- Not to be confused with Reactor 1.0 or Reactor 2.0
If anything in this doc is unclear, please just ask on the Hanabi Central discord or edit it yourself!
ReactorZero prioritizes simplicity and clarity — it's powerful, yet one of the simplest convention systems.
No knowledge of Reactor 1.0, Referential Sieve, or H-group required!
We first discuss 3-player games (including variants) before discussing 4/5-player games and convention proposals/discussions.
ReactorZero has only two types of clues: stable and reactive. Every clue given to the next player is stable.
We name the player taking an action Alice and the subsequent players Bob, Cathy (and Donald/Emily in larger games). All clues from Alice to Bob are stable.
Play reveal: Any stable clue that touches a card and makes it become obviously playable merely tells Bob that he may play that card.
For simplicity, when determining global knowledge, only consider the empathy of cards and the crossed out empathy given by cards on the stacks, cards in the discard pile and cards with full empathy.
Otherwise, the meaning of a stable clue depends on what the clue was:
| Clue type | Meaning |
|---|---|
| Color (touching new cards) | Play leftmost newly-touched card |
| Rank (touching new cards) | Discard the first unclued card to the right of the leftmost newly-touched card |
| Either (touching no new cards) | Play leftmost touched card |
Stalling: If a card that was called to play is not immediately playable, the clue has no meaning beyond the information it provided to the hand.
Finally, rank clues that touch new cards have different meanings in specific cases. In order of priority:
For more on variants, see the variants page.
Every clue not given to the next player is reactive. (All of Alice's clues to anyone not Bob are reactive.)
For a reactive clue, the clue recipient is called the receiver; every player whose turn is after the clue-giver up to the recipient is called a (the) reactor.
The clue gets an action from the reactor followed by an action from the receiver. When Bob takes that action, we say that Bob reacts to the reactive clue.
The type of action (discard or play) in both hands is determined by the type of clue:
| Clue type | Meaning |
|---|---|
| color | odd number of plays from the reactor and the receiver (1 discard + 1 play, or 1 play + 1 discard) |
| rank | even number of plays from the reactor and the receiver (2 plays, or rarely 2 discards) |
The total of a reactive rank clue is the numerical value of the rank clue. (E.g. If the clue was a rank 3 clue, the clue total is 3.)
The total of a reactive color clue depends is the position of the color in the interface. (The leftmost color gives a clue total of 1, the 2nd-leftmost color has clue total 2, et cetera.)
Clue totals (clue values) depend only on the clue itself, not the cards or slots touched.
The actions (play or discard) will come from one of Bob's slots and one of Cathy's slots while respecting the equation
Clue total = Bob's slot + Cathy's slot (modulo 5)
An equivalent way for step 1 and 2: Solve the equation X+4=1 using negative numbers. Thus X=-3. Thus Bob should discard slot -3. Slot 1 is slot 1, slot 0 (going to the left and wrapping around) is slot 5, slot -1 is slot 4, slot -2 is slot 3 and slot -3 is slot 2, so Bob discards slot 2.
An equivalent version that does not involve the modulo operator is:
Bob's slot + Cathy's slot = Clue total (or Clue total + 5)
This table gives Cathy's slot based on the clue value and the slot Bob plays.
| Slot Bob plays | |||||
|---|---|---|---|---|---|
| Clue value | 1 | 2 | 3 | 4 | 5 |
| 1 or 6 | 5 | 4 | 3 | 2 | 1 |
| 2 or 7 | 1 | 5 | 4 | 3 | 2 |
| 3 or 8 | 2 | 1 | 5 | 4 | 3 |
| 4 or 9 | 3 | 2 | 1 | 5 | 4 |
| 5 or 10 | 4 | 3 | 2 | 1 | 5 |
This table gives Bob's slot based on the clue value and the slot Cathy will play.
| Slot Cathy will play | |||||
|---|---|---|---|---|---|
| Clue value | 1 | 2 | 3 | 4 | 5 |
| 1 or 6 | 5 | 4 | 3 | 2 | 1 |
| 2 or 7 | 1 | 5 | 4 | 3 | 2 |
| 3 or 8 | 2 | 1 | 5 | 4 | 3 |
| 4 or 9 | 3 | 2 | 1 | 5 | 4 |
| 5 or 10 | 4 | 3 | 2 | 1 | 5 |
The 5 in "modulo 5" corresponds to the number of cards in each hand.
