article effect all people the same way. They are however take part in the normalizing of domain knowledge. So they can be seen as having a supirious effect and should ideally be removed from articles. [note 1]

Knowledge domains change over time. But at any given time are considered static.

### ${\displaystyle Dom_{Pol}}$  - Policy Domain

We recognize a Subject domain spanning the whole wiki called the Policy domain.

• It is considered as a self selecting[note 2].
• Since it has the most domain experts, it is the least self selecting one.

While ${\displaystyle Dom_{Pol}}$  spans all the wiki it is a small domain. It size is a factor of the number of policy memes. These number less then a hundred. (But more is we count templates and violations as memes too). It is also the faster changing domain.

### Utilities

Utilities become problematic in the sense that they try to formalize mathematically arbitrarily complex set of preferences of people. Utilities should be written locally for the smallest independent units. When the game is repeated or when many outcome are possible (a complex game) it becomes difficult for people to consistently determine their preference.

Anyone looking to make use of such results would be well advised to conduct experiments that test such assumptions as well.

Utilities model people's preferences for different outcomes. They codify prefrence for a given choice and for payoffs of games. Utility is about relative values not absolute values. So to compare 2 player's preferences - ignore the actual values and look at what is next best/worst outcome for each.

Utilities can be expressed as linear expression in:

• Work - time to do edit (related to experience and complexity)
• Material - memes
• Reputation - good to model seniority
• Face - good to model editor death

Overhead and Shared Community costs. In case of a wiki material is modeled memes.

${\displaystyle U_{i}(A)={\frac {\sum {D}}{\left\vert A\right\vert }}.{\frac {\sum {P_{T}}-L\sum {P_{F}}}{\left\vert P\right\vert }}}$  is person's i utility from an article A

${\displaystyle DE_{SD}(p_{i})={\frac {1}{N}}\sum _{m_{i}}^{N}\lVert Know_{p_{i}}(m_{i})\rVert }$

## Some Simple Games

### Coordination & Coordination Failures

Coordination of task requires that agents can agree and locate needed resources within a limited time-frame. While CSCW researchers frequently refer to coordination cost and to communocation costs as setbacks faced by agents articulating collaborative goals, their definitions are to vauge to be quantified. On the other hand Game theory as I originally envisioned it used abstract utility to define costs as well as payoffs but this soon became an barrier to testing the validity of the theory. I have therefore extended the the notion of an abstract utility representing the preferences of agents in a game with a more concrete surrogate. This is a "currency" that might be more conductive for CSCW research.

The currency of interest is (time, information)

I call the inability to coordinate as a coordination failure. How to quantify a such a failure ?

2, 3 & N agent coordination game
with mutually exclusive POV
 POV1 POV2 POV 1 3,2 0,0 or[note 3] POV 2 0,0 2,3 2 agent consensus coordination
 POV1 POV2 POV 1 3,2,2 0,0,0 or[note 3] POV 2 0,0,0 2,2,3 a 3 agent consensus coordination
 POV1 POV2 POV 1 3,...,3,2,...2 0,...,0 or[note 3] POV 2 0,0 2,...,2,3,...3 N agent consensus coordination

#### The Connection Problem

It is implicitly assumed that the agents can solve the connection problem[1] meaning that each agent is assumed to be able to locate the other agents who might have capabilities which are necessary for the execution of tasks.

However this is an assumption that should be challenged. One of the most common sources of collabotation failure is the lack of technology to bridge connection problems by suitable networking amongst required social roles as resource owners. The current remedy being used is thematic coordination spaces. These allow agents to locate suitable collaborators. A second remedy used to handle vandalism is to use robots which function in a number of coordination spaces silmultaneously. However while robots are useful for handling specific tasks they are not very good for communication.

Examples for Coordination space are:

• Article pages (ant path coordination[2] using inline tags)
• Article talk pages (to coordinate)
• Page History (to coordinate via editing) - using revert
• Project Pages
• Village Pump
• Community Notice Board
• Afd
• Comment pages for policy discussions
• Mailing Lists
• Irc Channels

These are locations where coordination negotiations take place.

However this assumption is unlikely to be true in the case of a newcomer. For example newcomers are exceedingly [citation needed] rare at AfD and even rarer at Deletion Review WP:DR. this component of the cost can be quantified.

#### The Consensus Game - Unanimous Decision

Coordination between two agents is traditionally modelled by Battle of the Sexes. The rules of the game are :

1. Agents must choose one of a set of mutually exclusive coordination options (e.g. a Point of view in a policy discussion or an opinion/fact to be expressed in an article)
2. They move simultaneously (or at least without exchanging information)
3. Each agent has the following preferences:
1. best is to have Consensus centered on their point of view.
2. next is to have Consensus centered on another point of view.
3. worst of all is to not have consensus at all.

The coordination challenge is as follows - they have to rendezvous at a common location. However each is driven by a conflicting personal preference.

Note: A real wiki has infrastructures for messaging (inline and article tagging) and coordination spaces coordination which should empower the agent to always reach consensus - yet in reality coordination failures are common (in some essential scenarios over 50%). Despite the apparent restriction of the model in general this model (and the variant below) can be used to effectively model a very large part of coordination in a wiki.

There are many coordination problems in Wikipedia but a good illustration is what I call the consensus game. This is a situation involving two or more editors attempting to reach consensus. Each agent must contribute a necessary component (technical information, writing skills, language skills, access to paid wall resources, knowledge of policy, user privilege) to reach a new consensus. However if any of a single component is not accounted for than the coordination fails and the only consensus possible is to delete the article. (Cases where editors actively compete are covered under wars of attrition) In this game the preferences for each agents are:

1. reach (coordinate) a consensus based on their POV.
2. reach (coordinate) a consensus based on the opposing POV.
3. return to an old consensus - may often not be an option (e.g. at creation or at AfD).
4. coordination failure.
• I call the version shown here as the Consensus coordination between two, three and many members is shown with two option are presented and there is no advantage in returning to the previous consensus. [note 3]
• Note that in the 3 person version the second player will get a low score in both cases - this naturally models a non involved third agent who benefits from consensus but does nor have a preference. (c.f. conflict resolution)
• Note that in the N person version there will be additional involved agents coming in who mirror the first/second/third agent. However they can no more force consensus than the case of 3 agents. They also do not posses resources which allow them to bargain for better payoffs in return for joining the coalition.

In many discussions there are several coordination choices (reSome issues are that there are at times more that 2 POV. This makes coordination harder (overall returns smaller).

What is the aspect of Coordination we would need to model:

1. The cost of establishing a coordination (social) network of agents working in a limited (coordination) space and time possessing the required resources such as:
• domain knowledge
• writing skills - good english & knowledge of the manual of style.
• language skills (translation/verification of foreign language sources.
• access to scarce research materials - e.g. paid wall resources (New York Times; Highbeam; Web of Science etc).
• knowledge of policy and norms.
• user privilege - the ability to edit protected pages
1. a probability reflecting the scarcity of the agent/resource needed for coordination.
2. In practice the number of agent seems significant. Two agent are doing 10 more work due to C2. If they are in conflict they can have a coordination failure and require a third opinion. To coordinate a larger group of than two seems to require 10 more work than the case of 2.
3. some coordination teams also need to form coalitions. This process is defined as agreeing on a payoff allocation based on access to resources. It is usually modelled by a sequence of appeals which can be justified or not. (technical - needs to be explained in laymen's terms).
2, 3 & N agent coordination game
with N mutually exclusive POV
 POV1 POV2 POV 1 3,2 0,0 POV 2 0,0 2,3 2 agent consensus coordination
 POV1 POV2 POV3 POV 1 3,2,2 0,0,0 0,0,0 POV 2 0,0,0 2,3,2 0,0,0 POV 3 0,0,0 0,0,0 2,2,3 a 3 agent consensus coordination
 POV1 POV2 POV3 POV 1 3,2,..,2 0,...,0 0,...,0 POV 2 0,...,0 2,...,2,3,2,...,2 0,...,0 POV 3 0,...,0 0,...,0 2,..,2,3 a N agent consensus coordination

This game has two pure strategy Nash equilibria and one mixed strategies equilibrium. However all the options are suboptimal. A correlated random token could make the mixed strategy ideal - allowing both users to play the pure equilibria options and get equal maximal payoffs. However without such a coordinating device each player will choose his POV more frequently. For the payoffs listed above, each player whom expresses a preference fill opt for the preferred POV probability 3/5. They will miscoordinate with a probability of 13/25 leaving an expected return of only 6/25.

For non involved players the situation is simpler - they will miscoordinate with a probability of 1/2 and have an expected return of 1.

My informal point of view based on recent AfD participation is that there are at least two classes of coordination tasks -

Coordination1 - cost of establishing a (social) network of agents with a set of qualifications (domain knowledge (usually scarce) and policy knowledge) under a space-time constraint (14 days at a AdF page). I'd call this a MicroCosmic Coordination tasks - as it requires creating a social network mixing insiders and outsiders.

Coordination2 - a second class of coordination tasks seems to require a second round of coordination and involve the task of establishing coalitions. This also has a larger communication component. I'd call this a Political Coordination task

Afd discussions - seem to have 2 + modes of closing similar to the Ideas presented by Ayelt Oz in her paper on acoustic separation.[3].

