Hanabi : A beginner's guide to conventions and strategies

Discussions about BGA (all languages)
beroberokun
Posts: 7
Joined: 12 January 2015, 15:35

Hanabi : A beginner's guide to conventions and strategies

Post by beroberokun » 14 May 2015, 23:19

Hanabi : A beginner's guide to conventions and strategies to win all the time (how to keep your friends and make new ones)

Glossary
Chop: the right most card you would likely discard.

0-Always scud over to the right
This is done automatically on game sites like BGA. After a play, the remaining cards will be pushed to the right, and a new card will be placed leftmost in the hand. This card is “fresh” and will often be used in finesse, reverse and bluff as it is often the one every player is waiting for.
If you play the physical game, get used to do that, absolutely! It simplifies the way you play your cards. It is also an essential component of the other conventions described below for, this way, the cards fall right toward the chop where they will be saved (or not).
If you don't, you're a masochist and I respect the fact that you dream those 30 points playing like that :-)

1-Play left. Discard right.
When given a multiple clue, always play the leftmost card! That's the one other players want you to play. After playing that first card, the second is up to your judgement. It will often be clued a second time to indicate when it becomes playable.
You can safely assume that a card recently clued to you is playable, if it's high in you hand (left) and there is at least an unclued card after it (see 6-Saves).

On the opposite, if you choose to discard always do so with the rightmost card that doesn't have any clues yet. Following “scud right” you will notice that useless cards fall right, as everybody is waiting for fresh playable cards.

The psychological effect: Discarding is the best move, not cluing!
Hanabi is a game where players have to trust each other. If nothing is said about your hand, it means you contribute nothing and it will remain like that until you draw a fresher card.
There is a recurrent beginner's mistake: cluing, cluing, cluing without ever discarding. If you trust your teammates any unclued card is just another discard waiting to happen. Discarding recycles your hand and makes the game move faster, while giving a clue token for the team to use.
If your gutt makes it hard for you to discard, think about it this way: You play others' hands for them and they reciprocate. If you discard something important, it's THEIR fault, never yours!

The discard convention.
Always assume a player will discard! Never assume s/he will use a clue token if there are some available. Sometimes the game doesn't give the players any playable cards to clue...unfortunately.

2-Finesse (3+ players).
This is a way to play two cards with only one clue. Those clue tokens are vital to the win so why not saving money!
Finesse is carried out as such: A player will give a clue, one card too far (1->3). Since players can see that there is a gap, beginners will assume a mistake, although it's not. If you trust your teammates, check other hands to see if you see the gap card (here the 2). If you don't see it, YOU have it!
Generally it will be the leftmost card, the freshest, the one everybody has been waiting for.
Or if you have a numeral clue, it could be that clued card.

3-Reverse Finesse (3+ players).
This is the same as finesse but backwards. The player with the gap card plays after the player with the card ending the sequence.
The consenquence of a reverse finesse is that the player that is hinted the clue it will have to discard or clue someone else to “pass” his/her turn.

4-Bluff (3+ players).
This strategy uses the knowledge of finesse. A player will clue a card hinting a finesse to another player. Yet, when the player that recognizes it plays, it's not the card he though he would put in play, but it fill another color as the card was technically playable.
A word of caution with bluff: Someone has just given you a bad clue for the sake of having your teammate play a good card. When it is your turn to play and the card that was hinted with a bluff turns out not to be the same color as the clue given to you, you should prevent yourself from playing it, on a reflex. You have to be aware of the clue given and played card! Following the game closely will prevent an unwanted fuse resulting from your lightheadedness.

5-1's without color clues are always playable.
In the beginning, players will often clue multiple 1's. Yet, beginners will sometimes refrain from playing the others as the “left-right” dictates the second (or third) card is up to their judgement.
To simplify and in order so save clues, players consider that all 1's are playable unless they are clued a color to reveal a mistake or a doublet.

6-Saves.
When an already discarded card is in the chop of a player, it is customary to indicate it to him/her with a number. The player must then understand that because the card is in the chop, it's a save and it should be kept for later.
In some situations, saving a card will result in cluing a card higher in the hand. Always consider a save before playing left-right.
!!!Sometimes there will be two indentical cards in two players chops. At least one must be saved but it will look like it's a playable card because its counterpart is not yet in the discard. When clued a number on the chop always check others' hands!!!

