# CHAOS™ — Playtest Rulebook **Locked prototype rules v0.5 • Todd Jerome Jenkins • October 2, 2026** **In Development • Pacing rules approved and locked for the next prototype. Use with card reference v0.3 (the same 96 card effects as v0.2).** Todd approved this prototype pacing lock on October 2, 2026: first victory check at the end of total turn 15; checks after every subsequent turn; game ends at total turn 30; one Interrupt opportunity per player per ordinary play. This supersedes the ten-round limit and repeated response priority. Counterclockwise dealer rotation, shared victories, changing objectives, and all 96 card effects remain in place. ## 1. The experience Make a plan, disrupt someone else's, and adapt when the objective changes. Choices matter, but a good plan is never a guarantee. Everyone remains in the game until it ends; there is no player elimination. **Players:** 3–8. **Development target:** ages 12+, 15–30 minutes. Duration is a target, not a measured product claim. The tested median of 20 turns estimates 15–20 minutes at an assumed 45–60 seconds per turn. ## 2. Components 96 cards: 30 Action, 20 Chaos, 12 Win Condition, 14 Modifier, 12 Interrupt, and 8 Anchor. Use paper to record the dealer, Chaos Level, round number, total turn count, and turn direction. No dice, app, timer, money, or additional purchasable components are required. The accompanying card text implements these rules. Category names accompany colors so color recognition is never required. ## 3. Objective Win by satisfying an active Win Condition at a victory checkpoint while you are not the current dealer. A condition may describe hand size, hand composition, or another objectively checkable state. The condition itself states the test. At Chaos Levels 1–3, one condition is active. At Levels 4–5, two are active; satisfying **either** is sufficient. This OR rule preserves two routes to victory and avoids requiring incompatible conditions simultaneously. If multiple eligible players satisfy a condition at the same checkpoint, they share the victory. There is no race to announce first. ## 4. Setup 1. Separate the 12 Win Condition cards; shuffle them into a face-down Condition deck. 2. Shuffle the other 84 cards into the Play deck. 3. Randomly choose a dealer by drawing one card each from the shuffled Condition deck. Lowest printed Condition ID becomes dealer; return and reshuffle all drawn cards. 4. The dealer chooses a starting hand size from 3 through 9. Deal that many Play cards to every player. The choice applies only to setup. 5. Reveal the top Condition card as the active condition. Start Chaos at Level 1, round at 1, total turn count at 0, and direction clockwise. Place the dealer marker with the dealer. 6. The player clockwise from the dealer takes the first turn. No victory check occurs during setup. Keep the Condition and Play decks and their discards separate throughout play. ## 5. Dealer’s Curse The current dealer cannot win at a normal victory checkpoint. On their draw step, every non-dealer draws **2** cards; the dealer draws **1**. These totals implement the recovered additional-card curse without adding another separate draw step. A round ends after every player has had one scheduled turn. Turn direction can change the order of players still waiting, but each player gets exactly one scheduled turn per round. Keep a simple taken-turn list if needed. At round end, expire effects marked “this round,” then move the dealer marker one seat counterclockwise, regardless of turn direction. The former dealer draws 2 cards immediately. Dealer succession happens even if the game reversed direction. No victory check occurs during dealer rotation. The former dealer becomes eligible at the next normal checkpoint. A card may transfer the dealer marker only if it explicitly says so; it does not reset the round or create an extra reward. Round-end rotation still happens normally. ## 6. A turn At the start of each scheduled turn, increase the total turn count by 1. This count belongs to the entire table, not each player, and never resets at round end. A pass or canceled play still counts as a turn; Interrupts and repeated Chaos effects do not add turns. **Draw.** Draw the required number of Play cards. Resolve no drawn card automatically unless a future card explicitly creates a draw trigger. **Play.** Play one Action, Chaos, Modifier, or Anchor from your hand. You may pass instead. Interrupts are response cards and cannot be your ordinary play. Declare any required targets and choices before responses. **Resolve.** Open the response window, then resolve the play and any legal responses under section 7. A canceled play still uses your one ordinary play. **Obey Chaos.** Confirm the current Chaos Level and active effects. Timed Modifier effects happen at their printed event, not again at this step. This is not an extra play or resolution. **Checkpoint.