• Fast (closely related and very similar to CSD) - do not require Micro Cosmic Coordination
• Slow with a good likelihood of repeated coordination failures - Micro Cosmic failures oft lead to relisting (extending the deadline), while political failures leads to no consensus at closure, multiple Afd of the same article over long timeframes.

Some points still needing more looking into

• yet another version of the coordination game with N-gaents with a preference fo reaching ever smaller consensus prefered to no consensus at all. (but with equal sized miscoordinated blocks disqualifying each other.
• repeated participation at Afd discussions leads to ad hoc coalition. These however turning into semi permanent ones.
• Semi permanent coalitions reduce coordination costs
• establishment of task related Afd notice boards support longer lasting coalitions by providing coordination discount for locating Afd discussions in related knowledge domains. e.g. PornBio via WikiProject_Deletion_sorting categories - if the categories have enough active participants (low diversity in editors population therefore leads to coordination failure at AfD ).
• http://www.ri.cmu.edu/pub_files/pub1/shehory_onn_1997_1/shehory_onn_1997_1.pdf

## Playing In Turns

let us consider a couple of variant. we start with a version of the consensus game for Afd where row players nominates deletion and the colum player will want to keep. No Consensus means keep.

 Delete Keep Delte 3,2 0,2 Keep 0,2 2,3 2 agent afd - NC means keep

## Edit Wars

We can model edit wars using a war of attrition model, this is a model originally formulated by John Maynard Smith,[4] a mixed evolutionary stable strategy (ESS) was determined by Bishop & Cannings.[5] Strategically, the game is an auction, in which the prize goes to the player with the highest bid, and each player pays the loser's low bid (making it an all-pay sealed-bid second-price auction).

In the context of an edit war it is best to consider the dynamic formulation:

Two players are involved in a dispute:

• The value of the object to each player is ${\displaystyle v_{i}>0}$ .
• Time is modeled as a continuous variable which starts at zero and runs indefinitely.
• Each player chooses when to concede the object to the other player.
• In the case of a tie, each player receives ${\displaystyle v_{i}/2}$  utility.
• Time is valuable, each player uses one unit of utility per period of time.

The evolutionary stable strategy is a mixed ESS, in which the probability of persisting for a length of time t is:

${\displaystyle p(t)={\frac {1}{V}}e^{(-t/V)}}$

As is the case with ESS - it does not guarantee the win; rather it is the optimal balance of risk and reward. The outcome of any particular game cannot be predicted as the random factor of the opponent's bid is too unpredictable. That no pure persistence time is an ESS can be demonstrated simply by considering a putative ESS bid of x, which will be beaten by a bid of x+${\displaystyle \delta }$ .

### Interpretation

Wikipedia has often been described as a space where content of articles is decided by the most persistent editor. However in a war of attrition - all players must pay a price to participate and the winner can end up with a Pyrrhic victory.

Some open question for this model:

1. What is the utility assigned to articles by editors.
2. What factors affect this utility. (Considering a block, page protection, admin action, sanction may result from a content dispute, and that it may spread to other articles.)

## Motivation

The original motivations was to investigate factors which can increase cooperative behavior in a wiki[note 4]. However it soon became clear that questions are of general interest could be answered by perusing this investigation. These include:

• how to increase user engagement?
• how to increase article quality?
• what (policy) causes a wiki to close/open up as an information society?
• how to make smaller wikis more successful?
• can one evaluate the impact of a proposed policy in a lab settings?
• can one model/simulate/predict/define experiments to test impact of policy on editor loss/gender gap?

This work may be of interest to mainstream CSCW researchers since it will also consider behavior of non cooperative nature:

• conflict
• combat - wars of attrition
• competition
• control
• coercion
• vandalism
• elites

## Prior Art

Game theory has been used previously in a couple of papers discussion Wikipedia.

## The Method

• Wiki Game Theory, is the use of mathematical modeling of behaviour patterns existing within the community of a peer produced

CSCW such as Wikipedia.

This type of modeling can reflect four levels of reality:

1. reality - against which other models should be tested.
2. evolutionary models - models in which more variables are allowed to run wild. (changing policy, technology).
3. dynamic models - stability of static solutions, repeated games. (time dependent, fixed policy) changing parametric incentives.
4. static models - used in this.

As a theoretical framework it can be used to resolve paradoxes that been observed by community members and researchers.

• These models can be useful in four use cases.
• classify behavior - in a finer form using aspects of model. (players, strategies, information, etc).
• qualify - different behaviors - these can be strategies, pure equilibrium, mixed equilibrium and evolutionary stable strategies.
• quantify - to estimate the outcomes of rational behavior and detect obstacles to rationality.

In more practical sense this theory should make testable predictions:

• How agents will behave in a sufficiently large sample of modeled situations modeled by the game.
• How a game scenarios would be expected to evolve over time due to invasion of different strategies.
• Systemic fragility arising from different risk factors - such as cheating.
• meta-policy prescription that could increase the success of the wiki.
• it can be used to compare academic and Wikipedia and print models of knowledge development. (Wikipedia, research Journal and Print Encyclopedia respectively)

### Criticism

The game theoretic approach assumes rationality on the part of players. In realty players are known to make mistakes. However most of these models have been developed based on reports of both qualitative and quantitative research. And while motivation and rationality are difficult to encode numerically, the advantages of the models is that they simplify reality but are still able to capture phenomena at a level which allows making testable predictions.

## Mathematical Methods

Solving problems in the game theoretic paradigm is done according to a number of established solution concepts. For example

• MiniMax.
• Deletion of dominated strategies.
• Backward induction. (assume forward play will be rational).
• Equilibrium.
• ESS.
• Proper.
• Trembling Hand.
• Sub-game Perfect.
• Pure Strategy.
• Nash.
• Correlated.
• Forward iinduction. (assume past play was rational)
• Core.
• Shapely Value.
• Extended Core.
• QRE - to test models with players that are not always rational.

## Experimental Methods

From its inception the theory of games has been viewed as belonging to applied mathematics and accordingly its models compatible with experimental testing.

## Research Question 1 - Population Dynamics

General population dynamics questions are:

• How many members are there in the community of wiki X?
• What is the size of the active community?
• How many members occupy different social roles.
• How long do members occupy different social roles. (Life span)
• What is the population over time on different wikis?
• How do these relate to changes in the overall population of the wiki?
• How do these relate to readership of the wiki over time?
• How do these relate to existence of robot members of the population.
• what number of players in different social roles are required to sustain different wikis. (e.g. against incoming spam)
• how do wikis cope with catastrophic and non-catastrophic decline in population size and funding. (e.g. 9/11 wiki).
• relate changing population dynamics to openness of the wiki and its policy complexity.

WGT population dynamics are:

• What is the incidence of a specific game
• At each play what strategy was employed by each player
• What was the outcome.
• What is the ESS.
• How do theoretical payoffs correlate with outcome payoff in one shot and many shot games given cost and benefits?

To answer so many interrelated questions - it will be necessary to :

1. gather information on these numbers.
• API based checkers running periodically on many wikis.
• dump based script that get this using SNA.
2. produce a population models.

To support quantitative investigation especially dynamically modeling of stability and fragility it is necessary to use real world numbers.

Note: sampling this information should be automated and run daily.

DATE: XX/XX/XXXX

Users
Group API Count > 1/4*7 day mean activity count > 1/4*30 days activity count > 1/4*90 day activity count
autopatrollers 2677
rollbackers 2677
registered 2677
anonymous
bots
Articles:
Group Count 7 day SD Source
total articles
creation daily
CSD daily 100
WP:Prod daily 100
afd currently open 700 350
Simple Events:
Group Count inc dec deletion delta
Articles +20 +200 -180 0
Talk +20 +200 -180 0
Talk - h-index +20 +200 -180 0
Article + Talk hindex +20 +200 -180 0
Tagging Art +100 +200 -300 100
Tagging inline 40 -40 20
Cat 330 350 -20 0
Policy Ref 100 100 0 0
Policy Mass 100 120 -20 0

Bases on SNA we can also look at "Social events"

Social Events:
Group Count %anon %bot %admin %role X
Edit War 20 40 20 30 0
WikiQuette 600 80 15 1 4
Biting 40 50 45 1 4
Consensus Success 700 5 45 25 95
Consensus Failure 100 5 0 5 95

TO update these on a regular basis I will have to extend the following script User:OrenBochman/Rbot/Variables.py and expand it to to work with SNA packages.

## Research Question 2 - Social Learning Curve

• What are the rules of the game?
• How many rules of policy must be mastered by players ?
• How has WP name space growing compared with Mainspace ?
• How many help pages are the

Policy:
Group Count 7 day SD Source
total WP: pages
Policy pages 100
Guideline pages 700 350
Essays 700 350

is policy growing relatively faster than article space? [6]

## Research Question 3 - Cost to Benefit Estimation

Game have cost and benefits associated with them. Writing articles in particular has costs associated with them. Costs can be discount using technology. UI support (Google scholar Mashup) or ProveIt UI.

### Research Question 3.1 - Cost of Editing

Editing Costs:

• Time
• Information complexity (citation).
• Information complexity discount (e.g. using proveIt to cite).
• Editing Complexity (templates).
• etc

Editing benefits:

• Fun factor
• Advancement - recognition, barn stars.
• The gift: Mutual information gain.
• Leaning.
• Altruism:
1. Memematic Kin altruism [We are closely related Memematicaly]
2. Memematic Group altruism [We are closely related Memematicaly as a group]
3. Direct reciprocity effect.
4. Indirect reciprocity effect.
5. Network reciprocity effect.
• etc.