In order for Bob to know which action he should take, Alice and Bob need to agree on which action Cathy should take. This is achieved by the following targeting priority.
Targeting priority in Cathy's hand:
Obviously playable: Whether a card is obviously playable is determined before the reactive clue is given.
Immediately playable: When Cathy is signaled a play, the targeted card in Cathy's hand must be immediately playable after Bob reacts.
Re-targeting: If the clue does not work (makes an unplayable card play or a critical card discard), the target is moved to the next possible target in the priority list. To compute the next possible target, consider what card would be prioritized if the impossible targets are ignored.
Plays (priority 1, 2, 3):
Finesses (priority 2): Only possible with rank clues as they require two plays.
Dupes (priority 4,5):
Leftmost Trash (priority 4):
Sacrifice (priority 6): To compute the furthest-away from playable, look only at the distance between each potential target and their corresponding stack. If multiple cards are furthest-away and equally far away, choose the card with the highest rank. If some of these have the same rank, choose the leftmost card. Moreover, do not sacrifice/target critical cards!
The Gameplay section applies at all player counts
Which cards are play-signaled?
Play signals are not generally erased if the play-signaled card might have been duped. If it is known that the play-signaled card has been duped, the player marks that card as trash.
Play-signaled cards that were actually duped can be fixed by:
A player without playable cards can discard the card with the most recent discard signal. (If no cards have discard signals, they can discard their leftmost unclued card ("chop").)
After a player plays or discards a card or receives a lock clue, they erase all discard signals in their hand.
A player with multiple trash cards typically discards them from left to right.
Discard of known cards without discard signals do not have any special meaning beyond giving a clue back to the team. They can exceptionally be done to generate a clue.
A card whose empathy contains both playable cards and trash cards is not expected to play on its own (unless absolutely necessary to get maximum score).
The meaning of a clue cannot be overriden by contextual information.
Contextual information is any information that does not come from the empathy of the cards (e.g. pink promise and signals are considered contextual information).
It is useful (but not expected) that the players infer contextual information to make more informed decisions.
It is Alice's job to save Bob's chop. Alice assumes that Bob will discard his chop (if Bob is allowed to do it) even if there is a critical card on Cathy's chop. Indeed, Bob and Cathy might know that a card in Cathy's hand is playable or trash and this is unknown to Alice.
It is permissible for Alice to not save Bob's chop if it is not critical and Alice has reasonable expectation that Bob will clue. It is a gamble for Alice to not save Bob's chop if it is critical and Alice has reasonable expectation that Bob will clue.
At pace 0, everyone must play a card on the last round to achieve the maximum score; at pace x, x players don't have to.
The endgame starts when the pace is strictly less than the number of players.
Positional discards are only possible in endgame and get the last possible player to play the matching slot. A card gotten by positional discard is not considered play-signaled.
A positional discard is not always possible, e.g. if it would look identical to discarding chop. Assume generally that Alice is discarding her leftmost locally known trash over a positional discard. Positional discards are less clear than cluing, prefer giving a clue if possible.
The players should try their best to avoid lines that do not get maximum score.
To that end, Reactive clues should be reinterpreted in an unconventional way such as re-targeting to get the maximum amount of plays in a 4-player game.
Sometimes, playing a card from contextual information only might also be necessary.
The next player must ignore the reaction and proceed as if they were locked for that turn. The entire reaction chain is off and the reactor should not try to correct their mistake by reacting later.
While receivers occasionally delay their actions, only rarely should reactors delay.
The receiver can delay their action by giving a clue or by playing a different card. In this case, the targeted card in the receiver's hand is signaled.
A reactor can delay their action by giving a clue, but in this case, the signal to later players is lost (e.g. Cathy does not try to figure out what was the meaning of the clue when Bob reacts later).
The reactor marks their card and is expected to react to the clue later, but the reactor's card is not considered signaled for targeting purposes.
