Reactor: a different way to play hanabi 3 to 5 players

Forum rules
Please DO NOT POST BUGS on this forum. Please report (and vote) bugs on : https://boardgamearena.com/#!bugs
Post Reply
_mg6maciej_
Posts: 33
Joined: 07 January 2025, 22:43

Reactor: a different way to play hanabi 3 to 5 players

Post by _mg6maciej_ »

If you like to try new stuff with your old beloved stuff or maybe wondering if there might be a way to play hanabi with more cards playing blind, never losing the first copy of any 2 (or 4k) and very rarely allowing for a discard of first copy of 4...

A surprisingly simple convention system where almost all clues are resolved before the next turn of the clue giver!

* most clues give a safe action (play or discard) to at least two other players (called reactive clues)
* it's always possible to save the whole hand of the next player, i.e. locking them, with a single clue
* unlike BGA or HGroup conventions we discard the newest unclued card if there is nothing else to do (scary, right?)
* good touch is a secondary principle and it's very simple to make someone discard a marked card that they don't know is trash while also making someone else play

If you've read most or some of the text below, thought it seemed interesting and would like to try this thing out in practice, contact me via private message.

Four major types of clues are:
* reactive to make 2 or more players discard or play specific slots (positions) in any combination, the most useful and frequently used
* stable color push (similar to trash push in BGA conv) to ask the next player who doesn't have a known playable to play some card
* stable fill-in to fully reveal a previously marked card is either playable or trash
* stable number save to make the next player without a known playable save their newest card (which is the default discard) or lock their entire hand

Instead of trying to explain the whole ruleset, I feel it's much better to only explain how reactive clues work, because they are the most useful, and show a game explaining the rest with examples.

Reactive number clues make even number of plays, usually 2 plays and 0 discards, but it's possible for a reactive clue to be 0 plays and 2 discards. reactive color clues make odd number of plays, so in 3 player game always 1 play and 1 discard. Cards to be played or discarded are targeted from left. The sum of slots is defined by color or number used in 3 player games:
1 or red -> sum [s]1 or[/s] 6
2 or yellow -> sum 2 or 7
3 or green -> sum 3 or 8
4 or blue -> sum 4 or 9
5 or white -> sum 5 or 10
basically increasing in order they are put in stacks.

For example if a clue 4 is given to your Cathy and Cathy has 2nd left card playable, it means Bob also has to play 2nd left card (2+2=4) and after Bob plays 2nd left, Cathy also knows her 2nd left is playable (4-2=2). If instead Cathy's 5th card is playable, on 4 clue to Cathy Bob plays 4th card (4+5=9) and Cathy knows her 5th card is playable.

********************************************************

In 4 and 5 player games reactive clues change the meaning:

1, 5, red and white -> sum 5, 9 (or 13 only in 5p)
2 and yellow -> sum 2, 6, 10 (or 14)
3 and green -> sum 3, 7, 11 (or 15)
4 and blue -> sum 4, 8, 12 (or 16)

When deciding between 1 and 5 or red and white, it's best to use the one that either gives more information or follows good touch principle.

If playing 5 player game with each player holding only 3 cards, the sums would be slightly different.

********************************************************

Let's observe an example game: https://boardgamearena.com/archive/repl ... s=97166359

My hand: 1m 2g 4w 4m 2m
Turns 1, 2 & 3: reactive clue 2, to make slots 1 & 1 play
My hand: 2y 2g 4w 4m 2m
Turns 4, 5 & 6: reactive clue 3 (value 8), to make slots 5 & 3 play
My hand: 1r 2y 2g 4w 4m
Turns 7, 8 & 9: reactive clue 2, to make slot 1 & 1 play
My hand: 1r 2y 2g 4w 4m

note: on turn 9, e4e2 can (and should) prevent hua's 2k discard, because he cannot see the other copy; this is the only move that I would consider a mistake in this entire game, the game cannot be won if the other 2k is last in the deck (2% chance)