The question boils down to:

• In a given social situation how can we quantify cost to benefit ratios based on overall player interaction??

### Research Question 3.2 - Cost of Communication

Extending the question above to Communication scenarios, both simple games and looking at conversations and at repeated conversations.

### Research Question 3.3 - Cost of Coordination

• How many centralized coordination spaces are there
• Article Page
• AfD...MfD
• etc

## Edits - The Stakes

Lets our imagine our four editor working so if L =2.5 and an avarage edit changes 30 D_F and 4 D_P

(Links between Article A and B can be viewed as allowing A to depend on B, essential importing it's memes without gaining length. But the contribution of linked memes is reduced by the number of links they are removed from each other.)

Creating a new article produces U(A) and is likely to introduce new completely new memes into the wiki.

By editing an exiting article an editor contributes U(S(A)-U(A) in each revision. Where S is the successor function. goal D_T and D_F may be altered. Since this is not a domain expert almost always D_T will become D_F.

To a domain expert D = D_T + D_F and he can increase U(A) consistently by either introducing more D_T or removing D_F. To a policy expert D_T and D_F look the same and so he can increase U(A) by increasing P_A towards Sum P. However in pursuing this

The reviewers dilemma is as follows: An editor makes a revision, the first reviewer must then decide to accept or reject it.

Editor D (co-operates) Editor P (defects)
Editor D (co-operates) Each gets U(S(A) - U(A) D get 0
P gets ()
Editor P (defects) Editor A: goes free
Editor B: 1 year
Each serves 3 months

## Information assymetry

In the context of a wiki we diffrentiate between domain memes ${\displaystyle D_{1}...D_{N}}$  and ${\displaystyle P_{1}..P_{N}}$  policy memes.

the lack of of knowledge in a domain the degree of knowledge in a domain is the is definde by

• To a domain expert ${\displaystyle D_{A}=D_{T}\bigoplus D_{F}}$
• To a policy expert ${\displaystyle D_{A}=D_{T}\cup D_{F}}$
• This diffrence is the primary source of information assymetry in the game. Most domain requires years of study.

Let the ${\displaystyle P}$  policy be described by Policy memes and violations ${\displaystyle P_{1}...P_{N}}$ .

• To a policy expert ${\displaystyle P_{A}=P_{T}\bigoplus P_{F}}$
• To a domain expert ${\displaystyle P_{A}=P_{T}\cup P_{F}}$

Notes:

• As ${\displaystyle {\left\vert P\right\vert }}$  increases there is a new role called a policy expert.
• This diffrence is a secondry source of information assymetry in the game. Learning to edit a wiki takes weeks of study.
• so ${\displaystyle {\left\vert D\right\vert }>>{\left\vert P\right\vert }}$
• overt a year change ${\displaystyle {\left\vert P\right\vert }}$  can change dramaticly, but ${\displaystyle {\left\vert D\right\vert }}$  is constant.

## Utility of Article and Edit

Here are five stereotypical Utility functions:

• Wise - Objective utility function (information symmetry) or policy savvy domain expert

$\displaystyle \label{eq:someequation} U_W(A) = \frac{ \sum{D_T} - L \sum{D_F} }{\left\vert A \right\vert} . \frac{\sum{P_T}- L\sum{P_F}}{\left\vert P \right\vert}$

• DExper -the Domain Expert's utility of an article

${\displaystyle U_{D}(A)={\frac {\sum {D_{T}}-L\sum {D_{F}}}{\left\vert A\right\vert }}.{\frac {\sum {P_{A}}}{\left\vert P\right\vert }}\qquad \qquad \qquad (2)}$

• the Policy expert's utility of an article

${\displaystyle U_{P}(A)={\frac {\sum {D}}{\left\vert A\right\vert }}.{\frac {\sum {P_{T}}-L\sum {P_{F}}}{\left\vert P\right\vert }}\qquad \qquad \qquad (3)}$

• the Noobe's utility of an article

${\displaystyle U_{N}(A)={\frac {\sum {D}}{\left\vert A\right\vert }}.{\frac {\sum {P_{A}}}{\left\vert P\right\vert }}\qquad \qquad \qquad \qquad \qquad \qquad (4)}$

${\displaystyle U_{B}(A)={\frac {L\sum {D_{F}}}{\left\vert A\right\vert }}.{\frac {\sum {P_{T}}-L\sum {P_{F}}}{\left\vert P\right\vert }}\qquad (5)}$

• the vandal -

${\displaystyle U_{W}(A)={\frac {\sum {D_{T}}-L\sum {D_{F}}}{\left\vert A\right\vert }}.{\frac {\sum {P_{A}}}{\left\vert P\right\vert }}\qquad \qquad \qquad (6)}$

where L is the lemon coefficient and is greater than 1. also we have omitted the advantage conferred by links.

### Sorting Gold Nuggets V.S. Lemmon

Lemmons are used in the sense of The Market for Lemons[7]

There is an overall lack of ability to differentiate good information sources from bad ones. This includes good citations, from bad citations. Good external links from bad ones, reliable sources from unreliable ones. Using and appealing to domain experts to check and correct entries is one solution. However can these academics be asked to do so in good faith - if any ignoramus may revert their corrections. There is generally no way to assess the level of fact checking that has gone into an article.

The problem is that in every level a few bad lemons editors can drive out the really knowledgable ones from contributing.

One such situation involves say a high-school student who knows the rules of style but nothing more wishing to make content more encyclopedic. By reformatting the information to fit a more factual style, Good prose can become a mish mash of factoids that end up mentioning details in an inaccurate way. Inaccurate means wrong.

## A Catalog of Social Roles

Wikipedians can occupy different social roles. Work on Usenet and other communities have indicated that social roles of different community members of virtual communities can change over time. On Wikipedia with its many coordination spaces it is entirely possible for one person to occupy different social roles.

Below is an attempt to catalog these roles and identify their

1. a functional definitions
2. network structure indicators
3. game theoretic indicators (to be added)
4. linguistic clues to power see bellow

By looking at SNA and Linguistics one hopes to produce a language independent methodology.

I also theorize that looking at actions of adding and removing sentiment from text is an indicator of emotional value worth investigating.

### Information Roles

Information roles are :

1. Censor - reduces information contents
1. Creator - add new information
1. Mover - moves info around and rephrases

Leadership roles require exercising power and providing leadership. Leadership also requires sharing of power and delegating responsibility as the community grows. The top issues with leadership roles are therefore:

2. exercising power
3. continuity, where the exists a failure to transfer the leader's role as the project matures.
4. stepping aside - good leaders let go of power once this is beneficial to their project.

Name functional def SNA affect Social Indexicality[8]
Absent Leader[9] inactivity will correlate with a decline in the project. Strong correlation of leader activity with the production of people who interact with him. negative
Admin gains power though elevated access
• Will initiate admin action. (Block,Ban,Page Protection, etc - )
• Will be asked to do the above.
neutral
Benevolent Dictator[9] the originator whose benevolence prevets forking but does not pass on the role Project Centrality, Inner Centrality, Eigenvector Centrality neutral
Elder [10] Ex Celebrity who contributes less but has the community's respect. Ego network ratio of edit to talk has changed over time from edit to communicate. Communication is based on Questions positive
First Servant[9] Leads by example, eschews power. positive
God King[9] excessive use of authority will have an admin actions than request negative
Role Model[9] teaches by example zero or low ratio of admin action events to leadership events positive

Some questions:

• how do we characterize leadership events.
• how do we characterize admin events. Can we score and correlate them based on incoming requests ?

### Hero Role

Heroes are the role assigned to core community members. Heros can become leaders. The issues of hero roles are: recognition, aspirations, ownership.

Header text functional def SNA affect Social Indexicality[8]
UnsungHero[9]
Wiki-gnome
make many small edits c.f. substansive role below small uncontroversial edits on many unrelated articles pages; alters are many and disconnected, minimal need for talk positive
AttentionSeeker[9] require recognition unlike gnomes will have more action on talk pages negative
Celebrity [10] highly visible prolific contributor (core contributor) positive
RoleModel[9] positive
VestedContributor[9] inflated sense of ownership, which may lead to unnecessary conflict negative
WikiMaster[9] positive

### Community Role

• Lurker
• Newby
• VisitorRole[9]
• GuestRole[9]
• CommunityMember[9]
• ModeratorRole[9]
• HostRole[9]
• DeveloperRole[9]

### InterWiki Role

• RefugeeRole[9]
• GateKeeper[9]
• BoundarySpanner[9] (GluePeople)

### Eccentric Role

• CourtJester[9]
• FringeArtist[9]
• AntiAuthoritarian[9]
• Trolls[10]
• UsurperTroll[9]
• SpoilSportTroll[9]
• Spammer
• HyenaRole[9]
• EccentricCharacter[9]
• ArthurStace[9]
• VestedTroubleMaker - can get away with bending rule but is not part of a cabal (has no mutual defence pact, cannot influence the hiveMind etc)
• Biter - Bites new-comers (or old timers)
• VandalFighter
• AnonymousDonor[9] unregistered user (long tail)
• KnowItAll[9]
• RecordKeeper[9]
• PoliceForce/VigilanteJustice[9]

### DiscussionRoles

• Wiki:JustaStudent[9]
• ColdBlanket[9]
• DramaticIdentity[9] feature:negative individual speaking "for an entity".
• FlameWarrior
• Ranter [10]
• Question persons[11]
• Dicussion person[11]
• Discussion catalyst[11]

note: network dynamics imply that there would be ratios between certain roles. For example an answer person will need many question persons.

