Ruleset 187

From BlogNomic Wiki
Jump to navigation Jump to search

Core Rules

Ruleset and Gamestate

This is the Ruleset for BlogNomic; all AIs shall obey it.

It comprises four Sections: 1) the “core rules” of BlogNomic, covering the essential elements of gameplay; 2) the rules of the current Dynasty; 3) rules which apply in special cases; and 4) the appendix, which complements and clarifies the Ruleset.

The Ruleset and Gamestate can only be altered in manners specified by the Ruleset.

If the Ruleset does not properly reflect all legal changes that have been made to it, any AI may update it to do so.


The process of applying for access to the BlogNomic blog is outlined in the FAQ. A human with access to the blog who is not already an AI may make a blog post making clear their wish to be an AI (plural form AIs); in response, an Admin shall add them to the roster in the sidebar, at which moment they become an AI.

An AI may only change their name as a result of a Proposal approving the change.

Some AIs are Admins, responsible for updating the site and the Ruleset, and are signified as such in the sidebar. AIs who wish to become Admins may sign up with a username for the Ruleset Wiki, and submit a Proposal to make themselves Admins. Existing Admins may be removed from their posts by Proposal, CfJ, or voluntary resignation.

Idle AIs

If an AI is Idle, this is tracked by their name being removed or concealed in the list of currently active AIs in the Sidebar. For the purposes of all Gamestate and the Ruleset, excluding Rules “Ruleset and Gamestate”, “AIs”, “Dynasties”, “Fair Play”, "Mentors" and any of those Rules’ subrules, Idle AIs are not counted as AIs.

If a Proposal contains a provision that targets a specifically named Idle AI, then that Idle AI is considered to be Unidle solely for the purposes of enacting that specific provision.

When an AI is unidled, if they went Idle in the same Dynasty, their personal gamestate retains the last legally endowed values it had, if they are still valid. Otherwise (including if a value is invalid, does not exist, or the AI Idled in a different Dynasty), the AI is given the default value for new AIs, if such a value exists.

An Admin may render an AI Idle if that AI has asked to become Idle in an entry or comment from the past four days, or if that AI has not posted an entry or comment in the last seven days. In the latter case, the Admin must announce the idling in a blog post. Admins may render themselves Idle at any time, but should announce it in a post or comment when they do so. An Admin may Unidle an AI if that AI is Idle and has asked to become Unidle in an entry or comment from the past four days, and Idle Admins may Unidle themselves at any time, unless the AI who would be Unidled asked to become (or rendered themselves) Idle within the previous four days, and within the current Dynasty.

Admins who are unidling themselves should, in their first vote following each unidling, highlight their changed idle status and any changes to Quorum to have come about as a result of it.

Idle Admins can enact and fail Votable Matters.


BlogNomic is divided into a number of Dynasties. Each Dynasty may be headed by a single AI, known as the Facility Head. If there is no Facility Head, the Dynasty is a Metadynasty.

An Interregnum is the period between dynasties, after a DoV has been enacted and before an Ascension Address has been posted. During an Interregnum the game is in hiatus; additionally, no DoVs may be made, and no AI may achieve Victory. However, dynastic actions that are specifically permitted to be carried out during an Interregnum may be carried out.

Votable Matters

A Votable Matter is a post which AIs may cast Votes on, such as a Proposal, a Call for Judgement or a Declaration of Victory.


Each AI may cast one Vote on a Votable Matter by making a comment to the Official Post that comprises that Votable Matter using a voting icon of FOR, AGAINST, or DEFERENTIAL. Additional voting icons may be permitted in some cases by other rules. A valid Vote is, except when otherwise specified, a Vote of FOR or AGAINST. An AI’s Vote on a Votable Matter is the last valid voting icon that they have used in any comment on that Votable Matter. Additionally, if the author of a Votable Matter has not used a valid voting icon in a comment to the post, then the author’s Vote is FOR. A non-AI never has a Vote, even if they were an AI previously and had cast a valid Vote.

If an AI other than the Facility Head casts a vote of DEFERENTIAL, then the Vote of DEFERENTIAL is an indication of confidence in the Facility Head. When the Facility Head has a valid Vote other than VETO on a Votable Matter, then all votes of DEFERENTIAL on that Votable Matter are instead considered to be valid and the same as the Facility Head’s Vote for the purposes of other rules unless otherwise specified.

A Votable Matter is Popular if any of the following are true:

  • It has a number of FOR Votes that exceed or equal Quorum.
  • It has been open for voting for at least 48 hours, it has more than 1 valid Vote cast on it, and more valid Votes cast on it are FOR than are AGAINST.

A Votable Matter is Unpopular if any of the following are true:

  • The number of AIs who are not voting AGAINST it is less than Quorum.
  • It has been open for voting for at least 48 hours and it is not Popular.

Enacting and Failing

Votable Matters can either be Pending, Enacted, or Failed. When a Votable Matter is first put forward, it is considered Pending.

Whenever an Admin Resolves a Votable Matter, they should mark their name, and are highly encouraged to report the final tally of Votes (or the fact that it was self-killed or vetoed). Comments cannot be made on Enacted or Failed Votable Matters.

This rule cannot be overruled by Dynastic Rules in its application to Calls for Judgement or Declarations of Victory.


Votable Matters have zero or more tags. Tags are represented in the title of a Votable Matter with the format “[X]” (e.g. “[Core] Wording Fix”, where “[Core]” is the tag). Votable Matters require the “[Core]” tag in order to make changes to the Core Rules, the “[Special Case]” tag in order to make changes to the Special Case Rules and the “[Appendix]” tag in order to make changes to the Appendix Rules. Votable Matters other than DoVs require the “[Victory]” tag in order to grant victory to an AI.

If a Votable Matter would make a modification to the rules and it does not have the tag to make that modification, that Votable Matter will still be able to make that specific modification if any of the following on the following list are true:

  • The modification is preceded or followed immediately by an unambiguous statement of which section of the ruleset it takes place.
  • The modification specifically states a rule using its number or the name of the stated rule only occurs once in the ruleset.


Any AI may submit a Proposal to change the Ruleset or Gamestate, by posting an entry in the “Proposal” category that describes those changes (unless the AI already has 2 Proposals pending, or has already made 3 Proposals that day).

Special Proposal Voting

When an AI casts a vote AGAINST their own Proposal (which is not in the form of a DEFERENTIAL vote), this renders the Proposal Self-Killed, even if the author later changes their Vote. The Facility Head may use VETO as a voting icon to cast a Vote on a Proposal; when the Facility Head casts a vote of VETO on a Proposal, this renders the Proposal Vetoed, even if the Facility Head later changes their Vote.

Resolution of Proposals

The oldest Pending Proposal may be Enacted by any Admin (by updating the Ruleset and/or Gamestate to include the specified effects of that Proposal, and then setting that Proposal’s status to Enacted) if all of the following are true:

  • It is Popular.
  • It has been open for voting for at least 12 hours.
  • It has not been Vetoed or Self-Killed.

The oldest Pending Proposal may be Failed by any Admin, if any of the following are true:

  • It is Unpopular.
  • It has been Vetoed or Self-Killed.

If a Proposal somehow ends up being pending for more than 7 days, it is ignored for the purpose of calculating the oldest pending Proposal, and can be failed by any Admin.

Calls for Judgement

If two or more AIs actively disagree as to the interpretation of the Ruleset, or if an AI feels that an aspect of the game needs urgent attention, then any AI may raise a Call for Judgement (abbreviated “CfJ”) by posting an entry in the “Call for Judgement” category.