7-The next color clue is in sequence.
If you've just noticed a reverse finesse but the sequence has not ended yet, you might be surprised that a continuation of this reverse is played on you...reversed. The gap 2 has been played but another player gives you a color clue before the end of that sequence (You are seated before the player who has the 3). You could assume that you also have the second 3 of that color and treat it as a discard (or worse, play it). But you'd be wrong. This color clue indicates you have the one following the 3 in that color (thus the 4).
Following the minimalist perception of these conventions, nobody will give you a discard clue on something that will be done so naturally.

8-Color is better than number unless...
Giving a color clue is quite straight forward. You should play that card as it often is the next in sequence. But sometimes (especially in harder versions) people will prefer giving you a number.
This number is a clue by itself! It tells you that the higher card(s) might be the same color and it was impossible to give you a color clue. Thus you can start speculating on the left part of your hand.

Strategy overall: keep it minimal.
Players should always try to clue as to never give an unneeded or useless clue. There is no reason to give four 4's to a player when it's the beginning of the game (unless you want a good laugh). Moreover, once clued, a card won't be discardable until told so (using a precious clue token).
Also it moves the chop up, clogging the hand eventually.
This is also a reason why some players don't receive any clues. We like them too, but their hand is rubbish and it should be recycled accordingly. Remember that discarding is never a bad play!

Special strategies:

Fuses are ok if they allow to save something.
Fuses are never fun to get. A card is sent to the discard without the benefit of getting a new clue token (and it angries your teammates).
The only situation for which everybody will be happy that you'd play a fuse is when there are no other ways to clue a card that should be saved for later. Count it as a half-clue for what you play in the discard has one of the qualities (number or color) of the card that is now safe in hand.
Three fuses is plenty. To give the game a fighting chance, try playing without them (one is too many).

Strategy for opening 1's.
Since you should always assume a player will discard, giving the 1's to a player that has a 1 in his/her chop is a good way to avoid waiting for another 1. Thus this players becomes a primary cluing target if no other 1's of the same color are in hands.

End game unconventional discard
A discarded card that doesn't follow conventions can mean the place of a playable card in another player's hand if there aren't sufficent clues to indicate it.

Middle game unconventional discard
A player discarding a visibly playable card, could do so to indicates another payer s/he has the exact same card. This is a strategy to avoid keeping clued cards in hands that will serve no purpose afterwards.

Well I hope you have as much fun as all of us playing hanabi.
This game is unique. It has something of poker, counting cards, yet it's communal. It looks like child play but it requires a learning process and some analytical logic. It's fun with friends but unforgiving to people who don't follow. It's always the same until that deck decides to show you otherwise. It's sometimes frustrating, but who could be mad with so many colors. It's unexpensive, yet you know deep in your heart you would pay 50$ for it (which you can). Most of all it makes me want to trust you all, a human quality of the highest regard.

Congrats to Antoine Bauza for his simple yet powerful realization. I hope to see you around the tables ;-)

Zamiel
Posts: 9
Joined: 06 September 2015, 21:55

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Zamiel » 23 July 2017, 09:17

My group has been playing Hanabi for over a year now and we have developed 9 first principles, as well as a bunch of advanced strategies.
I've documented it all here: https://docs.google.com/document/d/1sl_ ... V0jyKYVnXk

This is not a document for beginners. It's more of a document for intermediate to advanced players in order to take their game to the next level. So it doesn't exactly fit the purpose of the thread in providing general newbie advice, so for that I apologize!

To my knowledge, no other group has created a formal document like this that attempts to enumerate every move and strategy. Its up to 39 pages now! I figured I would leave it here in case other players or groups want to cherry-pick any of our strategies. (Of course, I'm sure a lot of strategies are also common between all of the groups.)

If you love Hanabi, my group is always looking for more teammates to play with, so feel free to join our Discord server (chat room) and say hi: https://discord.gg/FADvkJp
Last edited by Zamiel on 24 July 2017, 21:15, edited 3 times in total.

User avatar
RicardoRix
Posts: 487
Joined: 29 April 2012, 23:43

Re: Hanabi : A beginner's guide to conventions and strategies

Post by RicardoRix » 23 July 2017, 12:38

You know all the fun of this game is figuring out all this stuff on your own and with your own playgroup.

Rather than providing a guide for beginners, you've actually just ruined the fun for beginners.