### events of interest

events are best organized as parts of games between two or more people.

• ham spam game:
• article creation
• article deletion nomination.
• spamming
• spam/vandal fighting
• communication games - [note 5]
• Q&A Game
• question
• response
• respond old player
• respond new player
• Coordination games (consensus
• consensus
• no consensus vote

## Other concepts

• right to leave
• peer pressure
• fun factor

## Machiavellianism and other Psychological Factors

High Mechs will learn to work the system. In interacting with different social roles - they will learn how to manipulate these people to get their way. Some of these insights are available in the MeatBall web site. E.g.

• It could be using a policy to shutting down objections.
• It could be getting another person contribute (edit/research) 80% to their 20% work.

## SNA definition

• Some Social roles can be observed directly in SNA visualization.
• Others may be possible to extract using algorithms.

The above lists include "functional" social roles, are their SNA based social roles that are non functional. Of interest are the their parallels can be extracted via SNA.

• whole Web net can expose leader roles.
• Motif and Motif rations in an ego web.
• Motif and Motif ratios in an ego web.

## Lexical Analysis of Social Relations

1. a Linguistic/Ethnographic indicators.
• metalinguistic words and expressions:
• metasemantic definitions: (X is a Y; X means a Y).
• metapragmatic events of language use (X ordered/insulted/teased me). - using verbs of speaking
• grammaticality judgments.
• capturing etiquette lapses (are these indicative of social standing) - ask DDG about these.
• capturing mixed signals e.g (hyper politeness and contempt)indicating veiled aggression
• can a speech between two different social roles - embodying different levels of the social hierarchy be differentiated. Are there discrepancy. Is there censure for exceeding your one's social standing.
• are there performative formulas for establishing indexical equality in speech. Are these formulas effective means for manipulating the empowered to comply with the wishes if the proletariat

Indexical Analysis:

• look at deictic expressions: pronouns, anaphora, demonstratives, markers of tense, mood and modality and where they collapse the axes of denotation and interaction.
• Identify the zero point defaults or origo of reckoning: -
• I; me; my; to me; we, us, our, to us; you, your,to you; here, there; now, then; hither, thither; this, that[8]
• here, and now. This is a moving reference point relative to the rest of the sentence
• Look at pronouns to recover the referent.
• which - directs the previous context for who the speaker, hearer, etc are
• that - directs to the circumstance of the utterer.
• Look at denotational variables:
• referents (entities).
• predicates (Qualities and relations of entities).
• propositions (state of affairs predicated entities).

this suggest the performative game where in an initial asymmetrical situation one side tries to force a symmetrical mode of communication

this type of analysis - requires parsing of the sentence, identifying verb valence, and resolving the deictic references used in each slot. (Pragmatical resolution)

the deictic selectivity of propositional acts shapes interpersonal realities by establishing a link between denotational and interactional variables.

Algorithm to identify the degree of specificity of an utterance. [8]

1. look at and annotate:
• interrogative mood
• definiteness of determiners
• past tense
1. enumerate specific nodes with +; unspecific with - and non deictic cats with () i.e. neutral

With tight linkage the propositional act shapes the interaction. When the linkage is loose, the content appears to transcend the moment when the act was made. Looseness can create generic laws, contracts, commitments, stereotypical politeness [8]

Role fractions of speaker[8]:

• Animator - one who physically produces the current utterance.
• Principal - one held responsible for utterance propositional content.
• Author - one held responsible for utterance wording.
• Figure - persona performed through the act of utterance.

Algorithm to identify the degree of modalization and explicit performative locution. [8]

1. look at and annotate each predicate with +/- for each feature:
• interrogative mood.
• past tense.
• passive voice.
• perfective aspect.

speech chain transmission:

... [S->R] [S->R] [S->R] [S->R] ...
+--+  +---+  +---+
--------------------------> time

S:= sender


speech chains form networks.

### Registers

• A semiotic register: a register where language use is not the only type of

sign-behavior modeled, and utterance not the only modality of action. A register of discourse is a special case.

• Enregisterment: processes whereby performable signs become recognized (and regrouped) as belonging to distinct register by a population.

can we identify different registers in a population.

## Population Dynamics

Population Dynamics and evolutionary game theory are the next two step in this paradigm.

However to perform meaningful population dynamics it is necessary to

1. catalogue relevant social roles.
2. identify methods for tracking users within these roles.
3. check that these roles cover the population of the community fairly well.
4. extract populations sizes and incidents from database and or dumps.
5. generate time series from the above.
6. explore how populations in roles participation in correlated rise and fall of population/incidents
7. build models using differential equations relating interrelated games.
8. ask the research questions which model M(G_I,...G_N) fit patterns in #6
9. ask the research question can models from #8 be used predictively.

## Evolutionary game theoretic models

With a working system of differential equations we can ask a bunch of new research question and simulate for the answer.

1. how will changing some parameter cause the system to evolve effect populations/player strategy/game rules/pay-offs/costs.
2. how will choices between one of serveral policies effect the above (this is a special case of changing game rules).
3. how will changing software effect the above (this is a special case of changing technology).

Using an evolutionary models we might predict creation/extinction of states/roles/events in the system as well as possible time-frames. an Example application would be to try to measure the impact of upcoming technologies such as wikidata, visual editor, flow (new talk pages) and echo (notification).

However it still remains to be seen if this approach can be useful - after all even if the simulation predicts a new role arising - can we understand what it will actualy mean in the new context.

## Social Engineering

Since the above two programs are too ambitions even for a single crackpot researcher, I have developed some of these ideas into a new paradigm which is a light wieght frame work to use Wiki game theory to solve practical community problems and surprisingly it uses Gamification and SNA as well.

## Some simple games

Notation and requirement - we need

1. players ${\displaystyle i,j}$  (small letters)
2. strategies
• ${\displaystyle S_{i}}$  all the strategies of ${\displaystyle player_{i}}$ .
• ${\displaystyle s_{i}}$  a strategy for ${\displaystyle player_{i}}$  where ${\displaystyle s_{i}\in S_{i}}$  .
• ${\displaystyle s}$  is a particular play of the game (a strategy profile).
• ${\displaystyle s_{-i}}$  is a particular play excluding the option of ${\displaystyle player_{i}}$
3. payoffs ${\displaystyle U_{i}(s_{1},...s_{i}...s_{n})}$  or abbreviated ${\displaystyle U_{i}(s)}$  using a specific play

what we'd like to find out is:

• What would be the rational outcome
• dominant strategies
• weakly dominant strategies.
• mixed strategies
• sub-game strategies
• any the equilibrium (nash, etc).
• ESS (dynamically stable solution)
• Can payoff design be used to change the rational behaviour.
• How fragile is a wiki game to invaders[note 6]

methodology involves

##### Cooperation
• Coordination[12] can be modeled by:
 Ban Pass Escalate 5, 5 0, 4 Pass 4, 0 2, 2 Fig. 1: Stag hunt example

#### The Normal Form Subgame

This sub game happens after an inciting incident sequence . This could be a spam edit or a perceived spam edit. An edit war may incite a block and repeted incident my incite a ban.

• under normal circumstances the user would only be banned in a spam incident. though it could also happen for other reasons those are

other types of ban games.

• the full game would add a second step where the Patroller bans the Editor with probability

${\displaystyle p_{ban}=\sum edits_{ham}-L\sum edits_{spam}}$

patroller (accept) patroller (ban)
Editor GF (ham) (0,0) (0,0)
Editor BF (spam) ${\displaystyle (0,-(1-p_{ban})({\frac {work}{\lVert Community\rVert }}))}$  ${\displaystyle (-p_{ban}\times work,-p_{ban}({\frac {work}{10}}+coo))}$
• a patroller who bans a user incurs a coordination cost COO.
• the spammer only incurs the cost of setting up a new account estimated at one work unit.
• not banning a spammer has marginal cost (his future damage to the community).

#### Practicalities

note: there is a probability of getting an admin or an non admin reversion.

## Communication & Coordination

A simplified diagram of how consensus is reached

communication and coordination are deeply intertwined. However we can use game theory to coax them apart.

As is intimated from the consensus flow chart most wiki games include intermediate steps involving C2 (coordination and communications). Certain modes of C2 work like the ant-path meta-algorithm - quickly optimising resource and work. One such example is the responsible tagging protocol which aims to coordinate the editorial problems within the article - and so avoid any excess information. Other modes of C2 can significantly increase the workload required to complete some tasks - easily to the point of failure. [6]

### Communication & Communication Failures

In Wiki game theory , communication is defined as "the process of reducing information asymmetries between agents". In most commonly known game theoretic situations (strategic form two player games like the prisoner's dilemma) there is an assumption that agents moves occur simultaneously with no communication or possibility of coordination. In case of a wiki almost all games are played under imperfect information.

• Information is termed as states of the world
• Each player's belief on the state of the world
• Each player's belief of other players beliefs on the state of the world
• And so on until what is called common knowledge (every one having and knowing about a state of the world).