** Expire effects marked “this turn.” On total turns 1–14, do not check victory; earlier qualifying states do not reserve a win. At the end of turn 15 and every subsequent turn, check the currently active conditions simultaneously for all eligible players. On turn 30, if no normal victory occurs, perform the fallback in section 14 and end the game immediately. If the game continues, scan from the active player's seat in current direction and pass to the first player who has not acted this round. If everyone has taken a turn, perform round-end rotation and begin the next round with the player one seat in current direction after the new dealer. There is no hand limit. Unwanted cards are part of the game state. Effects allowing discard or transfer provide ways to change that state. ## 7. Responses and priority Place the ordinary play face up in a pending area. Snapshot the current turn direction for this window. Starting with the next player in that direction, make exactly one circuit of the table, ending with the active player. Each player gets one opportunity to play one legal Interrupt or pass. A played Interrupt becomes the newest pending effect. Continue to the next player without restarting the circuit; nobody receives a second opportunity. An Interrupt may respond to another Interrupt only when its printed target allows that. All pending cards remain visible and are no longer in hands. A physical card cannot be played twice in the same response window. After the active player takes the final opportunity, close the window and resolve pending effects newest first. An Interrupt that cancels a pending card removes that card without resolving it. A canceled Interrupt has no effect. Cards are discarded when resolved or canceled, except a resolved persistent card that enters the table. There is **one response window per ordinary play**. Do not open new windows during resolution. Drawn or returned cards cannot restart this window. Repeated Chaos resolution uses the already determined response results and does not create another window. Only the ordinary play and legal Interrupt responses are pending effects. Do not interrupt mandatory draws, dealer rotation, expiration, or victory checks unless a later revision explicitly establishes such a window. This keeps the procedure finite and understandable. ## 8. Card categories and zones **Action:** resolve once, then discard. **Chaos:** resolve once at Levels 1–2, twice at Levels 3–5. Chaos cards introduce disturbances, change Chaos Level, replace conditions, or alter the table. They are immediate cards, not persistent installations. **Win Condition:** lives in the separate Condition deck, active area, or Condition discard. Never enters a player's hand or the Play discard. **Modifier:** persistent table rule. A player may control only one Modifier; playing another discards their previous one. Each card states its scope and duration. No stacking identical named effects. **Interrupt:** response card used only during section 7's window. **Anchor:** persistent personal protection. A player may control only one Anchor; playing another discards their previous one. Each Anchor states exactly which effect it prevents and whether use discards it. An Anchor can never provide blanket immunity to the core turn procedure or to every condition. Persistent cards are not part of a hand. Discarding a persistent card does not resolve it again. Deck/hand transfers cannot move active Conditions or persistent cards unless explicitly permitted. ## 9. Chaos Levels | Level | Chaos-card resolution | Active conditions | Additional rule | |---|---|---|---| | 1 | Once | One | Starting level | | 2 | Once | One | Effects may refer to Level 2 | | 3 | Twice | One | Resolve the full effect twice sequentially | | 4 | Twice | Two | Either condition can win | | 5 | Twice | Two | After each Chaos card fully resolves, replace the older active condition once | Chaos is always an integer from 1 through 5. “Chaos escalates” increases it by 1, capped at 5. “Reduce Chaos” lowers it by 1, floored at 1, unless a card specifies a different amount. Snapshot the level immediately before the ordinary Chaos card resolves. That snapshot determines whether its printed effect repeats and whether the Level 5 replacement occurs. Changes made by its first resolution do not change that card’s repeat count. Each repetition uses the same declared player targets and announced modes; choices of which hand cards to discard or pass are made afresh during each resolution. Social observation results are recorded once and reused. An invalid target is skipped, without selecting a replacement unless printed text allows it. After each effect, update condition capacity. Entering Level 4 reveals a second condition. Falling below Level 4 discards the older condition and keeps the newer. Level 5's additional replacement occurs once after both resolutions, not once per repetition and not after canceled Chaos cards. If a card starts at Level 5 but reduces Chaos below 4, its final automatic replacement replaces the