Ankeszu
Posts: 66
Joined: 06 January 2014, 15:29

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Ankeszu » 23 July 2017, 14:27

@Zamiel,
Hello,
I've been reading your Hanabi Strategies document and I admit it's pretty well done. I like the way you name some strategies, I would never figure to call these things "Sarcastic Discard", "Gentleman's Discard" etc. xD I don't agree on some strategies (as too far-fetched logic), but this system surely looks internally consistent.
However I have few issues. The little ones about wording I suggested to change in the file. And there are few bigger ones, let me tell about them here. I know that many groups play differently and please don't take it as an attack, it's an invitation for discussion.

In all cases below, I refer to players as A,B,C,D (in the given order of the play) and to colors as r,g,b,y,w,m (red, green, blue, yellow, white, multi).
First Principles #2 - Good Touch Principle
Safe discards should not be clued unless there's an important reason to.
Very intriguing statement. Obviously true. However, what is important enough to clue it? Statistically speaking*, most games of experienced players don't reach 30 because of ... losing the first copy of some 3 or 4. In many cases losing that card may be avoided exactly by cluing safe discards (and proper discard management, reading between the lines ~ "I thought he would clue it. He didn't. Hm... He want me to clue it. So he doesn't want me discarding for some time. As the next player has trash chop, I'll make clues instead of him if I can."). In your opinion, is that experience enough reason to avoid discarding cards that doesn't have to be discarded yet - to clue safe discards in hope proper card come that allows to clue chop card to play or along with some other playable?
*Statistics based on my own experience, assuming last 5000 games as games of experienced player.

Intermediate #7 - Avoiding Double Discarding
If the player before you just discarded a card (say the red 4), and you don’t see that card in anyone else’s hand, there is a possibility that you could have that same card on your chop. In this example, if you discarded, both red 4s could be lost and your team would not be able to get a perfect score. This is called double discarding.
Since double discarding can “lose” the game, you should make any decent clue if you can to avoid it. Even a low-value clue might be better than discarding. After a go-around, you can safely discard - if your chop was indeed the same card, it would have been given a save clue by your teammates.
(...)
The plus side of this strategy is that, because players should not generally double discard, you do not need to be overly worried about the same card being on two simultaneous chops.
I am very surprised to see that point. Making a whole "avoid twin discard" convention without mentioning other players responsibility? Except for 2 player games, there is always at least one person to watch other players' chops and make sure they don't discard valuable cards. If anything important is on your chop - it's your teammate's job to tell about it.
Without your convention (ergo: players can discard one after another even if they don't have enough negative info that their card isn't the same), there are many easy ways to avoid twin chop discards, including changing the game flow via "stealing" clues, cluing things sometimes less efficiently, sometimes not cluing things in time, or making tokens to 8 to make discard impossible, etc... And if the plan failed, you can always clue them directly. Instead of checking each time if your chop can be the same as the previous guy, trust your teammates that they would warn you. The player that is given a chop clue (let's say 4 for your 4r example) SEE that it can be the same card as the next/previous player's. He can't know whether it's finesse, play or save, so he has to assume save anyway - and all team, as experienced players, should be aware of possibility of such that situation. In your First-Principles 3&4 we see "When a card that needs to be saved is at immediate risk of being discarded, it must be indicated with a save clue." and "A player receiving a clue should give precedence to a “save” clue interpretation over a “play” clue interpretation.". Leting players discard one after another if no clue is done to stop them allows to have more tokens for later (and, IMO, it's more clear for finesse possibilities - because in your case, a person after useful card discard can't discard, thus any clue from them would be treated less seriously).
(Yeah, I disagree with the Clue Flowchart around "Look at all of the cards in the trash" - I would add "...and other endangered cards: on chop of the player before and after you.")

Advanced #8 - Double Cluing the Same Cards
If you are clued the exact same thing twice in a row, it means that you can play ALL of the cards that the clue applies to, but in the opposite order than normal (right-to-left, since you would normally play cards from left-to-right).
- If one of the cards includes the chop card, it means to play all of the cards from 2nd oldest to newest, and then the chop last.
- If you are clued the exact same thing twice with some delay in the middle, it means that you can play ALL of the cards from left-to-right.
If I understand this correctly, that's very... counterintuitive and I don't really get the situation in which it would be useful, tbh (nor the logic it came from). I mean the varying order. Could you provide any real game example?
(And what means "some delay in the middle"? Like, "twice in a row" means two consecutive players clue you the same, and delay means "A clues you, B does not, C clues you too"? Or what exactly?)

Expert #16 - The Trash Finesse (strong form)
(...) It is possible to double trash finesse; by touching a trash 2 with TWO 2’s remaining, it means that you see BOTH the 2s, so this can cause two people to blind play their newest card (or one person to blind play twice).
Actually, following the logic of the bluff, it shouldn't have to mean that both 2s are playable.