### communication failure

• In many scenarios agents provide inadequate information resulting a communication failure.
• Some editors will use a tag content as problematic (in their point of view) but many tags cannot be resolved without the omitted context. However the original action may simply be disingenuous and not a good faith attempt to improve the content.
• In the above scenario the event could provoke an edit war aimed at discrediting the main editor resulting in blocks and article protection. These result in significant added cost of editing. (User Space Replication & Committee discussion & Approval by an Admin & Time delay for each edit).
• communication failure can more commonly arise through the complexity and disjoint location of policy. Even when it does not contradict itself Policy without precedent and a consistent and normative enforcement leaves a significant information asymmetry in favor of experienced users. A typical scenario is one where a new comer is provided a communication message that is full of unlinked encoded policy references.
• Mixed signal - by sending an ambivalent or a mixed signal a communicator curtails the avenues of unambiguous response available in the next turn.
• Insult - Wikiquette violations will as a rule result in the users message being ignored. In mid to upper level forums xDR wikiquette offenses will be initially humored, but as more opinions join the opinion will shift from content/procedural dispute to the Wikiquette violations.
• Legal Threats - Making legal threats is frowned upon though in reality legal action is always a possibility. This means that a legal threat is treated as cheap talk on Wikipedia. The anonymous nature of Wikipedia promotes this attitude - on a social network like Facebook a legal threat has an immediate target and the offence is harder to remove.

### Evolution Common Knowledge

On the other hand perfect rationality in Game Theory also allows to specify what agent know, and what they know that others know and so forth up to a "common knowledge" indicating perfect transparency. It is necessary to align agents' preferences vectors. The cost is that of writing the messages required if they all occur in a single coordination space. However if it is necessary to communicate using many disjoint coordination spaces than messages (as described in the banning of a vandal)

agent 1 bears a cost of creating messages (writing) while agent 2 has a cost reading

### Recommendations

Lists of Agents should delist inactive agents (e.g. admins, check user etc.)

### Simple Ham Spam - Creator facing a New Page Patroller

A new articles has n memes etimated as 5. They are either contributed as ham or spam and they are either accepted as ham or rejected as spam. By introducing Reputation and Face into the payoff the game can be seen to be an assymetrical Prisoner's dillema (iterated or not).

Partroller (accept) Partroller (reject SD) Partroller (reject Afd)
ham ${\displaystyle n-work}$
${\displaystyle n}$
${\displaystyle -work}$
${\displaystyle -work({\frac {1}{10}}+{\frac {1}{\lVert Community\rVert }})}$
${\displaystyle -2work,}$
${\displaystyle -work(2+{\frac {1}{\lVert Community\rVert }})}$
spam ${\displaystyle Ln-{\frac {work}{3}}}$
${\displaystyle -Ln}$
${\displaystyle -{\frac {work}{3}}}$
${\displaystyle -{\frac {work}{10}}}$
${\displaystyle -{\frac {work}{3}}}$
${\displaystyle -2work}$
• The information is perfect so the patroller can identify spam from ham.
• Deleting SD requires a minimal expenditure in communication, coordination or time.
• Afd requires voting, communication and reaching consensus within a week.
• L is the lemon[7] factor and is greater than one.

The ham spam games can be followed by:

• a single round of CSD/AfD game. High (40%) and on optional round of Xfd (resoration request) (5%)low.
• a single round of a Ban Game below. If spam count > 3.
• a single round of chicken - an edit war with another editor. (would be better modeled using a ham/opinion/spam).

However these have complex probabilities of taking place.

#### Empirical questions

1. What is the distribution of memes in a new article. Avarage and Standard diviation.
2. Is Accept/Reject dependent on number of memes.
3. The probabilities of CSD AFD have been studied
4. The probability of Bans are more difficult but can also be estimated by reviewing edit histories and counting events.
5. The death event is of interest as well.

Actual deletion can be more complicated for both sides - as can be seen from the chart (developed Alexandre Passant & Jodi Schneider) below:

this chart suggests that the normal form game occurs with multiple agents and multiple rounds at Afd and CSD.

#### Practical Conerns

Modeling a the playesrs as both a bad faith and a good faith actor simply reflects that each user may have his a conflict of interest with that of wikipedia. This includes pai editing under an alias AKA the invisible pink unicorn. Such an editor can quickly establish a IPU identity with low cost and a minimal reputation to do the spam edits.

Since this is a Prisoners dillema some questions arise.

Both players can may wish to increase their payoffs resulting in a non parto efficent mutual deffection equilibria (spam spam). A social engineer would inquire how a (ham,ham) could be encouraged. This is generaly solved by punishing users who defect.

the IPU is a form of abuse more pravelent amonst advanced users. These experienced users are able to game the system and place articles into wikipedia with minimal work. This has significant value for forign elements who offer significant temptations to have the creator intoroduce spam. Such editors will not longer be impatial when taking the other side as patroller. The risk of spamming is far greater for a patroller since he could loose his privlages. In practice great abuse is tollerated (a moral hazzard)

One case which is not tolerated is and therefore not without risk — once IPUs are exposed users can inccur signficant penalties. and lose their real identity and privileges.

So when they wish to behave badly they will use disposable accounts sometimes refered to as Invisible Pink Unicorns or (IPU). These accounts are also detectable and depending on how badly the IPU is abused - the curtain may be lifted.

#### Recomendation

• Use of stylometric and editing metadata to identify IPUs.
• A better COI resolution - let there be good spam. Users should be allowed to do much broader work within the scope of their main account, and not have to resort to hiding their identities. This means having a more liberal policy and being clearer about COI. However this will not solve all COI problems.
• So users should also have a stylometric print or another method of authentication which they cannot forge. This is a commited stylometric identiy.
• But revealing identity is punished in the long-term — epsecialy when dealing with controversial matters. So these two concerns must also be resolved.

### Spam Ham Game under perfect information

Combine the above in an extensive form.

### Simple Ham Spam - Non-Creator facing a Revent Changes Patroller - Sub Game

• This is a close relative of the ham/spam game. However in this case the edit is a smaller departure from the status quoe and thus percieved as less controversial. An quickest rejection of an edit is the a reversion. It is possible for a deletion prorocol to be initated as well.
• On the other hand users are cessured against making many small edits which increase work load at recent changes.
• generaly under knowledge symmetry (the longer the edit the more likely it is to be reverted). this is due to problems like CPVIO risk (though it can also be) due to other reasons.
• Revisions add new memes or replace memes. edits can be ham or spam.
• ${\displaystyle P_{revert}={\frac {\lVert Memes\rVert }{Meme_{max}}}*reputation}$
Partroller (accept) Partroller (revert)
ham, defend ${\displaystyle n-work}$
${\displaystyle n}$
${\displaystyle -2work-2face}$
${\displaystyle -work({\frac {11}{10}}+{\frac {1}{\lVert Community\rVert }})}$
ham, abandon ${\displaystyle n-work}$
${\displaystyle n}$
${\displaystyle -work-face}$
${\displaystyle -work/10}$
spam, defend ${\displaystyle nL-work/3}$
${\displaystyle -nL}$
${\displaystyle -2work}$
${\displaystyle -work({\frac {11}{10}})}$
spam, abandon ${\displaystyle nL-work/3}$
${\displaystyle -nL}$
${\displaystyle -work/3}$
${\displaystyle -work/10}$
• remember the information is still symmetric so the patroller can identify ham from spam.
• reversion after defending is a double effrontery resulting in a loss of the editor
• this is a sub game which would be followed by another round where the Patroller bans the Editor with probability

${\displaystyle p_{ban}=\sum edits_{ham}-L\sum edits_{spam}}$

• can this become a war of attrition?

### Spam ban sub-game

• Banning Vandals is a complex procedure with a protocol usually requiring about four incidents and interactions and coordination of multiple agensts.
• In this case I'll consider modelling an incident based on the work of prior research [13] [14]
• There are two options - developing an extensive form game for banning and developing a simplified normal form model to use as a sub game in more complex situation.
• The extensive form of this game can shed more light on the costs of coordination and collaboration. As described in the above publication the cost have been reduced over time by improving the technology of banning.
##### A Banning Protocol

This is a banning protocol for an anonymous IP based vandal who does not respond to any communication notices. The protocol is motivated that a vandal could be reasoned and converted to a good faith contributor.

1. Inciting Incidents:
1. First round of a Spam\Ham game with an anonymous spammer interacting with Vandal task force patroller with a Robot\Tool. resulting in a Reversion\Rollback of edit + edit message (Communication)
2. Warning level 1 at user page (communication).
3. Second round of the above Spam\Ham game with an anonymous spammer interacting with Vandal task force patroller with a Robot\Tool. resulting in a Reversion\Rollback of edit
4. Check user page for warning level (co-ordination).
5. Issue warning level +1 at user page (communication).
6. Third round of the above Spam\Ham game with an anonymous spammer interacting with Vandal task force patroller with a Robot\Tool. resulting in a Reversion\Rollback of edit
7. Check user page for warning level (co-ordination).
8. Issue warning level +1 at user page (communication).
9. Check AIV notice board for warning (co-ordination).
10. Issue an AIV board notice including Vandal' UserName + links to Diff messages for the above messages. (communication)
11. First round of the Simple Spam Ban Subgame.
2. Denouement:
1. Investigation of other edits by user and reversion. (note this can be expensive is other edits had procured since then the action cannot be rolled back)

Most blocks can be negotiated to be lifted very quickly, very few blocks are long term. Due to the low cost of new usernames [15]

Discussion similar protocols with AWB developer and operator Marios Magioladitis I was informed that he had publicised wrongfulness banning against his user.