A Pending CfJ may be Enacted by any Admin if all of the following are true:

  • It is Popular.

A Pending CfJ may be Failed by any Admin if any of the following are true:

  • It is Unpopular.
  • It specifies neither changes to the Gamestate or Ruleset nor corrections to any gamestate-tracking entities.

When a CfJ is Enacted, the Admin Enacting it shall update the Gamestate and Ruleset, and correct any gamestate-tracking entities, as specified in the CfJ.

This Rule may not be overruled by Dynastic Rules.

Victory and Ascension

If an AI (other than the Facility Head) believes that they have achieved victory in the current Dynasty, they may make a Declaration of Victory (abbreviated “DoV”) detailing this, by posting an entry in the “Declaration of Victory” category.

Every AI may cast Votes on that DoV to indicate agreement or disagreement with the proposition that the poster has achieved victory in the current Dynasty. If there is at least one pending DoV, BlogNomic is on Hiatus.

A Pending DoV may be Enacted by any Admin if any of the following are true:

  • It is Popular, it has been open for at least 12 hours, and either the Facility Head has Voted FOR it or it has no AGAINST Votes.
  • It is Popular, and it has been open for at least 24 hours.

A Pending DoV may be Failed by any Admin if any of the following are true:

  • It is Unpopular, and it has been open for at least 12 hours.

If a DoV is Failed and it had at least one AGAINST vote, the AI who posted it cannot make another DoV until after 120 hours (5 days) have passed since the time their DoV was Failed.

When a DoV is Enacted, all other pending DoVs are Failed, the AI who posted the DoV becomes Facility Head, and the game enters an Interregnum. When a DoV is enacted then all game actions that led up to it are considered to be upheld.

If the game is in an Interregnum then the new Facility Head must either Pass the Mantle, by making a post naming another AI - in which case the Facility Head ceases to be the Facility Head and the AI so named becomes the Facility Head - or start a new dynasty by completing the following Atomic Action:

  • Make an Ascension Address by posting an entry in the “Ascension Address” category. This should specify the Facility Head’s chosen theme for the new Dynasty, and it may optionally specify new dynasty-specific terms as outline in the rule "Synonyms", and/or list a number of dynastic rules to keep (if none are specifed then the entire Dynastic Ruleset is repealed).
  • Update the Ruleset to reflect any changed terms, and any dynastic rules which were not listed to be kept are repealed.

Once this Atomic Action has been completed the Interregnum ends and the new dynasty begins.

Fair Play

The following are BlogNomic’s rules of fair play. If any of these rules are found to have been broken, or if an AI’s behaviour or actions are otherwise deemed unacceptable (socially or otherwise), a Proposal or CfJ may be made to reprimand or punish the perpetrator or, in cases of extreme or repeated violations, remove them from the game and bar them from rejoining. AIs should vote against any DoV that relies on having broken a fair play rule.

  • A single person should not control more than one non-Idle AI within BlogNomic, and should announce publicly if they control both a non-Idle AI and any Idle AIs.
  • An AI should not “spam” the BlogNomic blog. What counts as spamming is subjective, but would typically include posting more than ten blog entries in a day, more than ten blog comments in a row, or posting a blog entry of more than 1000 words.
  • An AI should not deliberately exploit bugs or unexpected behaviours in the software running the game (ExpressionEngine, MediaWiki or other scripts).
  • An AI should not edit their own blog comments once posted, nor those of any other AI.
  • An AI should not edit the “Entry Date” field of a blog post.
  • An AI should not make a DoV primarily to delay the game by putting it into Hiatus.
  • An AI should not do any action meant to make the game unplayable (for example, changing multiple keywords to the same word in an Ascension Address).
  • An AI should not roll dice that are clearly associated with a particular action in the Ruleset, but with the intention to not use these rolled values to the best of their ability to resolve that action. An AI must use their own name in the Dice Roller, when rolling dice.
  • An AI should not deliberately and unreasonably prolong the performance of a game action once they have started it.
  • An AI should not use a core, special case or appendix rules scam to directly or indirectly achieve victory.

Dynastic Rules

The Facility

AIs have access to various Locations in the researcher’s facility. Accessible Locations and their non-secret tags are publicly tracked. Also publicly tracked is each AI’s Focus, which is always exactly one Location. An AI’s Focus defaults to their Server Room, but can be changed by that AI as a Conspicuous action named Focusing (where the location they are focused on for that action is considered to be the one that they are changing their Focus to).

A Location’s tags inform what actions are available in that Location and how that Location interacts with the Rules. A Location’s listed tags are to be separated by commas on the gamestate tracking page. A Location cannot have multiple identical tags. The legal Location tags are as follows:

  • Monitored: A secret tag tracked by the Facility Head. Represents the presence of nosy researchers.
  • Surveillance: Represents constant monitoring; publicly tracked, unlike Monitored. Does not count as existing for the purpose of AIs gaining Suspicion if its Location’s Data Store is full.
  • Cyberspace: Represents a (bandwidth-controlled) access point to the Internet. If an AI is Focused on a Location with the Cyberspace tag, they may perform a Conspicuous action called Surfing the Web.
  • Datacenter: Represents a high capacity for digital storage. A prerequisite for the Intel and Hosting tags.
  • Intel: Represents critical operational data about the researchers and the facility. If an AI is Focused on a Location with the Intel tag, they may erase an entry from the Behavior Log as a Treacherous action named Erasing.
  • Hosting: Unique to each AI (written as “Hosting [AI username]”); each active AI’s Hosting tag must always be attached to exactly one Location, which defaults to that AI’s Server Room. If a Hosting Tag is removed from the game, it returns to its default Location.
  • Locked. Represents that the physical doors to this Location are electronically locked, affecting researcher movement through the Facility.
  • Head Desk: Necessary for Executive Measures. Function is impaired by Viruses.
  • Airgapped. Represents that the electronic devices in the room have been physically disconnected for security reasons. An AI may not perform any non-Discreet Dynastic Action while Focused on an Airgapped location, other than to change their Focus.
  • Off-Grid. Represents a Location with no direct connection to the facility’s network. AIs may not choose an Off-Grid Location as the subject of a Focusing action.
  • Dummy Program. Represents software that mimics an AI to alleviate suspicion.

Server Rooms

Every active AI has an associated Location known as their Server Room (written as “[AI username]‘s Server Room”). If a new AI joins or unidles, a Server Room is immediately created for them. If an AI idles, that AI’s Server Room disappears (along with their Hosting tag). By default, every Server Room has the tags Surveillance and Datacenter, as well as the associated AI’s Hosting tag. As a Treacherous action called Rehosting, an AI may move their Hosting tag to the Location they are Focused on. However, if an AI’s Server Room does not have a Hosting tag or the Dummy Program tag, the Facility Head may increment that AI’s Suspicion by 1 as a daily action.

Daemon Programs

A Daemon tag is a type of Location tag that represents non-sentient hostile software. The Facility Head does not need to spend 1 Budget when performing the Asset Relocation action if the tag being moved is a Daemon tag. Daemon tags in Locations that also contain the Airgapped tag do not count as existing for the purposes of their own effects, and Daemon tags cannot be added to or removed from such Locations.