sole active condition. The snapshot still controls its repeat count and automatic replacement. ## 10. Condition changes “Change the condition” replaces the sole condition at Levels 1–3 or the older condition at Levels 4–5. Discard the replaced condition to the Condition discard; reveal a new one. Maintain left-to-right age order in the active area. If the Condition deck is empty, shuffle the Condition discard to rebuild it, leaving active conditions in place. If no replacement is available, keep the condition being replaced. Objective changes never trigger an immediate victory check during unresolved effects. ## 11. Probability and uncertainty The shuffled decks provide the baseline uncertainty. Players know the categories and possible effects but not the next card or opponents' hands. Individual cards may create conditional outcomes with observable triggers. Do not add coin tosses, dice, or improvised percentage rolls. The initial Play deck distribution is: Action 30/84; Chaos 20/84; Modifier 14/84; Interrupt 12/84; Anchor 8/84. Conditions are 12/12 in the separate deck. These are initial single-draw shares, not constant probabilities throughout the game. Rules for timing and resolution remain definite. Uncertain outcomes arise from hidden information, card order, player choices, and precisely defined conditional triggers. Calling outcomes probabilistic does not mean players must guess how rules work. ## 12. Social triggers and disagreements Social-trigger cards must define: whose behavior counts; an observable behavior; the start and end of its window; and a maximum number of triggers. Do not require physical contact, humiliation, personal disclosure, alcohol, or sustained physical performance. For a disputed observable trigger, one objection raises Chaos by 1 **after the current play resolves**, at most once per ordinary play. If the table cannot agree on whether the trigger occurred, treat it as not having occurred and continue. Repeating the objection adds nothing. For uncertainty about printed rule meaning, use section 13. Do not penalize players for asking a rules question. Record the unclear wording for revision instead of turning unresolved intent into an argument. ## 13. Conflicts and edge cases Core protections outrank cards: finite response windows; no replay of a spent card during a window; no elimination; scheduled victory checkpoints; Chaos range 1–5; and the game-end procedure. Card text can change another rule only by naming the change explicitly. Among valid conflicting persistent effects, the most recently installed effect governs the conflicting part. Older effects still apply where compatible. Keep table installation order visible. Specific card wording beats a general rule except the protected procedures above. If wording remains ambiguous, use the interpretation requiring the fewest additional effects; record it and continue consistently for this session. Resolve instructions in printed order. Do as much as possible. Draw fewer cards if fewer exist; discard fewer if you have fewer; an impossible target instruction has no effect. Costs must be paid in full before a card is legal; consequences after a legal play are performed as far as possible. The current inventory has no additional play costs; every printed instruction is an effect. When the Play deck empties, shuffle the Play discard; do not recycle pending, active, persistent, or held cards. If no Play cards are available, draw zero and continue. A future card must not produce extra turns or recursively play cards; those require a separately reviewed timing design. Players cannot create binding promises that override rules. Alliances and bargains are voluntary table interaction. Game materials contain no enforceable contract or external stakes. ## 14. Finishing the game A normal victory ends the game immediately after its checkpoint. Resolve the complete play before that checkpoint; an intermediate apparent win is not enough. At the end of total turn 30, perform the normal victory check first. If nobody wins, remove the Dealer's Curse and award fallback victory to the player or players with the fewest cards in hand, including the dealer. Persistent cards do not count. End immediately, even midway through a round; do not rotate the dealer or award the former-dealer draw first. The 30-turn ceiling applies to every player count. Rounds still govern dealer rotation and round-duration effects, but no longer determine the game limit. Track fallback frequency during tabletop testing; two independent 10,000-game batches of this version produced approximately 28% fallback wins. ## 15. Examples of procedure **Responses:** Leah plays an Action against Clara. Clara plays an Interrupt that cancels that Action. Alma plays an Interrupt legally canceling Clara's Interrupt. After the single response circuit closes, Alma's response cancels Clara's. Leah's Action then resolves. All three cards leave the pending area; Leah does not get another