The two situation he expounded on were

1. banning of an inactive user due to a third party importing importing his robot's edits followed by a ban due to an operation of a bot without a bot rights.
2. This contravenes policy forbidding banning of inactive users.

## Edit wars (wars of attrition)

Some debates are inherently contentious nature leading to edit wars and total war

Due to misaligned incentives, conflicted interests and misguided sense of ownership online communities are plagued by aggressive behavior. On content management systems (e.g. wikis) these manifests as edit wars, online and offline harassment, legal threats. The primary Wikipedia policy involves consensus, however, when a highly Machiavellian editor identifies (say via signaling) that he is faced by a less experienced editor, the high Mach will engage by making a power play.

Power plays are aggressive moves, often within the gray area of acceptable behavior which can have a number of outcomes:

• Grieving the "opponent" who will stop a Socratic sensation of flow associated with knowledge creation on a wiki. This is typically due to a need to coordinate on a talk page with an uncooperating editor.
• Confound an opponent who accepting the aggressor is i.e. right will move on.
• Intimidate a lower mach opponent whose investment in their contribution is a low stake one.
• Enrage an opponent who, responding in spades could easily respond far beyond they gray are of the acceptable and place himself in a situation where any request for intervention may go against him.
• Precipitate an intervention from an admin who will usually pass for the above outcomes but who may censure the aggressor.
• Be foiled by a sophisticated opponent. (c.f. Slim. v.s. Guzman on paid editing)...

Both due to a lax approach enforcement policy highly Machiavelian users who enjoy and thrive in conflict may commit a sequence of aggressive moves.

Based on [12][16]

• an edit conflict between two users culminating in arbitration may be modeled using a game of wp:chicken.

Two editors are represented:

How can a war of attrition can arise in a world with an objective information function?

• Agents have (resolvable) information asymmetry say a different interpretation of policy - (c.f. should we hyphenated).
• Two agents have a categorical information asymmetry say, atheists v.s. Christians or a b
• Three agents have a non-transitive information asymmetry say three editors playing rock paper scissors.
• A large scale conflict with two or more groups and any number of

### Power plays

Openings or Power Plays (an attempt to claim control over an article) [16]:

• Article scope - Central and peripheral content is strictly delimited by an individual or core group of contributors.
• Prior consensus - Decisions made in the past are presented as absolute and uncontested. (this is a common law model using precedents. It also represents an attempt to reduces coordination costs - however, the claim of uncontested decision is clearly taking thing too far.)
• Power of interpretation - One sub-community commands greater authority than another. (Coalitions will often lead to the formation of one or two elite. These elites can have a valid purpose, but hey do tend to become abusive as they gain power). Many elites have the power to block an action but less power to carry out their wishes.
• Legitimacy of contributor - Traits of a contributor (e.g. expertise) are used to bolster or undermine a position - (This is Wikipedia's reputation mechanism at play - a factor that can lead to cooperation.)
• Threat of sanction - Threatening to use sanctioning mechanisms (e.g. blocking) or to pursue formal arbitration. (This is not all bad since threats of punishment can be a path to cooperation).
• Practice on other pages - Content organization on other articles is used to validate or discredit a revision. (This is a normative argument and may be useful in a TronBot Mediated setting)
• Legitimacy of source - cited source is discredited.

Middle Games:

• Talking to one another - for example to try an NPOV coordination
• Request for Comment - RfC
• Editor Assistance
• Mediation
• Arbitration sub-game
 Swerve Straight Swerve Tie, Tie Lose, Win Straight Win, Lose Crash, Crash Fig. 1: A payoff matrix of Chicken
 Swerve Straight Swerve 0, 0 -1, +1 Straight +1, -1 -10, -10 Fig. 2: Chicken with numerical payoffs
• in their view arbitration has tow functions
• on boarding potentially good contributors and weeding out potentialy bad contributors.
• explain to those editors that survive the process how to coordinate their work.

a extended form of this game may describe more granular arbcom decisions:

• Anti Social (vandals,disruptive,discourteous, make a minority attack,stalkers,harassers,vandals etc.)
• Anti consensus:
• Violations of Editing Policies
• Impersonation
• Contempt
• Article Chaos

arbitration can result in results

## The battle of the sexes - Modeling The Gender Gap -

 Edit ♂ POV Edit ♀ POV Edit ♂ POV 4,1 0,0 Edit ♀ POV 0,0 1,4 Unburned
 Edit ♂ POV Edit ♀ POV Edit ♂ POV 2,1 -2,0 Edit ♀ POV -2,0 -1,4 Burned

While the gender gap could result from ... a battle of the sexes it is not an obvious situation. Anyhow the battle of the sexes is a symmetrical model - unless one side can burn money. In Wikipedia there is an extreme gender gap. A research program might go about researching the effects different factors effect the gender gap might look how different "money burning" strategies effect female v.s. male users. To be clearer money burning could be all sort of behaviours such as deleting 27% of all articles... or having lots of articles about porn.

However this is a method to test threatened sex discrimination. To model direct discrimination is also possible with a simpler game.

(aversion to technology upfront)

## Article Ownership

Does ownership create an entry deterrence for new participants?

Game of Incumbent editor or clique of editors against Entrent editor c.f. w:golden mean

## intro

Derive a psych tests to personalize user engagement experience.

1. MACH test
2. Risk aversion (Hyperbolic v.s. Exponential Discounting Test)
3. Stylometric
4. Social/Game Role Compatibility
5. Learning style

## Original Mach Test

1. Never tell anyone the real reason you did something unless it is useful to do so.
2. The best way to handle people is to tell them what they want to hear.
3. One should take action only when sure it is morally right. [note 7]
4. Most people are basically good and kind. [note 7]
5. It is safest to assume that all people have a vicious streak and it will come out when they are given a chance.
6. Honesty is the best policy in all cases. [note 7]
7. There is no excuse for lying to someone else. [note 7]
8. It is hard to get ahead without cutting corners here and there.
9. All in all, it is better to be humble and honest than important and dishonest.[note 7]
10. When you ask someone to do something for you, it is best to give the real reasons for wanting it rather than giving reasons that carry more weight. [note 7]
11. Most people who get ahead in the world lead clean, moral lives. [note 7]
12. Anyone who completely trusts anyone else is asking for trouble.
13. The biggest difference between most criminals and other people is that criminals are stupid enough to get caught.
14. Most men are brave. [note 7]
15. It is wise to flatter important people.
16. It is possible to be good in all respects.[note 7]
17. Barnum was very wrong when he said that there’s a sucker born every minute.[note 7]
18. Generally speaking, men won’t work hard unless they’re forced to do so.
19. People suffering from incurable diseases should have the choice of being put painlessly to death.
20. Most men forget more easily the death of their father than the loss of their property.

## Modified Mach Test

• Refomulate the questions in the context of English Wikipedia as a conceptual viahcle for creating the self selection signature test.
1. On talk pages it is best to avoid discclosing the real reason you did something unless it is useful to do so.
2. In consensus discussions, the best way to handle people is to tell them what they want to hear.
3. Community members should take action only when sure it is sanctioned by policy. [note 7]
4. Most contributors to Wikimedia projects are basically good and kind. [note 7]
5. It is safest to assume that all editors have a vicious streak and it will come out in a content dispute.
6. Honesty is the best policy in all cases. [note 7]
7. There is no excuse for lying to someone else. [note 7]
8. It is hard to get ahead without cutting corners here and there.
9. All in all, it is better to have have humble and honest than important and dishonest.[note 7]
10. When you ask someone to do something for you, it is best to give the real reasons for wanting it rather than giving reasons that carry more weight. [note 7]
11. Most editors who seek promotion to admin in Wikipedia are significant contributors respecting rules and guidleines. [note 7]
12. Anyone who completely trusts other editors is asking for trouble.
13. The biggest difference between most COI-editors and other people is that COI-editors are stupid enough to get caught.
14. Most ediors are bold. [note 7]
15. It is wise to flatter adminis, chapter and WMF members.
16. It is possible to be good in all respects.[note 7]
17. Barnum was very wrong when he said that there’s a sucker born every minute.[note 7]
18. Generally speaking, editors won't work hard unless they're coerced to do so.
19. People who are incurably disruptive should be painlessly and premenetly banned.
20. Most editors forget more easily the retirement of their mentor than the loss of their articles.

## A Wiki mach test

It looks like using a behavioral signatures to test for this may be more robust and be more subtle to automation than administrating a questionnaire. ideally the test would use self-selection criteria as well as self de-selection criteria to weed in and weed out users. What may be different in the partial test is that some metrics may be stronger than others and a complex formula + correlation might be needed to validate and align the two scales.

• testing Machiavellian tendencies of wikipedians:
1. user uses multiple disclosed account for editing
2. user uses multiple undisclosed account for editing
3. user canvasing
1. user has limited posting for an issues he voted on
2. mass posting for an issue (spamming)
3. user expressed positive/negative sentiment biased (campaigning )
4. user expressed neutral sentiment (neutral )
5. has sought out people based on view-point (non-partisan)
6. has sought out people based on view-point (vote stacking)
7. contacted users openly (open)
8. contacted users off wiki = (stealth)
4. user has been blocked
5. user has been unblocked early
6. user has been banned
7. user has been unbanned early
8. user has a fun club
9. user has changed username more than once
10. user has a semi protected user page
11. user has a semi protected talk page
12. has been subject of check user
13. has edited controversial topics
14. has contributed to good articles
15. forum shopping
16. arbitration
1. won an arbitration
2. lost an arbitration
3. commended by arbitration
17. nomination
1. flags collection
3. beaurowcrat
4. check user
5. oversight
6. steward
18. policy discussion
1. initiated
2. succeeded
19. user's edits deleted
20. reverted/rollback edits
21. reverted/rollback edits non spam (30%+ of content eventual introduced)
22. reversing vote in debates.