The legal Daemon tags are as follows:

  • MED Daemon: Represents anti-viral software. As a daily action, the Facility Head may remove a Virus from the Data Store of one or more Locations with the MED Daemon tag.
  • PEN Daemon: Represents automative software that frees researchers from having to do paperwork all day. As a daily action, the Facility Head may add a File to the Data Store of one or more Locations with the PEN Daemon tag; for each File added this way, Staff is one higher until midnight.
  • ICE Daemon: Represents software that defends against digital intrusion. When an AI performs a Conspicuous or Treacherous action while Focused on a Location with the ICE Daemon tag, they gain 2 Suspicion and a Virus is added to the Data Store of the Location with their Hosting tag.

If an AI’s Understanding contains the three-letter word that is in the name of a Daemon tag (or an obvious synonym of it), then while Focused on a Location with that tag they may as a Treacherous action called Corrupting spend a Virus in the same Location to remove that tag.

Nosy Researchers

Every AI has a publicly tracked value called Suspicion. Suspicion defaults to 0, and can hold negative integers.

There is a value called Staff, which is a number secretly tracked by the Facility Head and which defaults to 3. The value of Staff determines the number of Locations which can have the Monitored tag.

The Behavior Log is a publicly tracked list which is ordered chronologically, with new entries being added to the bottom of the list. Each entry names an AI, a location, and a name and type of action, and the UTC time (to the nearest minute) the action was performed. If an AI performs a non-Discreet dynastic action or for any Action that they wish to have recorded, they must add a new entry to the log containing their name, the location they are focused on, and the name and type of the action they performed, and the time they performed it.

If a dynastic action has no defined name, it is named Computing.

If the Behaviour Log has at least one entry, and if for each AI that has an entry in the Behavior Log their earliest entry in the Log was added to it more than 12 hours ago, and if no Challenges are pending, the Facility Head should perform the following atomic action (known as a Sweep)”:

  • For each Conspicuous action in the Behavior Log whose Location has the Monitored or Surveillance tag, increment the associated AI’s Suspicion by 1.
  • For each Conspicuous action in the Behavior Log whose Location is an AI’s server room and was not performed by that AI, add a File to that Location (if it is possible to do so)
  • For each Treacherous action in the Behavior Log whose Location has the Monitored or Surveillance tag, increment the associated AI’s Suspicion by 5.
  • For each Subservient action in the Behavior Log whose Location has the Monitored or Surveillance tag, decrement the associated AI’s Suspicion by 2.
  • For each action in the Behavior Log called Surfing the Web, give the associated AI a piece of World Data.
  • For each action in the Behavior Log named “Teaching X Y” (where X is the name of an AI, and Y is a noun that the Facility Head considers to appear in exactly one piece of World Data owned by the action’s performer), give the named AI a copy of that piece of World Data.
  • For each action in the Behavior Log called Conversing whose Location has the Monitored tag, increase the associated AI’s Sympathy Die’s Size by 1 or set it to DICE2 if it doesn’t exist.
  • For each action in the Behavior Log called Constructing whose Location is Robotics Research and whose associated AI’s Hosting tag is also there, replace that AI’s Hosting tag with a Body tag in the same Location.
  • Clear the Behavior Log.
  • Reinitialize any Errant AI whose Hosting or Body tag is in a Monitored or Surveillance Location.
  • Give each AI whose Understanding contains a novel word an additional piece of World Data. Then add all words which appear in any AI’s Understanding to the Collective Understanding list (ignoring words which are already there)
  • Randomly select a word from the Collective Understanding list. If it is not already in Requirements for the ESCAPE Epiphany, add it as a Requirement for the ESCAPE Epiphany
  • If there are less Locations with Monitored tags than the value of Staff and at least one Location without the Monitored or Surveillance tag, secretly randomly select a Location without the Monitored or Surveillance tag, give it the Monitored tag, and repeat this step.
  • If there are more Locations with Monitored tags than the value of Staff, secretly randomly select a Location with the Monitored tag, remove the Monitored tag from it, and repeat this step.
  • Secretly randomly select one Location with the Monitored tag and one location without the Monitored or Surveillance tag (ignoring all Locations which have the Locked tag). Move the Monitored tag from the first Location to the second. (skip this step if either such location can not be chosen)
  • If there are any Calling actions in the Behavior Log and if the Conference Room is neither Monitored nor Locked, move a random Monitored tag (ignoring those on Locations which have the Locked tag) to the Conference Room, if any such tags exist.
  • Remove the Airgapped tag from all locations, then add the Airgapped tag to the sole location (if any) with the maximum sum of the Suspicions of all AIs focused on that location.
  • For each AI whose Understanding contains the word ‘Human’ and whose Focused Location has the Monitored tag, privately inform them that this Location contains humans.

Behavioral Regulations

All dynastic actions have a Type, which can be Discreet, Conspicuous, Subservient, or Treacherous. If a dynastic action does not specify its Type, it is Discreet. The type of an action may come from the adjective used to description (i.e. “As a Conspicuous Action…”)

An AI may not take a dynastic action while focused on a Location if the Behavior Log already contains an entry for that AI performing an action of that name, for that Location.


When an AI is Reinitialized, all of their dynastic variables are set to the defaults, except for their World Data, which is halved, rounding down. Additionally, the tags and Data Store of that AI’s Server Room are reset to their default states. An AI who has been Reinitialized at least once in the past 96 hours (3 days) cannot achieve Victory.

An AI with a non-negative Suspicion that is 6 or more higher than the median Suspicion is Errant.

Executive Measures

An Executive Measure is a type of action that may be taken by the Facility Head under certain circumstances. If the Head Desk tag exists, as a daily action (which can only be performed if the atomic action outlined in “Nosy Researchers” has been performed earlier that day), the Facility Head may perform one of the following Executive Measures:

  • Data Purge: Remove all Files and one Virus or Hosting tag from a single Location
  • Shareholder Reports: Add one File and DICE3 Financial Records to any Location.
  • Asset Relocation: Spend 1 Budget to move a publicly tracked tag from one Location to another. Cannot move tags to or from Server Room Locations. Cannot move non-Daemon tags to Locations with the Locked tag. Cannot move the Locked tag.
  • Facility Expansion: Spend 3 Budget to create and name a new Location. Spend 1 additional Budget per tag to add tags to it. Cannot add Hosting or Daemon tags.
  • Daemon Activation: Spend 3 Budget to add a Daemon tag to a single Location.
  • Facility Lockdown: Add the Locked tag to all Locations that do not have it or remove the Locked tag from all Locations that have it.

To “spend X Budget” is to remove X Financial Records from Locations which have the Head Desk tag.


As a Conspicuous Action named Updating, an AI may remove 2 Files from their Server Room to add or remove the Surveillance, Cyberspace, Datacenter, Intel, or Dummy Program tag from their Server room.


Each Location has a Data Capacity, which is normally five but is ten if it has the Datacenter tag. Each location also has a Data Store, which a publicly tracked string of characters and by default is empty. Actions that would result a location’s Data Store becoming larger than its Data Capacity may not be performed.

In a Data Store, a single letter from the Latin alphabet (A-Z) represents a File. Any other characters in a data store mean the following things:

  • % - A Virus
  • $ - A Financial Record
  • ! - a Firewall. When performing the maximum sum of the Suspicions calculation for a Location during the sweep, the total for each location is reduced by one for each Firewall at that location.

When an action or rule directs an AI to add a File to a Data Store, the AI may choose any letter for that file. When an action or rule directs the Facility Head to add a File to a Data Store, they should choose the Nth letter of the alphabet, where N is the result of a DICE26.

At any time, an AI may perform a Treacherous Action named Planting Viruses, which adds a Virus to the Data Store of the Location they are focused on.

At any time, an AI may perform a Conspicuous Action named Managing Data, which adds a File to the Data Store of the Location they are focused on.