As a reactor, delaying is less problematic when
Only the most recently given reactive clue, if there are multiple reactive clues given before your turn, is "active".
4p and larger games can be significantly harder than 3p games.
As in a 3-player game, Alice's clues to Bob are stable.
Alice's clues to Cathy are reactive like in a 3-player game. Since each hand contains 4 cards in a 4-player game, the equation involving the clue total is changed to
Clue total = Bob's slot + Cathy's slot (modulo 4)
All clues to Donald are double reactive. They get an action from Bob (the first reactor) followed by an action from Cathy (the second reactor) followed by action from Donald (the receiver).
| Clue type | Meaning |
|---|---|
| color | odd number of plays (3 plays or rarely 1 play) |
| rank | even number of plays (2 plays or very rarely 0 plays) |
The equation is
Clue total = Bob's slot + Cathy's slot + Donald's slot (modulo 4)
The target in Donald's hand fixes targets in other hands to make Donald's finesse work. The connecting target(s) is assumed to be on the latest player(s) if multiple players have the same connecting card. The clue is then reduced to a reactive clue between hands with unassigned targets.
Re-targeting refers to changing the card you target. A play target is a card that would be signaled to play. This includes X-away cards. A discard target is a card that would be signaled to discard.
Do not re-target a play target for another play target to the right in any hand except Cathy's hand. If the clue does not work (makes an unplayable card play or a critical card discard) or if there are not enough plays remaining to make the target play as a finesse, change the targeted card directly to the discard target with highest priority.
Re-target inner reductions first. For example if a triple reactive clue (which does not work) contains a C-E finesse and a B-D finesse, try retargeting on Donald first and then retarget on Emily. In this example, do not retarget on Cathy as it is a dependency of Emily.
A playable different hand-dupe is generally considered playable. The card has the same priority as trash cards if its copy has just been play-signaled in the hand of a later player involved by the clue.
6-player games follow the same guidelines in principle.
Clues to Bob, Cathy and Donald have the same meaning as in a 4-player game.
Discard re-targeting, playable different-hand dupe from 4-player games also apply in larger games.
All clues to Emily are triple reactive. They get an action from Bob (the first reactor) followed by an action from Cathy (the second reactor) followed by action from Donald (the third reactor), followed by an action by Emily (the receiver).
| Clue type | Meaning |
|---|---|
| color | odd number of plays (3 plays or very rarely 1 play) |
| rank | even number of plays (4 plays or rarely 2 plays or almost never 0 plays) |
Clue total = Bob's slot + Cathy's slot + Donald's slot + Emily's slot (modulo 4)
The target in Emily's hand fixes targets in other hands to make Emily's finesse work. The connecting target(s) is assumed to be on the latest player(s) if multiple players have the same connecting card. The clue is then reduced to a reactive clue between hands with unassigned targets.
Level 0: Play a 2-player game with the variant No Variant. Only stable clues are required.
Level 1: Play a 3-player game of No Variant with the simplified targeting priority in Cathy's hand: play leftmost playable > play leftmost one-away > discard leftmost trash or same-hand dupe.
Level 2: Play a 3-player game of No Variant with the standard targeting priority in Cathy's hand.
Level 3: Play a 3-player game of Omnis. Familiarize your self with pink promise for locks and plays and the fact that lock has precedence over playable rank in pinkish variants.
Level 4: Play a 4-player game. Learn the targeting priority in Donald's hand, how to compute finesses dependencies in Bob's hand and Cathy's hand. If Donald has a one-away card, it is usually not possible to make Donald discard. Learn the discard re-targeting convention. Different-hand dupes are much more common in 4-player games.
Level 5: Play a 5-player game. Learn the targeting priority in Emily's hand and continue finding dependent targets for finesses and following the discard re-targeting convention for targets in Donald's and Emily's hand.
Level 6: Learn to recognize the reason why some clues were not given. For example, in some circumstances, if a stable clue instead of a a reactive double play is given, it indicates that Bob is playing the same card as would be signaled in Cathy's hand.
Proposals should prioritize simplicity/clarity over efficiency.
This space is reserved for optional conventions and proposals so they can be referred to before the game.