Let's assume that on the table there are 1b 1m 2g 2r 2y 2w. Players don't have any relevant cards marked nor negative information. In their last moves, B discarded; C discarded or played something. Then, A clues single trash 2 on C's newest card*.
B thinks: ok, I have to have 2 on my newest; I play. (Let's assume he played 2b.)
C thinks: B played in response to 2 clue. Obviously, my 2 is not playable - it's discardable. So A's clue let one card being played (2b) and one safe discard being pointed (trash 2). Due to B's response I know my 2 isn't multi without seeing 2m around - so the clue can be done without 2m around...
* - it doesn't even have to be single trash card, it can be three trash 2s - and due to response, obviously none is playable; a clue made one play and three safe discards. Not bad.


(Sorry for my english. First time I tried to write it down the forum erased all things I typed during preview and I am a bit tired of writing it again. I hoped I didn't miss anything while rewriting.)

Liallan
Posts: 1208
Joined: 26 May 2014, 07:01

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Liallan » 24 July 2017, 02:31

RicardoRix wrote:You know all the fun of this game is figuring out all this stuff on your own and with your own playgroup.

Rather than providing a guide for beginners, you've actually just ruined the fun for beginners.
I agree. I think this is very overwhelming for a beginner. I've played twice, making me a beginner, and I have no idea what you're talking about. How about re-name it to "Intermediate player guide for those who want to go further with this game." Even looking at something like this totally turns me off to the game because it makes it sound too complicated.

Every time I see conversations about this game, it's like everyone assumes that everyone else should all be playing that same way. I would find that a little boring. When I played with my group, I realized you have to know your group. If everyone else is not following your way of playing, then it's all sort of out the window. (And I'll never play on here because all the threads about it have totally turned me off.)

If I lose a friend because I couldn't get 30 points in this game, they weren't a friend.

User avatar
turtler7
Posts: 90
Joined: 23 December 2016, 00:01

Re: Hanabi : A beginner's guide to conventions and strategies

Post by turtler7 » 24 July 2017, 05:58

Liallan wrote:
RicardoRix wrote:You know all the fun of this game is figuring out all this stuff on your own and with your own playgroup.

Rather than providing a guide for beginners, you've actually just ruined the fun for beginners.
I agree. I think this is very overwhelming for a beginner. Even looking at something like this totally turns me off to the game because it makes it sound too complicated.

Every time I see conversations about this game, it's like everyone assumes that everyone else should all be playing that same way.
This game was an exceptionally fun start figuring out what makes a good clue, realizing that some discards are important to save while others though not yet played are not important to hold onto (4's), etc. Playing with a group of engineers made this go exceptionally far in the logic train and making it so every clue was trusted to be at least one if not more layers deep.

That said, my group of engineers have found it immeasurably better by using conventions similar to these as a baseline and when a clue is given that doesn't quite seem normal or immediately decipherable we can gather our wits, trust our friends, and make epic plays that include 3-5 cards down from a single clue. The optimization, application of logic to conventions, and decision making process for how to trick the other players into doing what you want when normal clues will backfire or lead to a deadly discard are what really make this game so great. For an audience of engineers.

From my talks with others at gaming clubs, it seems hanabi is a popular game for players who have been drinking and find it funny what crazy clues, plays, and things happen under those cirucmstances. Those type and more casual players will not want to optimize like this.

Zamiel
Posts: 9
Joined: 06 September 2015, 21:55

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Zamiel » 24 July 2017, 08:49

RicardoRix wrote:You know all the fun of this game is figuring out all this stuff on your own and with your own playgroup.

Rather than providing a guide for beginners, you've actually just ruined the fun for beginners.
Hello Ricardo. I'd agree with you in that there is a definitely a "meta-game" of Hanabi - the game of sitting down with your group and compiling a set of really good conventions, through trial and error, over the course of hundreds of games. It certainly takes a long time, and there's certainly some fun in fine-tuning them.

However, claiming that taking this aspect away "ruins all of the fun" may be going a little overboard. Part of the fun of playing Hanabi is also actually playing the game. Coming up with new conventions is fun, but I certainly don't intend to come up with a new convention after every single game. What I do expect is this: each deal is like a new logic puzzle - an optimization problem that I have to solve using complex chains of reasoning. Hyphen-ated, a member of my group, puts it quite eloquently:
Hyphen-ated wrote:Doing complicated chains of reasoning that relies on stuff like "he knows that she knows that I know that he knows this" is something almost completely unique to Hanabi in boardgames, as far as I know. It's like those logic puzzles with the 100 blue-eyed islanders, or the people in a line with colored hats. That's the stuff I really love about the game.
Last edited by Zamiel on 24 July 2017, 08:58, edited 1 time in total.