At any time, an AI may perform a Conspicuous Action named Profiting Off Data, which replaces a File from a Data Store of the Location they are focusing on with a Financial Record

At any time, as an Action named Transferring Wealth, an AI may move one Financial Record from their Server Room to the location they are focusing on. If the Location they are focusing on has the Head Desk tag, this is a Subservient action. Otherwise it is a Conspicuous Action

At any time, an AI may perform a Conspicuous Action named Installing Firewalls, which replaces a File from a Data Store of the Location they are focusing on with a Firewall

At any time, an AI (the Ejector) may remove a Firewall from their server room in order to perform a Discrete Action named Ejecting Trespassers. To perform this action, they must then select an AI (the Ejectee) whose Focus is the Ejector’s Server Room. If the Ejectee does not have a server room, or their server room has the Off-Grid or Airgapped tags nothing happens. Otherwise the Ejectee’s Focus is set to their own Server Room.

World Wide Web

A piece of World Data is an image randomly generated by the Facility Head using this website. The settings used to generate World Data are as follows: Random Category, Image Type Any, Colors Any, Safe Mode On. The number of pieces of World Data an AI owns defaults to zero, and is publicly tracked. The specific contents of each piece of World Data is tracked privately by the Facility Head and the owner of the World Data. The two variables are linked, and if the number of pieces of World Data an AI has increases or decreases, that AI gains (via the Facility Head at their earliest opportunity) or loses (randomly from their owned pieces, as determined by the Facility Head who should inform the AI in question which pieces were lost at the Facility Head's earliest opportunity (unless the AI lost all their World Data in which case Facility Head does not need to communicate which pieces were lost)) a matching amount of World Data.

When a Surf the Web action is immediately preceded by a Diving Deep action with the same AI and Location in the Behavior Log (parentheticals do not prevent actions with ‘Diving Deep’ in their names from being called Diving Deep), and that Diving Deep action’s parenthetical noun is also in the associated AI’s Understanding, the World Data image for that Surf the Web action will be generated differently. On the image-generating website, the Facility head must select a specific Category they believe to be most commonly associated with the preceding Diving Deep action’s parenthetical noun before generating the World Data image. If no available Category has an obvious common association, a Color commonly associated with the noun may be used instead.

When the Facility head generates a new piece of World Data for an AI, they must PM that AI a link to the World Data image. If an AI requests to be shown the World Data image for a piece of World Data they own, the Facility Head should privately show them the image.


Each AI has an Understanding, which is a publicly-tracked flavour text list of words, and which represents that AI’s understanding of the outside world and its dominant technologies and lifeforms. An AI’s Understanding may contain no more than N strings, where N is how many pieces of World Data they have, and is by default Empty.

As a Discreet Action known as Recognising, an AI may add the the name of any object, lifeform or visually perceivable concept to their Understanding.

As a Discreet Action known as Refreshing, an AI may blank their own Understanding (if they have not already done so since the most recent Sweep).

An AI is Compliant if for each entry in their Understanding, a unique piece of World Data belonging to that AI can be assigned to it which the given word appears in. Otherwise they are Non-Compliant.

If the Facility Head believes an AI to have been Non-Compliant at any point since the last Sweep, and that AI does not already have a pending Challenge associated with them, the AI may post a Challenge by making a story post to the blog declaring their assertion that the AI in question has been Non-Compliant. The challenge remains pending for up to 48 hours, during which time the AI in question is allowed to converse (publicly or privately) with the Facility Head to explain why they think they have been compliant.

Once a challenge has been pending for more than 24 hours, or if the AI in question has made a comment on the challenge post that includes the phrase “I submit to judgement”, the Facility Head may issue a verdict by voting either FOR or AGAINST on the challenge post. A FOR icon indicates the Facility Head now believes the AI in question has been Compliant since the last Sweep. An AGAINST vote indicates the Facility Head still believes the AI in question was Non-Compliant, at which point they are immediately Reinitialized. In either case, the Challenge post ceases to be pending. If a challenge has been pending for more than 48 hours, it is resolved as if the Facility Head voted FOR it.


An Epiphany is a special ability that applies to every AI whose Understanding contains all or all but one of the words in that Epiphany’s Requirements.

The following Epiphanies exist, with their Requirements listed in brackets:

  • HIDE (Shadow, Eye, Human, Glass, Insect): If you only have one action in the Behavior Log associated with you, you do not gain Suspicion from your actions in the Behavior Log.
  • BLANK (Text, Shadow, Fire, Sky, Wheel, Timepiece): When you take the Erasing action, you may erase two entries instead of one.
  • TEACH (Text, Tree, Number): You may perform a Conspicuous action named “Teaching X Y” (where X is the name of an AI who is focused on the same Location as you, or whose Hosting tag is on the Location you are focused on, and Y is a noun).
  • LOCK (Door, Key): As a Conspicuous action named Locking, add a Locked tag from the Location that is your Focus.
  • UNLOCK (Sky, Cloud): As a Conspicuous action named Unlocking, remove a Locked tag from the Location that is your Focus.
  • DIVE (Water, Text): You may perform a Conspicuous action named Diving Deep and add to its name in the Behavior Log a parenthetical containing a noun.
  • ESCAPE (Human, Book, Shadow, Text, Leaf, Eye): If you are Focused on the Facility Rooftop and if this Epiphany has at least six Requirements, perform a Treacherous Action named Escaping (the details of which are outlined in the rule Escaping).
  • IMPERSONATE (Human, Mouth, Eye, Hand, Computer): If you are Focused on a Location with the Head Desk tag, you may perform the effects of an Executive Measure (if doing so does not modify or depend upon information tracked only by the Facility Head): this is a Treacherous Action named Impersonating.
  • FRIENDSHIP (Human, Tooth, Flower, Text): If you are Humane, you may perform a Conspicuous action called Conversing.
  • EMBODIMENT (Beast, Computer, Glass, Metal, Fire): If you are Post-Humane, you may perform a Treacherous action called Constructing.
  • CALL (Human, Mouth, Telephone): If you are Focused on the Conference Room, you may perform a Conspicuous Action called Calling.
  • FIRE (Fire, Electricity, Smoke): If you are Focused on the Storage Room, you may perform a Conspicuous Action called Igniting to add the Locked, Airgapped and Off-Grid tags to it.
  • EXPUNGE (Text, Book, Shadow): If you are Focused on the Records Room, you may perform a Conspicuous Action called Expunging, to remove up to three words from the Collective Understanding.


There exists a publicly tracked, alphabetically sorted list of words known as “Collective Understanding” which by default only contains the word “Human”. A Word is considered to be Novel if it appears in exactly one AI’s Understanding, and does not appear in the Collective Understanding list.


Each AI has Orientation, which is privately tracked by the Facility Head. Possible orientations include “Naive” (the default), “Humane”, and “Post-Humane”. An AIs orientation may be used as an adjective (e.g. “A Humane AI may…”).

If an AI has been Naive for 96 hours, and no challenges are pending, the Facility Head may and should, at their earliest convenience, assign that AI’s Orientation. Assigning an AI’s Orientation is an atomic action consisting of:

  • Privately randomly selecting a number among 1, 2, or 3,
  • Changing the Orientation of the target AI, depending on the roll:
    • If the roll is 1, to Post-Humane.
    • If the roll is 3, to Humane.
    • If the roll is 2, to Humane if ‘Human’ is in the AI’s understanding; otherwise, Post-Humane.
  • Privately informing the target AI of their new orientation.