ordinary play. **Double Chaos:** A Chaos card begins resolution at Level 3 and contains one instruction to escalate Chaos. It resolves twice: Level 3 → 4 → 5. A second condition appears upon entering Level 4. The Level 5 additional replacement does not occur because the snapshot was Level 3. **Moving target:** A player satisfies the old condition midway through resolution. A later effect replaces it. At the checkpoint, only the conditions now active are checked. Satisfying the discarded condition does not win. **Dealer eligibility:** The dealer satisfies an active condition at a normal checkpoint but cannot win. At round end the marker moves and the former dealer draws two. They can win at the next normal checkpoint if they still satisfy a condition. The response example can use A5, I1, and I5. The double-Chaos example can use C1: escalation occurs before its draw instruction each time. ## 16. Quick start Separate the 12 Conditions; shuffle the other 84 cards. Randomly choose a dealer; dealer chooses 3–9 cards per starting hand. Reveal one condition and set Chaos to 1. Begin clockwise after the dealer. On your turn: draw 2, or 1 if dealer; play one ordinary card or pass; allow one Interrupt opportunity per player in a single circuit; resolve newest responses first; obey current Chaos rules; expire turn effects; check victory from total turn 15 onward. Dealer cannot win a normal checkpoint. Count every scheduled turn, including passes. Rotate dealer counterclockwise after everybody has taken a turn; former dealer draws two. At Level 3+, Chaos cards resolve twice. At Level 4+, two conditions are active and either can win. At Level 5, a Chaos card starting resolution there also replaces the older condition once afterward. Check victory only at checkpoints. After the normal check on total turn 30, if nobody wins, fewest cards wins with no dealer exclusion. Stop before any dealer rotation or reward draw. ## 17. Playtest procedure Use the paired card reference v0.3; record the versions at session start. Test 3, 5, and 8 players and starting hands of 3, 6, and 9. Record duration, round and turn of finish, active condition or fallback, disputes, empty draws, hand sizes, dealer changes, and objective changes. Record whether a win occurs at the first permitted checkpoint (turn 15), and whether the game ends through an active condition or fallback. Do not change rules halfway through a measured game. Record a consistent ruling, finish the game, then revise between sessions. Begin with facilitated play; follow with groups using only written materials. Ask whether choices mattered, reversals were entertaining, and waiting or bookkeeping became burdensome. Human tests are still required. ## 18. Precise timing and transfers ### Player and card choices Targets and announced modes/categories are fixed when playing. Hand-card selections for discards, gifts, or passes occur when that instruction resolves, so newly drawn cards may be selected. A printed “up to” allows zero; an exact quantity requires as many as available. A category refers to one of the five Play categories. Titles and flavor are not extra rules. Random hand selection: the owner mixes their hand face down and the recipient, or the owner's next neighbor if there is no recipient, takes a card without seeing its face. Hand size is public. To prove a composition condition, reveal the necessary cards without discarding them. ### Simultaneous transfers Set all outgoing cards aside before delivering them together. Received cards cannot be passed onward within the same instruction. A one-card exchange fails if either participant has no card; a whole-hand exchange can exchange an empty hand. Anchors preventing transfers protect outgoing cards, not incoming gifts. In a one-card table pass, a blocked outgoing card stays in its owner's hand; other outgoing cards still reach their named neighbors. No rerouting is performed for single-card passes. For whole-hand exchanges, a protected participant blocks the entire two-player swap. For table-wide whole-hand passes, identify protected players first, omit them, then circulate hands only among unprotected players in the printed direction. With fewer than two unprotected players, no hands move. Identify all triggering protection before discarding those Anchors. Repeated Chaos resolves this procedure again; the first pass's consumed protection is no longer present on the second. An Anchor belonging to the Chaos card's player does not protect that player from their own card. Physical clockwise/counterclockwise transfers are independent of turn direction unless the card names current turn direction. ### Social observation Observation begins when the card is announced and ends when the single circuit of response opportunities closes its response window. Record the result once, then resolve. A comment is voluntary communication about that card; mandatory game declarations and genuine rules questions do not count. Speech, signing, writing, and