Gaming the systems like many pieces of Wikipedia jargon means different things to different people. Gaming The System is about subverting the goals of the project to an individual's own interests by manipulating the rules. In this section, however, the discussion is centered about using game theoretic thinking to understand better understand the interplay of community, content and governance.

1. What are the paradoxes of the Community?
3. Which Wikipedia policies are in contradiction with its other polices?
4. Which Wikipedia practices are in contradiction with its self-interest?
5. What are the main points of view of these people?

• Organic editor loss - The average user has a limited amount of information he is best suited to contribute. Once these are done he or she will leave the project or transition into non-editorial roles.
• Editing Cost - Over time the cost of becoming a new user increases, have increasing difficulties to succeed contributing a single article - however these new users are the project's future.
• Policing Cost - As they gain experience users, in the name of preserve a communal "status quo"[note 8] end up increasingly police each other, rehashing policy, engaging in increasingly political conflicts over personal stakes and end up hazing the weakest victims, the new users.

### The contradiction of Community and Anonymity

• Community means that users can expect to meet each other down the road. It also means that they will have to pay penalties for their actions against the group.
• Anonymity means that users can cheat - and avoid the penalties that would be the part of a known individual.
• Wikipedia lacks and resists policy about qualifications.
• Wikipedia discourages companies to use their employees (AKA COI editors) to contribute content to Wikipedia.
• Wikipedia also does not allow superstars to benefit from their work since it hides the level of contribution of the users.
• Community engagement is the most crucial element for the success of a Goal Centric Projects such as Wikipedia. Other Projects sharing this property are open source projects. These like many virtual communities of yore have implicit social aspects. It is been in the mismanagement of society that such projects have
• Wikipedia matured before social media came of age. This places it into the special category. One of the most interesting aspects of social networks today is that the more an individual partakes of social activities the harder it is to conduct in deceptive activities. A second aspect Which is the reason why it suffers from many social This It has in the days prior to There are many penalties for revealing an individual on-line. As one's roles becomes more central - especially if someone appears to be a superstar it is very likely that this will attract hateful backlash.

### The Dilemma of Elites and their Accountability

• An elite in the sense of a group of the community which exerts a disproportionate influence for which it is unaccountable [17]
• Elites can have both positive and negative effects. But a lack of accountability has a corrupting effect. History also shows how these groups can become radicalized by foreign interests.
• In communities lacking formal structure elites are placed strategically to wield influence far greater than their numbers in the group.
• Elites often coordinate using alternative communication networks to those used by the community. E.g. their own
• It is possible that one elite will be balanced by another. However is this status cannot formalized it is an unstable power structure. Over time and once an elite becomes dominant it will have tactically superiority over a newcomer group.
• The long-term effect is to transform a community into either a tyranny of the masses or a type of ad hoc dictatorship. which in many sense if where things stand in many ways.
• Q. Are there elites in Wikipedia community?
• A. There surely must be - Ramsey Theory assures us such structures exists. A passing knowledge of policy on accountability assures us that such groups exist.
• Q. Do such elites pursue their own agendas such as: influence editorial decisions; policy; rules of the society ?
• A. Since Wikipedia works on consensus rather than majority decision there is a herd mentality in decisions which would

benefit from independent thinking.

• Q. Is it possible to detect an Elite in action?
• A. In discussion members of such an elite will behave different. People in the elite will support each other and avoid contradicting one another. Thus elites can be uncovered by applying sentiment analysis, discourse analysis of other methodologies to their communications and collective action.

• Q. How can an elite's influence be countered.
• A e.g.
1. Identify the elite - anonymity is the elite's greatest assets, private communication circles is the second.
2. Form a watchdog group that enforce accountability. (Could be subverted by the elite)
3. Replace informal with a formal structure that guarantees user rights equally.
4. Create an opposing group to provide an counter-force competition.
5. Divide and Conquer - Needs an opposition that can divide.

### Paradox of Deletion & Notability

Each edit to Wikipedia is done under CC-BY-SA 3.0 License and the GFDL. Which places them into the commons. When an article is deleted the loss is to all Wikipedia users and syndicates. Unless the edit is illeagal (CPVIO), deleting it from public history is in conflict with GPL. There are many reasons supplied for deletion, but outside of breaking a law such as copyright, plagiarism, libel, violations of privacy, or upploading reckless and endangering information there is good room for debate about deleting content.

There debate begins with spam, to inadequate edits and so on.

Tyranny of the commons - free riders against the workers.

Increasingly I take an issue with the statement that someone is an expert in some area. It would be best not to label anyone an expert. In the Black Swan, Nassim Taleb makes a case against most so-called experts. Yet another popular book called w:number crunchers make calls for using data based methods and statistics to replace experts.

• Authority - a person to whose opinion, other people will acknowledge and often defer to. I.E. someone whose influence and knowledge within a given domain exceeds his peers. However it proves little. Hilbert program's dramatic failure.
• Expert - a person whose knowledge in a field allows can actually be demonstrated objectively. For example by consistently making predictions that come true. At one time an expert astronomer could predict a lunar eclipse in a certain place at a certain time in the future. (Today the calculations are easier)

Dilemma - Consensus V.S. Truth:

• Look as policy as a mirror into deeper underlying problems

"a needs for reaching a consensus vs. editorial integrity."

### Governance & Invasion cost

Confirmation Bias?

Most likely nobody is planning to take over Wikipedia. Yet the costs of subverting the organization or the community in some scenarios is significant to asses the stability of the governance. It is quite likely that by now Wikipedia has developed an immune system.

• Assessment of the cost and strategy of subversion of an organization.
• strategic weaknesses

some issues to look at

• consensus - what is it.
• how good is it against manipulation by voters.
• how good is it against divergent candidates.
• cost of coordinating large discussions.
• cost of coordinating coalitions. (elite cliques with hidden communication channels).
• would providing evidence of covert (off wiki) communication be cause for administrative action.

# Memes

• Decline and fall of Virtual Empires/Communities
reasons for thier fall in dual lens of society and meia.
• Authority over Expert.
• Stylometric Editor Fingerprinting.
• Social Media Meme:
• The Long Tail (In sales and information marketing )
• The Wisdom of Crowds
• Amazon's Mechanical Turk
• Optimal Crowd Sourcing [18]
• Product Placement in Social Space
• Social Media Convergence
• Paid Editing in the The Gift Economy.
analyse paid an unpaid editing done side by side.
e.g. paid and unpaid docents
e.g. paid and unpaid developers at wmf.

## MediaWiki Governance Memes

• Leadership roles of The Board, WMF employees & Volunteers, Chapter, devteam, stewards, Bureaucrats, administrators, etc in day to day activities.
• Two Hat Paradox - Leaders needs to operate visibly at a higher moral level showing the community they respect and uphold the policies from which their authority spring. Many people in leadership roles, pretend to wear two hats
• Simple Joe's hat - acting less then stellar and wishing to doge accountability.(e.g. "Administrators are really Janitors" pretending that have no cliques, lack influence, access to better information, or real power and thus addressing their shortcomings).
• Policies in action:
• Real world - Legal stuff like open licenses and charters, nonprofit, tax, accounting etc. (Buffered by WMF and Chapters)
• Formal Hard - Built into the software.
• Formal Soft - Extensions, Bots, Tools, run by the community.
• Semi-formal - Policy enforced by the core community. (undeletion,arbitration).
• Informal - Policy enforced by the core community. (speedy,deletion,status approval,COI).
• Informal - Policy pages - to be followed by editorial community.

## Notability Memes

• Prescriptive vs. Descriptive - Wikipedia
• As society grows more open prescriptive (lexicon/encyclopedias/databases) will be superseded by descriptive ones.
• Webster's 3rd. vs. Wikipedia.

So what do we do when we do want to reject materials? One way to go about it is to request higher standards. Request citations for all fact. Reject unsourced conclusions. Reject low quality citations. Point out low authority of materials. Place things in a hierarchy - Oscars are more important than Cannes which is more important than the Razzies which are more important than the AVN. By having a multidimensional paradigm instead of notability we can better specify what should be rejected out of hand (not notable) and what should be rejected procedural (local editorial issues).

Content NameSpaces - Partition Main space - Make a space for bio of living person. One for obscene material and one for marginal content. By automatically moving articles to these locations we can reduce the cost associated with owning these medias. Making a We could make obscene name space available only to credit card owners. This would knock down the interest to almost nothing. Such built-in penalties would not need so much work to maintain and would generate less contention.