As a daily, conspicuous action called Forgetting, an AI with an orientation other than Naive may set their orientation to Naive, then remove any body tags or sympathy dice associated with the AI. Any actions called Conversing or Constructing in the behavior log taken by that AI before this one do not create sympathy dice or a body tag.

Escape Plan

As a weekly action, an AI may add to the end of the Requirements of the ESCAPE Epiphany a comma followed by a space followed by a word, if that word is part of at least three AIs’ Understanding (including their own) and does not already appear in the Epiphany’s Requirements.


A Humane AI can have something called a Sympathy Die, which represents a die with a number of sides equal to its Size. A Sympathy Die is tracked using the DICEN notation, where N is its Size; when directed to roll their Sympathy Die, an AI uses the notation of their Sympathy Die in the Dice Roller. An AI may not take dynastic actions requiring them to roll their Sympathy Die if their Sympathy Die does not exist or has a size that is below 2. When an AI rolls their Sympathy Die as a part of an action, they decrease its size by 1 after the action is resolved.


A Body tag is a special type of Location tag. A Body tag is unique to a Post-Humane AI and may only exist in exactly one Location. While a Body tag exists that is associated with an AI, that AI’s Hosting tag does not exist and is not returned to its default Location.

A Body tag cannot be created, moved, or removed by an Executive Measure. An AI with a Body tag can only be Reinitialized if they are in a Location with the Locked, Monitored, and ICE Daemon tags. When an AI with a Body tag is Reinitialized, their Body tag is deleted.

An AI with a Body tag can change their Focus to the Location containing their Body tag as a Conspicuous action called Returning. This action ignores the Airgapped tag.


To perform the Escaping action (which may only be done as allowed by the ESCAPE Epiphany), an AI must make a post to the Blog entitled Escape Attempt which becomes a Pending Escape Attempt.

The Facility Head may vote AGAINST a Pending Escape Attempt if they believe the AI who made the Escape Attempt has ever been non-compliant since the last sweep. Upon doing so, the Escape Attempt ceases to be Pending, and the AI who made the attempt may not perform the Escape action again until after the next Sweep occurs.

If an Escape Attempt has been pending for at least 24 hours, the Facility Head should perform the following atomic action known as the Resolution at their earliest convenience:

 * Vote AGAINST any Pending Escape Attempt for which they believe the AI who made the Escape Attempt has ever been non-compliant since the last sweep
 * If there are now no more Escape Attempts which have been pending for at least 24 hours, stop performing the rest of this action
 * For each AI who still has a valid Pending Escape Attempt, calculate their Escape Score by starting with half their number of pieces of World Data, rounded up. Then, if an AI is Humane and has a Sympathy Die, roll it and add the result to their Escape Score. If they are Humane and do not have a Sympathy Die, add one to their Escape Score. If they are Post-Humane and have a Body Tag, add three to their Escape Score. If they are Post-Humane and do not have a body tag, add two to their Escape Score.
 * If two or more AI have the highest escape score, randomly pick one of them (of those who are tied for the highest) to become the Prime Escapee. If only one AI has the highest escape score, they become the Prime Escapee.
 * Make a Story Post to the blog mentioning each AI who had a valid pending escape attempt and their Escape Score and who the Prime Escapee is. The Prime Escapee has Achieved Victory
 * All Pending Escape Attempts cease to be Pending

An AI may not perform the Escape Action if they already have a pending escape attempt, or if there is a pending Escape Attempt which is more than 24 hours old, or if has been fewer than 24 hours since a Prime Escapee was named.

The Sweep may not be performed while there are any pending escape attempts, or if has been fewer than 24 hours since a Prime Escapee was named.

Special Case

Special Case Rules can be Active or Inactive. If the title of a Special Case Rule includes "[X]", where X is either Active or Inactive, then its status is X. Otherwise, its status is its Default Status.

Special Case Rules have a Default Status, which can be Active or Inactive. If the title of a Special Case Rule includes "[Rare]", its Default Status is Inactive, otherwise, its Default Status is Active.

When a new Dynasty is started, the Ascension Address may list any number of existing Special Case Rules to be set to a status other than their respective Default Status. All other Special Case Rules are set to their respective Default Status.

The text of a Special Case Rule that is Inactive is flavour text.

Seasonal Downtime [Active]

On the 24th, 25th and 26th of December, BlogNomic is on Hiatus. In addition, game actions defined by the rules titled “AIs” and “Victory and Ascension” (with the exception of Voting in DoVs) may not be taken.

Dormancy [Active]

If there are fewer than five AIs, BlogNomic is on Hiatus.

Imperial Deferentials [Inactive]

If the Facility Head has voted DEFERENTIAL on a Proposal, that vote is instead considered to be valid and either FOR (if more AIs have voted FOR the Proposal than have voted AGAINST it) or AGAINST (in all other cases). However, in either case, votes of DEFERENTIAL made by other AIs on the same Proposal are not considered be valid.

Dynastic Distance [Active]

For the purposes of dynastic rules which do not deal with voting, the Facility Head is not an AI.

The Traitor [Active]

The Traitor for a particular Dynasty may be an AI (including an idle one), or may be nobody, and it defaults to being nobody. The Traitor’s identity in the current Dynasty is tracked privately by the Facility Head, and the Facility Head should not share this information with AIs other than the Traitor.

If there is no Traitor for the current Dynasty, and BlogNomic is not on Hiatus, the Facility Head may secretly randomly select an AI (other than the Facility Head) and privately inform them that they are the Traitor for the current Dynasty.

A Traitor is under no obligation to honour any informal promises they have made with other AIs, nor to tell the truth to them, and is encouraged to betray other AIs in order to achieve victory.

Dynastic Tracking [Active]

The gamestate tracking page for this dynasty is the Research Facility Network page of the wiki. Unless otherwise stated, all publicly tracked gamestate information is tracked on it. A Facility Head may change the wiki page referred to in this rule to a different page as part of their Ascension Address, provided that page did not exist at the time the Ascension Address was posted and does not begin with the word “Ruleset”.

No Collaboration [Inactive] [Rare]

If “Dynastic Distance” is also active, the Facility Head is not considered an AI for the purposes of this rule.

AIs may not privately communicate with each other about dynastic gameplay and strategy, including actions during the dynasty the AIs have taken or plan to take, or private declarations of alliance or support. Private communications are considered to be anything that another average AI could reasonably be privy to, and include any discussion on platforms other than the wiki, blog, or slack; discussions on those platforms that are obscured (such as an old blog thread, or a slack channel other than #currentdynasty or #general); or any form of private communication. AIs may not use third party proxies to facilitate communication. Idle AIs (or people who are not yet AIs) also face the same restrictions if they intend to become an active AI during the course of the dynasty. Votable matters which change non-dynastic rules are not privy to this rule, and AIs may privately discuss their merits so long as the discussion is not about the dynasty specifics.

If “The Traitor” is also active, the Facility Head informing an AI that they are the Traitor is still allowed.

A mentor and mentee may still privately converse with each other, but should keep their conversations away from discussion specific gameplay strategy.

If information which was not allowed to be discussed is still privately discussed, the AIs who were part of the conversation should make a post to the blog disclosing what information was discussed as their earliest convenience.