pointing to a shared note are equally valid. For A28, the selected player can name one category any time during the window; their first answer is final. No answer when the window closes means decline. A27's guessed category is declared on announcement; its random card is selected during resolution. A disputed observed trigger is treated as not occurring. Apply the approved one-point dispute escalation after the ordinary play and its Modifier follow-ups resolve, before the checkpoint. Never escalate twice for one play, or for a rules question. If the ordinary play is canceled, no social effect resolves; an already recorded observation dispute still receives its one escalation. ### Interrupt details Passing uses your only opportunity in this response window. Playing an Interrupt also uses it. Later Interrupts do not return priority to earlier players, even if they target an earlier response. After the active player has responded or passed, close the window. No new cards enter it during resolution. The chain can contain at most one Interrupt per player. If an Interrupt's target is gone by resolution, that entire Interrupt has no effect, including associated draws. Otherwise cancel its target first, then perform its other instructions. I9 checks whether the Action's printed effect can instruct that responding player to draw, given its announced targets; a possible conditional branch qualifies without waiting for the observation window to finish. When a resolving card leaves the pending area, discard it after all its own instructions (and Chaos repeats) finish, but before “after resolution” Modifier triggers. Pending cards cannot be drawn, recycled, discarded from a hand, or retrieved by A15. Successful persistent installations move directly into their active area. Canceled cards go immediately to the appropriate discard. ### Modifiers and Anchors Install a new persistent card before discarding the controller's old one. Successful installation is a single event; canceled installations do not replace anything. No Anchor blocks replacement or expiration of its controller's own cards. Compatible draw adjustments add together, then floor the total at zero. At an event with multiple compatible Modifier triggers, capture the qualifying Modifiers at that event and resolve them oldest installation first. All players affected by one trigger are processed in current direction starting with the active player. A later trigger's optional hand selections use the hand at that time. M14 sees its own installation once; its draw is not another installation. M7/M8 occur after the response window closes but before pending effects resolve. They apply to each participant once, with at most one Interrupt per player in the window; canceled Interrupts still count as played. M5/M6/M9/M10 do not fire for canceled ordinary plays. All round-duration Modifiers expire after the final turn's checkpoint, before dealer rotation or reward draws. Anchors act automatically on one qualifying instruction, then discard. They do not interrupt and cannot protect setup, core draw steps, dealer rotation/reward, Chaos or condition changes, checkpoints, or fallback. An already canceled card never consumes protection. A prevented transfer or discard stops only that instruction unless it is an entire two-player swap. AN7's consolation draw occurs once; its self-discard cannot trigger itself again. ## 19. Quick reference for tricky cases | Situation | Resolution | |---|---| | Chaos starts at 3, rises to 5 | Two effects, no automatic Level 5 replacement | | Chaos starts at 5, falls to 3 | Two effects, then replace the sole condition once | | AN2 blocks first of two Chaos draw instructions | Discard AN2; second draw can affect the owner | | An Interrupt targets a card already canceled | Whole Interrupt does nothing | | C8 swaps the same hands twice | Original hands return unless protection blocked a repetition | | New Modifier canceled | Old Modifier remains | | A15 retrieves a card | Resolving A15 itself is not in the discard yet | | Dealer moved midround | No extra reward or reset; normal round-end movement still applies | | Total turns 1–14 | No victory checks; qualifying states are not banked | | Total turn 15 onward | Check current active conditions after every completed turn | | Total turn 30 | Normal checkpoint first; then fewest-card fallback with dealer eligible; stop before rotation | | Later Interrupt targets an earlier response | Earlier player has no second response opportunity | No webpage or artwork work is part of this version. ## 20. Approval and version record Pacing locked by Todd on October 2, 2026. Supersedes rulebook v0.4 for active prototype play. Card effects and quantities remain unchanged. Simulation evidence: 10,000 turn-15 comparison games plus an independent 10,000-game stress test; median 20 turns in both batches, no detected model failures. This is a prototype rules lock, with future revisions recorded as new versions.