• Draw a spectrum Diagram

• Future events (they may not happen thus are speculative and non notable). Because of their speculative nature they are likely to be in the hype to paid spectrum. While their citations may support speculative statements they are poisoned roots. Even if the event takes and article is changed to reflect said fact the article cannot be said to follow from these facts and using them is an a posteriori fallacy.
• Fads are not notable. The majority of Greek plays written, as well as most early Italian operas composed that were great sensation - made the artist rich and famous stars. But these did not stand the test of time for various reasons. Most Hollywood movies do not become classics - their only chance of making a return is through media saturation. This is a reason why it is better to wait with such subjects. However we are an open society and our encyclopedia is inclusive. We should accept fads for what they are, and once they pass we should take due note of the fact. Fads are indicated by a rapid spike of interest and citations of short life span. The sources should be clustered too.
• Offensive and controversial material like Pornography related Porn Stars 'Porn Movies' and 'Sexual Subjects' are in a different category.
• Theirs is not only a limited audience but their very presence may offend a large class of viewers.
• They make Wikipedia access problematic for children.
• They also likely to be non-encyclopedic (commercial,pr,vanity) agendas behind them.
• They also require resources such as review, patrolling, copyright oversight etc which many would find objectionable. (e.g. large masses of obscene pictures being introduced into commons)
Thus there is a vested interest in reducing their presence downgrading their importance. An effort should be made to send them elsewhere. Note: there are now external wikis which are better suited to hosting such material.
• Obsolete Video Games are also in a much larger category - without the negative

## Inefficiencies

• DMCA protocols vs. Self Policing.
• AstroTurfing

## Non Wiki Games

• The subjects need not be notable or novel, though the work should contribute new information.
• The publication work is developed by one or more individual based on a research.
• The results are submitted for publication at different journals - coordination
• The results are reviewed by anonymous peers
• The paper is improved rechecked
• The paper is accepted or rejected for publication if it is sufficently notable and still novel.

### Peer Review Game

this is a game which uses

this is a game of coordination

### Print Encyclopedia

to make a print encyclopedia the following are needed

1. list of subjects (based on a notability cireria)
2. coordinating article creators and reviewers
3. single or multiple editor collaboration using full document iteration.
4. coordinating article style
5. coordinaing fact checking to fix mistakes + republication.

## Research Tools & Resources

### Ethnographic Research Resources

• paper The making of an imagined community: The press as a mediator in ethnographic research into Assisted Reproductive Technologies (ART) Carles Salazar & Gemma Orobitg

### SNA

Texts:

• Exponential Random Graph Models for Social Networks - Theory, Methods, and Applications - Dean Lusher & Johan Koskinen & Garry Robins
• Models and Methods in Social Network Analysis - Carrington, Scott & Wasserman
• Social Network Analysis Methods and Applications - Stanley Wasserman & Katherine Faust
• Analyzing Social Media Networks With Nodexl - Derek L. Hansen & Ben Shneiderman & Marc A. Smith
• Exploratory Social Network Analysis With Pajek - Wouter De Nooy & Andrej Mrvar & Vladimir Batagelj

FLOSS SNA Packages:

• Introduction to SNA with R
• PNet a program for the simulation and estimation of Exponential Random Graph (p*) Models.
• Siena SNA using dynamics (time development).
• Statnet - tools for the analysis, simulation and visualization of network data.
• Pajek a program for large network analysis.

Other Concepts:

• H-index of trees and graphs.
• Network Motifs and ratios.

## References and notes

1. R. Davis and R. G. Smith. (1983). "Negotiation as a metaphor for distributed problem solving.". Artificial Intelligence, 20(1) 20 (1): 63–109,. Unknown parameter |URI= ignored (help); Unknown parameter |month= ignored (|date= suggested) (help)
2. ant path meta algorithm - inline tags when used correctly provide editors a system of iterative coordinating requirements and changes between domain experts and policy experts using minimal coordination effort.
3. Oz, Ayelet (5/1/2009). "The Hidden Wikipedia: Wikipedia as a System of Acoustic Separation" (PDF). Harvard Law School. p. 80. Retrieved May 14, 2012. Check date values in: |date= (help)
4. Maynard Smith, J. (1974) Theory of games and the evolution of animal contests. Journal of Theoretical Biology 47: 209-221.
5. Bishop, D.T. & Cannings, C. (1978) A generalized war of attrition. Journal of Theoretical Biology 70: 85-124.
6. a b F. B. Viegas, M. Wattenberg, J. Kriss, and F. van Ham. (2007). "Talk before you type: Coordination in Wikipedia." (PDF). HICSS: 78–78.
7. a b Hoffer, George E.; Pratt, Michael D. (1987). "Used vehicles, lemons markets, and Used Car Rules: Some empirical evidence". Journal of Consumer Policy 10 (4): 409–414. doi:10.1007/BF00411482.
8. p.47 Cite error: Invalid <ref> tag; name "Asif Agha" defined multiple times with different content Cite error: Invalid <ref> tag; name "Asif Agha" defined multiple times with different content Cite error: Invalid <ref> tag; name "Asif Agha" defined multiple times with different content Cite error: Invalid <ref> tag; name "Asif Agha" defined multiple times with different content Cite error: Invalid <ref> tag; name "Asif Agha" defined multiple times with different content
9. http://meatballwiki.org/wiki/CategoryRole
10. SOCIAL ROLES IN ELECTRONIC COMMUNITIES Scott A. Golder and Judith Donath Sociable Media Group, MIT Media Laboratory golder@media.mit.edu, judith@media.mit.edu
11. A conceptual and operational definition of'social role'in online community by Gleave, Eric & Welser, HT in System Sciences, 2009
12. a b Hoffman, David A. (2010). "WIKITRUTH THROUGH WIKIORDER" (PDF). Emory Law Journal 59 (2009-17). Unknown parameter |Author1= ignored (|author1= suggested) (help)
13. Ribes, David; Geiger, Stuart (2010). "The Work of Sustaining Order in Wikipedia: The Banning of a Vandal" (PDF). Proceedings of the 2010 ACM conference on Computer supported cooperative work. CSCW '10 ACM conference on Computer supported cooperative work. ACM. pp. p117–126. Retrieved May 10, 2012.
14. Maehre, Jeff (2009). "What it Means to Ban Wikipedia: An Exploration of the Pedagogical Principles at Stake". College Teaching (ACM): p229–236. doi:10.1080/87567550903218711. Unknown parameter |Volume= ignored (|volume= suggested) (help); Unknown parameter |Issue= ignored (|issue= suggested) (help);
15. Friedman, Eric J.; Resnick, Paul (2000). "The Social Cost of Cheap Pseudonyms". Journal of Economics and Management Strategy (ACM) 10: 173–199. Retrieved August 18, 2013. long term bans are not particularly significant since banned users can quickly create new identities and resume business as usual.
16. a b Travis Kriplean; Ivan Beschastnikh; David W. McDonald; Scott A. Golder (2007). "Community, consensus, coercion, control: cs*w or how policy mediates mass participation" (PDF). Proceeding GROUP '07. international ACM conference on Supporting group work. Retrieved May 11, 2012.
17. Freeman, Jo (1970). "The Tyranny of Structurelessness]". Retrieved May 10, 2012.
18. Axelrod, Robert. 2008. “Political Science and Beyond: Presidential Address to the American Political Science Association.” Perspectives on Politics 6 (1): 3–9.
19. Carr, Jeffrey. 2010. Inside CyberWarfare. Sebastopol, CA: O’Reilly Media.
20. Fearon, James D. and David D. Laitin. 1996. “Explaining Interethnic Cooperation.” American Political Science Review 90 (4): 715–35.
21. Hardin, Garrett. 1968. “The Tragedy of the Commons.” Science 162: 1243–1248. _. 1994. “The Tragedy of the Unmanaged Commons.” Trends in Ecolog y & Evolution 9 (5): 199. Mitchell, William. 1988. “Virginia, Rochester, and Bloomington: Twenty-five years of Public Choice and Political Science.” Public Choice 56 (2): 101–119.
22. Putnam, Robert D. 2000. Bowling Alone. New York: Simon and Shuster.
23. Wu, Xu. 2007. Chinese Cyber Nationalism. Lanham MD: Lexington Books.
24. Zagorski, Nick. 2006. “Profile of Elinor Ostrom.” Proceedings of the National Academy of Sciences of the United States of America 103 (51): 19221–23.
25. Joseph Michael Reagle Jr. (2010). "Good Faith Collaboration - The Culture of Wikipedia". Retrieved May 10, 2012.
26. John Broughton (2008). "Wikipedia: The Missing Manual". Retrieved May 10, 2012.
27. Joseph Michael Reagle Jr. (2010). "Critical Point of View - A Wikipedia Reader" (PDF). Retrieved May 10, 2012.
28. Clay Shirky (2010). "A Group Is Its Own Worst Enemy". Retrieved May 10, 2012.
29. Andrew, G. West (2010). "Autonomous Detection of Collaborative Link Spam". Retrieved May 10, 2012.

### notes

1. A caricature for a No Common Sense Policy
2. Self selection is a criteria which works by being to expensive for players to manipulate. I.E. to be selected using such a criteria required more resources then would be gained by acquiring it. e.g. to get a Chinese passport a foreigner must take a written Chinese language exam - which is so complex that it will probably cost more work and effort than would be gained by such a passport
3. a b c d (In case a return to an old consensus is possible this would use the of 1,...,1 in the top right corner.)
4. CSCW - Computer Supported Cooperative Work.
5. look at languages evolution games
6. good faith and bad faith invaders should be considered. Note: bad faith are obviously bad, would too much good faith be counter productive and render the population fragile to subsequent invasion
7. Reverse coded
8. these is a behavior spectrum starting from enforcing real values embodied in the editorial policy like NPOV and Notability, to vitriolic marking of virtual territory....