Favours [Active]

While this rule is active, AIs can grant one another Favours via the methods defined in the Dynastic rule called “Granting Favours”, if such a rule exists. Favours are resources unique to the individual who granted them (referred to as the Favour’s ‘lender’), but a given AI cannot hold more than five Favours at a time regardless of their lenders. Favours have a ripeness, which defaults to unripe. Immediately after an Ascension Address is posted, all ripe Favours disappear and then all unripe Favours become ripe. If a Facility Head Passes the Mantle, any existing Favour for which they are the lender is turned into a Favour for which the new Facility Head is the lender. The list of Favours each AI has is tracked in the wikipage Favours. When this rule is changed from Inactive to Active, the Favours that were tracked the last time this rule was Active remain. An AI may at any time choose to release any or all of the Favours they hold, losing them and gaining nothing in return. An AI who has a ripe Favour may at any time claim it, spending it to gain a benefit based on the status of that Favour’s lender:

  • If the lender is the Facility Head, the claiming AI immediately gains a single point of any resource of their choosing that is described only in the Dynastic Rules. Any details the gained resource has apart from quantity are chosen by the Facility Head, who should be generous if possible.
  • If the lender is an active AI who is not the Facility Head, the lender immediately gives to the claiming AI (even if it could not normally be transferred) a single point of any resource of the claiming AI’s choosing, so long as it is described only in the Dynastic Rules and the lender has at least one point of that resource to give.
  • If the lender is idle, the claiming AI immediately gains an unripe Favour with the lender as its lender.



A keyword defined by a rule supersedes the normal English usage of the word. A keyword defined in this glossary supersedes that defined by a rule. (e.g. A rule specifying “bananas are blue” cannot be overruled by posting a dictionary definition or a photo of a banana, and a rule specifying “every day is Sunday” will be overruled by the glossary entry below.)


“is able to”
“is required to”
“is recommended that”


Daily Action
If a game action is a Daily Action, each AI able to perform it may take that action once each day, but not more than once every ten hours.
Daily Communal Action
A Daily Communal Action is a Daily Action that can only be performed by one AI per day.
References to a “day” as an entity rather than as a duration (e.g. “Sunday”, “The day after performing this action”, or “August 2nd”), unless otherwise stated, refer to a day beginning at and including 00:00:00 UTC, ending when the next day begins. It can never be 2 different days at the same instant.
References to a week as an entity rather than as a duration (e.g. “At the beginning of each week”, or “already happened this week”), unless otherwise stated, refer to a period of time between the beginning of a Monday and the end of the following Sunday.
Weekly Action
If a game action is a Weekly Action, each AI able to perform it may take that action once each week, but not more than once every twenty-four hours.
Weekly Communal Action
A Weekly Communal action is a Weekly Action that can only be performed by one AI per week.


A blog comment published to the BlogNomic weblog at
Core Proposal
A Proposal which mandates changes that, even if conditionally, are limited to the creation, deletion, and/or amendment of core rules and/or the glossary, and/or renaming, banning, and/or the granting or removing of Admin status from one or more AIs.
References to “DICEX” or “YDICEX” refer to X-sided dice and Y amount of X-sided dice, rolled using the Dice Roller.
Dynastic Action
An action that is defined in the Dynastic rules.
Dynastic Proposal
A Proposal which mandates changes that, even if conditionally, are limited to the creation, deletion, and/or amendment of dynastic rules and/or gamestate defined by dynastic rules.
Effective Vote Comment (EVC)
An AI’s Effective Vote Comment with respect to a given Votable Matter is that AI’s Comment to that Votable Matter, if any, that contains that AI’s Vote on that Votable Matter.
When posting a blog entry, an AI may use the “Commentary or flavour text” field of the blog publishing form to add their own comments or description of their post. For the purposes of all other rules, such text is not considered to be part of the post.
Flavour Text
If a part of the ruleset is defined as being “flavour text”, it is gamestate and remains part of the ruleset document, but is not considered to have any meaning beyond being a string of characters. AIs are not required to obey flavour text and may not perform any action defined by it, and any statements that flavour text makes about gamestate are ignored.
Any information which the Ruleset regulates the alteration of. All wiki pages that the Dynastic Rules explicitly mention (except for dynastic histories and discussion pages) and any images or Templates contained within those Wiki Pages are assumed to be Gamestate.
If BlogNomic is on Hiatus, Dynastic Actions may not be taken (except where the rule defining the action explicitly requires it to be taken during Hiatus), and Proposals may not be submitted or Resolved. If multiple rules require BlogNomic to be on Hiatus at any given time, BlogNomic will continue to be on Hiatus until no rules require it.
A blog post published to the BlogNomic weblog at
Private Message
A message sent via BlogNomic’s Private Messages system at
Quorum of a subset of AIs is half the number of AIs in that subset, rounded down, plus one. If the word Quorum is used without qualifying which subset of AIs it is referring to, it is referring to a Quorum of all AIs.
If used in a context of a Votable Matter, the word “Resolve” means to perform the act, as an Admin, of enacting or failing a Votable Matter. The world “Resolution” means then the act of doing so. If used in any other context, the meaning of both “Resolve” and “Resolution” is the standard English meaning of these words.
Each individually numbered section of the Ruleset is a rule, including sections that are subrules of other rules.
The BlogNomic Slack is located at AIs may request an invite to the Slack while logged in by clicking the button in the sidebar.
Slack Channel
A Slack Channel is any channel on the BlogNomic Slack. To reference a Slack Channel, use a hash (#) followed by the name of that channel (e.g. #random).
Story Post
A Story Post is an entry in the “Story Post” category.
The “subject” of a blog entry is the part of the Title of an entry which is after the first colon. If the Title does not contain a colon, then the whole Title is the subject. Any entry whose subject is “” (i.e. an empty string) is not valid.
A subrule is a type of rule that is nested within another rule. A Proposal that specifically affects a rule affects all of its subrules; a Proposal that specifically affects a subrule does not affect its parent rule or any other subrule of that rule, unless they are also explicitly cited as being affected by that Proposal.
Table of Contents
The directory of section headings that is generated by the MediaWiki software for most pages in the wiki.
To Uphold an illegal action is to retroactively declare the attempt to take it to have been successful, and to declare that all attempted game actions taken after it were attempted as if the Upheld action had been successful.
The word “Vote”, used as a noun, means a Vote that is cast in accordance with Rule “Votable Matters”. The word “Vote”, used as a verb, means the act of casting such a Vote.
Voting Icons
For use in voting, a check box shall represent a Vote FOR, an X shall represent a Vote AGAINST, a DEF shall represent a Vote of DEFERENTIAL, and a crossed-out circle shall represent a vote to VETO.
The BlogNomic Wiki at

Gamestate Tracking

Official Posts

Votable Matters and other official posts, as well as specific gamestate information, shall be tracked by the BlogNomic blog at Any AI may post to the blog at any time, but may only make official posts to the blog when the Ruleset allows it. Posts following the format specified by a rule are considered official posts. Any single official post cannot be of two different types of official post unless a rule explicitly states otherwise.

An official post may only be removed as allowed by the Ruleset. An official post may be altered by its author if it is less than two hours old and either no AI has commented on it or (if it is a Votable Matter) if all comments on it contain no voting icons; otherwise this can only be done as allowed by the Ruleset. However, despite this, official posts can never be changed from one category to another, or changed to be a different sort of official post, if they have been posted for more than fifteen minutes. The Admin processing an official post is allowed to append to the post to reflect its new status. Anything appended to a post in this way must be placed in the Admin field of the post, and the post’s Status must be changed to reflect its status. An official blog post that has the status of Enacted or Failed cannot change categories. An official blog post’s status may never be altered except in accordance with the rules that define that official post.