Turn 10: discard leftmost card (default action)
Turns 11, 12 & 13: reactive clue 2 (value 7), to make slots 5 & 2 play
Turns 14, 15: reactive clue red (value 6), to make slot 4 play and slot 2 (leftmost card that is trash, dupe 4b) discard
Turns 16, 17, 18: reactive clue 4, to make slots 3 & 1 play
My hand: 5b 1r 2y 4w 4m
Turn 19: stable number save, protecting 5b from discard and asking to discard the card to the right of it (trash 1r), essentically moving chop from default slot 1 to slot 2
Turns 20, 21 & 22: reactive clue blue (value 4), to make slot 3 discard and slot 1 play
Turns 23 & 24: stable fill-in of 4g to make it play
Turn 25: finally discarding 4b trash from a red clue on turn 14
Turns 26, 27 & 28: reactive clue 1 (value 6), to make slot 1 & 5 play; finesses in reactor are quite rare and bluffs even more so
Turn 29: discard slot 2 (default action), chop was moved to this position on turn 19
My hand: 3b 5b 2y 4w 4m
Turn 30: discard slot 1 (default action)
Turns 31, 32 & 33: reactive blue clue (value 4), to make slot 3 discard and slot 1 play
My hand: 3r 3b 5b 4w 4m
Turn 34, 35 & 36: reactive 5 clue, to make slots 1 & 4 play
My hand: 4w 3b 5b 4w 4m
Turn 37: discard (not globally) known trash over leftmost card; this also tells e4e2 he has 1k in hand and hua should know his now slot 2 is most likely trash, because e4e2 would feel compelled to protect it
Turn 38: discard slot 1 (default action); 4w discard tells me I should have the other copy of it somewhere in hand
My hand: 4k 3b 5b 4w 4m
Turn 39: discard slot 1 (default action)
Turn 40: play 5r from good touch
Turn 41: discard slot 1 (default action)
My hand: 3k 3b 5b 4w 4m
Turns 42, 43 & 44: reactive 1 clue, to make slot 1 & 5 play
My hand: 1w 3k 3b 5b 4w
Turn 45: stable number save
Turns 46 & 47: discarding chops (default action)
My hand: 4r 3k 3b 5b 4w
Turn 48: old version of a stall clue at 8 clues (or when locked), after this game it was decided that empty clue to the next player (hua) should be performed instead, because this clue would normally be reactive to play slot 1 & 3 (finesse)
Turn 49: default discard action
Turn 50: legit stall clue at 8 clues to the next player, fill-in of an unplayable card, info clue
Turn 51: discard slot 1 (default action)
Turn 52: old version of a stall clue at 8 clues
Turn 53: discard chop at slot 2 (default action)
My hand: 2k 4r 3b 5b 4w
Turn 54: at 8 clues or when otherwise locked you are allowed to do stable color push to Cathy, this is one of those rare situations and asks me to play slot 1
Turn 55: when the next player already has a known play, you can clue a stable clue to the player after, in this situation green wraps around and pushes 1k to play
Turns 56 & 57: execution of the two stable play clues
My hand: 2b 4r 3b 5b 4w
Turns 58, 59 & 60: reactive 5 clue, to make slots 1 & 4 play
My hand: 3w 4r 3b 5b 4w
Turn 61: play from good touch and from the fact the other copy of this card was reactively told to discard
Turn 62: play of a fully known playable card
My hand: 3m 3w 4r 3b 4w
Turns 63 & 64: stable color push to make 2w play
Turn 65: discard marked trash; if this card was 5m (and there is not enough negative clues on the card to say it's not 5m), it would have been clued for play much earlier: turn 48 would be a reactive double play clue instead of a stall
Turn 66: play of a fully known playable
Turns 67, 68 & 69: reactive 1 (value 6) clue, to make slots 5 & 1 play
Turn 70: play from good touch
Turns 71 & 72: stable color push to make 5w play
Turn 73: play of a fully known playable card
Turns 74 & beyond: the struggle of too much efficiency, just cluing empty when at 8 clues & discarding to draw the last playable, 5m

The game ended with a loss of only a single first copy of a useful card, the 2k very early in the game, but it could have been easily prevented. It also doesn't show how locks look like (clue number to the rightmost unclued card).

If you have made this far and watched the entire game, congratulations! Now you should be ready to play your first reactor game, so don't hesitate to PM me if you'd like to try. And if you have any questions, just ask here.

Documentation of this convention system is available at: https://hanabi.wiki/conventions/reactor-1, it looks lengthy mostly because it covers exceptions and adjustments when playing many different variants on hanab.live.
Post Reply

Return to “Hanabi”