Zamiel
Posts: 9
Joined: 06 September 2015, 21:55

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Zamiel » 24 July 2017, 08:53

Liallan wrote:I agree. I think this is very overwhelming for a beginner. I've played twice, making me a beginner, and I have no idea what you're talking about. How about re-name it to "Intermediate player guide for those who want to go further with this game." Even looking at something like this totally turns me off to the game because it makes it sound too complicated.
Hello Liallan. That's fair - I've edited my post to say that the guide is explicitly meant for intermediate to advanced players.

Zamiel
Posts: 9
Joined: 06 September 2015, 21:55

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Zamiel » 24 July 2017, 08:57

turtler7 wrote:That said, my group of engineers have found it immeasurably better by using conventions similar to these as a baseline and when a clue is given that doesn't quite seem normal or immediately decipherable we can gather our wits, trust our friends, and make epic plays that include 3-5 cards down from a single clue.
Hi Turtler7. Yup! That is one of the best things about Hanabi. It sure feels great when huge finesses happen! I'd love to play some games with you or your group. Drop in our Discord server and say hi sometime!

Zamiel
Posts: 9
Joined: 06 September 2015, 21:55

Re: Hanabi : A beginner's guide to conventions and strategies

Post by Zamiel » 24 July 2017, 11:36

Ankeszu wrote:However I have few issues. The little ones about wording I suggested to change in the file.
Hello Ankeszu. Great to hear that you found the document interesting. And glad to hear you want to contribute to making it better, so thanks for that. Ideally, I'd ask that you coordinate document updates with me with through PMs on Discord, as using the Google Docs "suggestion" feature is really quite a pain in the butt, so hopefully we can do that going forward. (I also rejected a few of your suggestions, and I'd be happy to go through it line by line on Discord with you if you like.)

Here's all of the discussion points you brought up:



Cluing Safe Discards

This is something we don't do. The reason is that it becomes difficult to distinguish between a clue that means "play this card" and a clue that means "you can safely discard this card". We don't want players having to puzzle over the exact meaning of a clue, because, well, that information is asymmetric. Thus, a player may deduce that another player did something for a completely opposite reason!

We hate ambiguity. We want all clues to have specific meanings. In our framework, (almost) every clue means "play this card" or "save this card". Why is it good to get rid of as much ambiguity as possible? Well, it is extremely valuable to have a strong predictive model of what each player will do on their turn. Since "floating ambiguity" is minimized, it gives each player a Hanabi super power: they can perform high-level multiple-turn-in-the-future predictions to see "lines" of how the game state will be 5-10 turns from now, or sometimes even longer! This is invaluable!

Of course, as you mention:
Ankeszu wrote:most games of experienced players don't reach 30 because of ... losing the first copy of some 3 or 4
Definitely true. So that's why, on top of our existing framework, we've added a bunch of tools to help us save juicy 3's and 4's, sometimes without ever cluing them directly. These are things in the document like the "Bluff" move, the "Chop Move", the "3 Bluffs" convention, the "Tempo-Clue Chop Move", and so forth.

Hopefully that gets at the meat of your inquiry here.



Double Discarding

This really ties in closely to the previous section. Consider that in our framework, (almost) every clue has to be a play clue or a save clue. And you are only allowed to give save clues on 5s, or cards that are already in the trash, or cards that are playable.

Thus, say for example that you see two subsequent yellow 4s on two peoples chops in a row. In your framework, this is not a problem. Just clue one of the 4s! The player will puzzle over it, but after a few turns they will probably be able to look at the history of the game and see why you did it. Problem solved.

But in our framework, this would be quite dangerous! Cluing the 4 would be a signal to play the yellow 4 right now, so either someone else would misplay a card or the person would interpret it as a self-finesse and misplay another card. So this conundrum is the reason why we have a section in the document to cover the "double discarding" situation at all.

So, why do it this way? Well, let's discuss that.