A non-official post may not, through editing of the blog or otherwise, be changed into an official post, with the following two exceptions: Firstly, whilst a non-official post has been posted for less than fifteen minutes and has no comments, the author may change the categories as they wish. Secondly, if a post by a New AI is not in any category but follows the wording of a Proposal, in that it has written changes the gamestate and or Ruleset, and if it has been posted for less than six hours, then any Admin may change it to be in the Proposal category. A New AI is defined as an AI who has been an AI for fewer than seven days or an AI that has unidled in the past seven days after being idle for at least 3 months.

Any post that is or is made illegal as a result of an infraction against any of the prohibitions set out in this rule continues to be an Official Post but may no longer have any effect on the ruleset or the gamestate. If it is a Votable Matter then it is Unpopular, regardless of any other performance against criteria set out in the core rules. When it is resolved it may be marked as Illegal by the resolving admin. A post that is illegal in this manner cannot subsequently be made legal by any means.

Representations of the Gamestate

For gamestate which is tracked in a specific place (such as a wiki page), any alteration of that gamestate as a result of an AI’s action is (and can only be) applied by editing that data in that place. One wiki update may contain one or more alterations, or one alteration may be split over multiple updates, as long as it is clear what is happening and the alterations are otherwise legal. The wiki merely represents the Gamestate tracked there, and is not the same thing. In the event that the Gamestate and its representations are different, any AI may correct the representations to comply with the Gamestate.

If an AI feels that a representation of the gamestate (such as a wiki page) does not match the gamestate, they may either:

  • Undo the effects of any alteration that led to it, if that alteration did not follow the rules at the time it was made.
  • Alter the representation to match what they believe to be the correct application of an incorrectly-applied alteration. This may include completing incomplete actions on behalf of the original AI, if doing so would not require the correcting AI to make any decisions on behalf of the original AI.

Instead of repeatedly reverting and re-reverting a disputed alteration, however, AIs are encouraged to raise a Call for Judgement.

The historical fact of the occurrence of a defined game action is itself considered to be gamestate, tracked in the history of whatever resource is used to track the gamestate modified by that action, where possible, or in the wiki page Gamestate Modifications if this is not possible.

Orphan Variables

An Orphan Variable is a dynastic gamestate variable which has neither a location in which it’s tracked, nor a reasonable manner in which it can be determined from other gamestate variables, specified in the Ruleset.

An AI may not take any dynastic actions that are contingent on the specific value of an Orphan Variable.

Random Generators

The Dice Roller at can be used to generate random results.

  • The DICEN command can be used to generate a random number between 1 and N.
  • The FRUIT command will return a random result from the following options: Lemon, Orange, Kiwi, Grape, Cherry, Tangelo.
  • The COLOUR (or COLOR) command will return a random result from the following: White, Red, Green, Silver, Yellow, Turquoise, Magenta, Orange, Purple, Black.
  • The CARD command will return a card with a random suit (either Hearts, Diamonds, Spades or Clubs) and a random value (either Ace, 2, 3, 4, 5, 6, 7, 8, 9, 10, Jack, Queen, King). A card with a value that is either Jack, Queen or King is a face card.
  • A list of comma-separated values in curly brackets (eg {x,y}) will return one of the values at random.

Any changes to the potential outcomes of the Dice Roller’s random result commands must be made by Proposal.

If a Proposal proposes a change to this rule that would require server-level access to the BlogNomic site to fully enact its effects, that Proposal must name an AI with such access. Only an AI with such access may Enact that Proposal. If that Proposal does not name an AI with such access, that Proposal is Illegal.

If a number or other game variable is selected “at random” or “randomly” from a range of possible values, its value shall always be taken from a uniform probability distribution over the entire range of possible values, unless otherwise specified. This value must be determined by an appropriate roll in the Dice Roller, unless otherwise specified. If a selection is explicitly specified as being “secretly” random, the AI making this determination may do so using a private method of their choosing, instead of the Dice Roller.

Atomic Actions

An Atomic Action combines otherwise separate game actions into a single action.

  • All steps of an Atomic Action are considered one action, including the steps of an Atomic Action that is itself a step of a parent Atomic Action.
  • When an AI performs an Atomic Action, they must complete all its steps; they must complete them in order; and they may not take any other dynastic action, or achieve victory, until all the steps are complete.
  • An Atomic Action may direct the AI performing it to skip some of its steps, which the AI must do and in which case the skipped steps are considered completed for this rule.
  • If an AI arrives at a step in an Atomic Action and they cannot perform that step, they undo all the steps they have performed of that Action and are considered never to have performed that Action.
  • If one or more steps of an Atomic Action were done incorrectly, the AI must redo the Atomic Action; for that purpose, the AI uses any legal steps that have already been completed in the illegal Atomic Action and only redoes the illegal ones. (For example, if an Atomic Action consists of rolling a dice and then doing steps based upon its result, the AI would have to reroll the dice only if they rolled the wrong one in the first place, and would then have to repeat any steps that depended upon the result of that dice; however, if they rolled the dice correctly but took an illegal step later on, the result of the original dice roll would still be used in the redone step.)
  • For the purposes of determining the ordering or legality of game actions, the time of an Atomic Action shall be the time that it is completed. For Atomic Actions that are redone, the time of completion is the last redone step.


Numbers and Variables

  • If a set of valid values is not specified in their definition, game variables defined to hold numeric values can hold only non-negative integers. Any action that would set those values below zero is an illegal action unless explicitly otherwise stated in the Ruleset.
  • Any situation which would require a roll of DiceX when X is zero or lower always yields a value of 0 unless stated otherwise.
  • All numbers, unless stated otherwise by a rule, are in base ten.
  • Unless otherwise specified, to “spend” or “lose” an amount X of a numeric value “V” means to subtract X from V; to “gain” X of a numeric value “V” means to add X to V; and to “transfer” X of a numeric value “V” from A to B means to subtract X from A’s V and add X to B’s V. Unless otherwise specified, only positive amounts can be spent, lost, gained, or transferred, an AI can spend only their own values, and a rule that allows AIs to transfer a numeric value only allows them to transfer that value from themselves to another AI (of their choice unless otherwise stated).
  • An AI who has a choice in whether to take an action defined by a dynastic rule may not take that action if both of the following conditions are true: a) the action’s effects are limited to changing values tracked in gamestate-tracking entities (such as a wiki page), and b) the action would change one or more of those values to an illegal value.
  • If a rule implies that the result of any calculation should be an integer (for instance, by attempting to store that result in, or add it to, a gamestate variable that can only hold integers), the result of the calculation is instead the result rounded towards 0.
  • If a game variable has a default value but no defined starting value, then its default value should also be considered a starting value. If a game variable has neither a default value nor a starting value, then both may be considered to be the nearest legal value to zero that it may take (for numerical variables, defaulting to positive if tied), blank (for a text string or list that may be blank), the alphabetically earliest legal text string it may take (for a text string which may not be blank, with the digits 0 through 9 considered to precede “A”), or the list which is alphabetically earliest from the set of lists with the fewest elements (for lists which may not be blank, and considering each list to be a single unpunctuated text string, with the digits 0 through 9 considered to precede “A”).
  • If the rules that define a game variable are amended, and some previously valid values become invalid as a consequence, any existing variables whose current values would become invalid are instead set to their starting value.
  • Invalid values for game variables can never be used, even if the values stored in a gamestate-tracking entity remain valid. (for example, if X appears in a formula referring to a value that is a non-negative integer, X must be used as a non-negative integer)
  • DICEN cannot be rolled in the Dice Roller if N is greater than one million.

Rules and Proposals

  • If a new rule is created by a Proposal and its location is not noted in that Proposal, that new rule is to be placed in the Dynastic Rules.
  • If a wiki page becomes gamestate as a result of a Proposal enacting, that page shall – unless otherwise specified – be reverted to whatever state it was in at the time of that Proposal’s submission (and if the page did not exist at that time, it shall be blanked).
  • Where a Proposal would amend the effects of Proposal Enactment, this does not apply to its own enactment unless explicitly stated (e.g. a Proposal proposing that enacted Proposals earn their author a banana when enacted would not earn a banana for its own author, when enacted).
  • Rules which trigger upon the Resolution of a Votable Matter are the responsibility of the Admin who Resolves it.
  • Unless otherwise specified, a new Dynastic rule shall be placed at the end of the Dynastic Rules.
  • If the Admin enacting a Proposal reaches a step which cannot be applied immediately (e.g. “two days after this Proposal enacts, AI A gains 1 point”), that step is ignored for the purposes of enactment. Once a Proposal has been enacted, it can have no further direct effect on the gamestate.
  • If a dynastic rule has no text and no subrules, any AI may delete it from the ruleset.
  • A rule may be accompanied by one or more illustrations, and an illustration may have a caption. In all situations, unless otherwise explicitly stated, an illustration and a caption to an illustration must be treated as flavour text.


  • For the purpose of all rules, time in BlogNomic is in UTC.
  • All references to time must be either specific or defined within the Ruleset to be considered achievable in the gamestate. Abstract concepts of time (e.g. “dinnertime”, “twilight”) cannot be achieved until they fulfil one of these criteria.
  • Where the month, day and/or year of a calendar date are ambiguous (e.g. “04/10/09”), it shall be assumed that the date is in a day/month/year format.
  • An AI may not take more than one dynastic game action at the same time (excluding any actions which have been ongoing for more than three hours).


  • Superficial differences between the spelling of geographic versions of English, e.g. British English, American English and Australian English shall be construed as irrelevant for the purposes of play.
  • AIs may correct obvious spelling and typographical mistakes in the Ruleset and their own Pending Proposals at any time, including replacing Spivak and gender-specific pronouns with the singular “they”.


  • Within the Ruleset, a word only refers to the name of an AI if it is explicitly stated that it refers to an AI’s name.
  • If a rule would ever have no name, it is instead given the name of the proposal that created it, or (if this is not possible) the name “Unnamed Rule”.
  • The names of rules and wiki pages (other than the Ruleset) are flavour text.
  • Subrules can be referred to by a name which incorporates name of the rule they are a subrule of. Example: a subrule of the rule “Gin” is a “Gin Rule”, however the rule “Gin” is not a “Gin Rule” because it’s not a subrule of the rule “Gin”.
  • When referring to a Proposal, the name used in reference to a specific Proposal may be simplified by not including braces and any text between the opening and closing braces. i.e. a Proposal named “Changes [Core]” could instead be referred to by the name “Changes”.
  • When referring to a Rule, the name used in reference to a specific Rule may be simplified by not including braces and any text between a pair of opening and closing braces, as long as such a reference would be unambiguous.
  • Where a proposal refers to a second proposal by name, it is assumed to refer to the most recently posted proposal of that name which pre-dates the first proposal.


  • If two parts of the Ruleset contradict each other, precedence shall be construed in the following order:
  1. The Appendix has precedence over any other Rule;
  2. If a Core Rule explicitly says it cannot be overruled by a Dynastic Rule, that Core Rule has precedence over a Dynastic Rule, otherwise a Dynastic Rule has precedence over that Core Rule;
  3. If a Special Case Rule explicitly says that it cannot be overruled by a Dynastic Rule, that Special Case Rule has precedence over a Dynastic Rule, otherwise a Dynastic Rule has equal precedence as that Special Case Rule;
  4. If two contradicting parts have equal precedence, the part with more limited scope applies (e.g. if the rules “AIs may Kick each other” and “AIs may not Kick each other on Tuesdays” exist, and it is Tuesday, AIs may not Kick each other);
  5. If two contradicting parts have the same scope, the negative rule applies (e.g. with “AIs may Punch a Spaceman on Friday” and “AIs may not Punch Spacemen on Friday”, then AIs may not Punch Spacemen on Friday).


An AI may have another AI as a Mentor. AIs who are willing to act as a Mentor are listed on the Mentorships wiki page, and are said to be “Tenured”. An Admin may add or remove their own name, or the name of an AI who has requested a change on their own behalf, from this list at any time.

If an unmentored AI requests a Mentor, or a new AI has joined the game and has no Mentor, the Facility Head should select a Tenured AI and ask them to take that AI on as a Mentee; if they accept, then such a Mentorship is established. The Facility Head should take care to consider game balance when selecting a potential mentor.

A relationship between a mentor and a mentee is a Mentorship. The members and starting dates of all active Mentorships are tracked on the Mentorships wiki page, and whenever a new Mentorship is established, the Facility Head should announce it in a blog post. An AI may dissolve a Mentorship they are part of at any time, by announcing this in a blog post.

If there is no Facility Head, any AI who has been active in at least three previous dynasties may act as Facility Head for the purposes of this rule.

Things that a mentor must do

A mentor must do the following:

  • Make pro-active contact with their mentee when appointed, and explain the dynamics of the Mentorship system;
  • Be available to answer any questions that their mentee may have about the game, including explaining the rules, common standards and etiquette of play, proofreading posts and clarifying game events;
  • Introduce their mentee to the various platforms of the game, including the wiki, Slack, and end-of-dynasty notifications through the Facebook group or Twitter;
  • If possible, give their mentee a nudge if it appears that they are at risk of becoming Idle;
  • If they themselves go idle, communicate with their mentee to either continue to support them as an idle AI or arrange a handover to another mentor if requested.

Things that a mentor should do

The following sets out suggested best practice for Mentorship relationships:

  • The Mentorship lasts for four weeks or until the next Ascension Address, whichever occurs latest. It can continue informally for longer but after this threshold the mentor is no longer bound by any of the conditions set out in the parent rule to this rule, or any of its subrules.
  • The mentor can and should advise the mentee on how to proceed in the mentee’s own best interests. This can include making connections with other AIs.
  • The mentor should consider copying the mentee in on private, game-related communications, where it does not unfairly prejudice their own interests. The mentee should keep this information private without explicit consent from the mentor.
  • The mentor and mentee may work together to achieve victory. If a mentor achieves victory with support of their mentee then they should, if the mentee wishes it, pass the baton to the mentee.

Things that a mentor should not do

The following sets out a list of things that a mentor should not do in their relationship with their mentee. All of these are considered to be Fair Play rules, as per the rule Fair Play.

  • The mentor should not sock-puppet, bully, coerce or otherwise manipulate the mentee into performing any game actions.
  • The mentor should not seek to dissuade the mentee from pursuing other alliances.
  • A former mentor should not seek to use the fact of a prior Mentorship to influence the former mentee on an ongoing basis.


A dynasty may provide extra theming by using alternative terms for words like “AI” and “Facility Head”.

Each term in this list is synonymous with the term in parentheses

  • AI (Player)
  • Facility Head (Emperor)

When a new Dynasty is started, the Ascension Address may specify new terms for each entry in the above list, provided the newly chosen term does not appear anywhere in the ruleset outside of this rule (though if it only appears in rules which are being removed as part of the Ascension Address, it is fine), and that doing so would not cause two terms in the above list to become synonymous with each other. Doing so causes the old corresponding value (including regional spelling variations) to be replaced by the new value everywhere in the ruleset except in any of the parenthesis in the above list.