Consider that, in our framework, when you get (what looks like) a save clue on a 4, you assume the following set of possibilities for the card:
1) Every 4 in the trash
2) Every 4 that is playable immediately or through someone else's clued/finessed cards

Essentially, you are posing the question: wouldn't it be better to "tune" the conventions such that we INCREASE the total set of possibilities to be:
1) Every 4 in the trash
2) Every 4 that is playable immediately or through someone else's clued/finessed cards
3) Every 4 that is on the chop of either the player in front of you or the player behind you

On the face of it, this is certainly a reasonable proposal. It does solve the double discard problem in a clean way; players who are in a double discard situation no longer have to (potentially) waste clues to stall.

But of course, it carries with it its own set of downsides. With the lists of possibilities growing larger, each player has more ambiguity about their hand. Each player is incrementally disempowered to perform finesses themselves, since they are heavily disinclined to duplicate cards (because of "Good Touch" principle). Furthermore, normal old play-clues (and self-finesses!) are less likely to work, because it turns clues that would otherwise be valid play clues into possible saves (because we always prioritize save clues over play clues). The lost tempo from this is a fairly big downside.

So, we must balance the benefits with the tradeoffs. We can ask the pertinent question: How often do double discard situations occur in which players must waste a clue? And the answer is that they are quite rare, only happening once or twice a game (and often not at all). So "tuning" for this situation isn't worth it when expanding the set of interpretations for save clues would apply to EVERY save clue in the game!

Whew! Hopefully that was coherent enough. Putting the matter of policy aside,
Ankeszu wrote:there are many easy ways to avoid twin chop discards, including changing the game flow via "stealing" clues, cluing things sometimes less efficiently, sometimes not cluing things in time, or making tokens to 8 to make discard impossible, etc.
An excellent point, and I agree with this. We can summarize it as such:
Players should try to "architect" a line such that it avoids an impending double discard situation, without cluing either card directly.
I'm sure that both of our groups keep this important concept in mind when planning out lines mid-game. However, acknowledge that architecting a clever line is not always possible in all game states! (At least, with the restrictive set of clues that my group is allowed to give.) That's why it is nice to have our current policy as a backup.



Double Cluing the Same Cards

To be clear, if you are clued the exact same thing twice before it gets to be your turn, it means that you can play ALL of the cards that the clue applies to, but in the opposite order than normal. If you are re-clued about a bunch of cards in your hand after you have already taken a turn, it means that you can play all of them from left to right. (I'll update the document to explicitly say this.)

Now, you requested an example. Imagine that a player has, in his starting hand, the following cards, from oldest to newest:
R4, R1, R2, R3, B5

Now, also imagine that none of these red cards are present in any of the other hands. So obviously, we want all of these red cards to play. However, with our framework (or with any framework?), it will be quite awkward to get all of them out without communicating the wrong thing. However, this "double cluing" thing is one of the tricks in our toolbox to work around this awkward situation. By just cluing them reds twice in a row, it gets all 4 cards out in the exact right order. Numerically, we would classify this as a 4 for 2, meaning that we used 2 clues to get 4 cards played, which is quite efficient!

Of course, this is just a simple example. I'm sure you can see how this tool can be useful in a variety of situations; not all cards have to even be the same suit (in a Rainbow game).



The Trash Finesse

The situation you describe is not a Trash Finesse. It is actually a Trash Bluff! This is because player A is in what we call "bluff position" for player B. In simpler terms, this means that player A is allowed to lie to player B. So when player B plays a card, everything stops, because player C can see that player B was lied to. So, you are completely correct in your analysis here. =)

As a matter of policy, we disallow "long-distance" bluffing. This means that player A is ONLY allowed to lie to (or, in other words, to bluff) player B, and not anyone else. (This is first principal #8, Good Lie Principle. And this is key for understanding the next part.)

So now you must be wondering: if that was a trash bluff, then what in the world is a trash finesse? Hold onto your hats, because here we go.

- Imagine a 5 player game with player A, B, C, D, and E.
- The play stacks are: 1B 1M 2G 2R 2Y 2W
(same as your scenario from earlier)
- Player A clues the previous player, player E, about one 2. It is a trash 2.
(same as your scenario from earlier)
- Player C and player D have the two remaining 2's on their finesse positions.

See how this is different? The key difference is player A is not in "bluff position" for player C or player D. So therefore, player A is not telling a lie, like they were before. Player A is telling the truth - they see ALL off the copies of the 2. And the entire table can see this, so it will get BOTH 2s.



Play Time

It sounds like you are really great at Hanabi. I would love to play a few games with you (and your group). I hardly use BGA anymore, so PM me your contact information, or drop in our Discord server and say hi sometime!

Post Reply

Return to “Discussions”