Recommend
87 
 Thumb up
 Hide
31 Posts
1 , 2  Next »   | 

Hanabi & Ikebana» Forums » Strategy

Subject: The Elusive 25 point Game - Tips for effective Hanabi play rss

Your Tags: Add tags
Popular Tags: [View All]
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Introduction:
I started typing up tips as a response to a message in another thread and it quickly ballooned out of control. So I decided to just make a strategy post and collect some of my thoughts. I do want to make a quick disclaimer though, one of the things I love about this game is that there are no hard and fast rules, there is no right way to play that always works in all situations, which is why it's such a fascinating and rewarding experience. So a lot of what I talk about below should be taken as general principles to live by, but not "the right way to play" or anything along those lines. I also fully accept the possibility that I do stupid things and may have missed some important concepts or points, and welcome anyone chiming in on anything I've said or to add more points to the item.

Important:
I'd also like to add that coming to some of these conclusions and styles of play organically with your own group is one of the enjoyable aspects of this game as well, so I wouldn't recommend reading this before you have played a bit of Hanabi, as you may be shortchanging your own enjoyment.

I also know that there is a fair bit of discussion on what conventions and strategies are within the spirit of the game, this guide is meant to discuss things that would be considered within the spirit of the game by most, although I don't say that to denigrate conventions, just to let you know where this is coming from.

Hanabi Strategy:
Basic Topics:
- Clues to give and their value
- Don't get discard happy
- Pay Attention
- The more data you can impart the better
- Token Management
- Negative Inferences
- Process of Elimination
- Something weird just happened

Advanced Topics:
- Why are clues given in weird ways some times?
- Play if you can
- Card Age
- Let things take care of themselves
- Responsibility
- Actions speak louder than words
- Cluing through the process of elimination
- Endgame Deductions
- Random Strange Play
- Endgame Considerations
- Handling Redundant Cards (avoiding failure)

Expert Topics:
- Finese Play
- Wait, why didn't that go horribly wrong?
- This is going to work, wait, what? You bastard!
- The Free Discard

BASIC TOPICS:
Clues to give and their value:
I just want to start with a few comments about the mindset of the table. In general, the goal is to play cards; this makes clues that enable things to happen the most important clues by far. It may be tempting to tell someone "You have all red cards" because you feel like you've given them information about a lot of cards with just one clue token, but that's simply not how this game functions. Forget the suits and numbers of the cards, you are primarily trying to communicate one of three things with any clue that you give (Ideally you can communicate a lot of things with one clue, but these are the simplest). They are the following:
- A card is safe to play (and should be played).
- A card is dangerous (it may result in a fail token or worse, make a perfect score impossible)
- A card is safe to discard

This is the data you are trading in, and the trick to the game is how to encode that data within the constructs that Hanabi provides; all of a number, or all of a color. Now, that being said, which of the clues is more important? I would argue that you always want to give clues that enable safe plays whenever you can. First of all, that's the goal, to play cards, but more than that, getting cards on the table opens up other player’s hands to further clues that result in plays, it also makes more cards safe to discard which makes your game safer in general. A play accomplishes two really good things.
1) It makes a new card playable
2) It makes a new card discardable.

When you play the red one, suddenly the red two is a valid card to play *AND* red ones can be discarded painlessly. This is a very good thing.

Finally, things work better if everyone thinks "They are telling me this so that I will play it" unless you have a compelling reason to think otherwise based on the state of the game. After clues that enable plays, I would say clues that identify dangerous cards are more important just because if your goal is that perfect score, you need to keep these in the players hand until such time as they can be played or that goal becomes unachievable. When it comes to cards that can be discarded, try to do this when you get a lot of cards clued in for 1 token, like telling someone he has three ones when all the ones have already been played. Telling a player about a single card they can discard often just burns through the deck and should be avoided when possible.


Don't get discard happy:
It is very tempting to fall into the trap of telling people about discards, discarding tons of cards, and getting a lot of clue tokens, which allows you to easily identify and play cards. The problem is that invariably, you will run out of cards and you will be out of plays *well* before your perfect score. The size of the deck is every bit as limiting to a perfect game as the number of clue tokens, and needs to be carefully managed. This is another reason why it's always best to target card plays when possible, and not card discards.

If we look at the numbers a somewhat startling revelation can be seen. There are 50 cards in a Hanabi deck. If we look at a 4 player game (the best Hanabi experience IMHO) then 16 of those cards are already in the players hands. 25 cards need to be played and replaced for the perfect game. That leaves just 9 cards left in the deck to refill your discards. Another way to think about this is that every card you discard after the 9th discard will need to be a play in that last go round when the deck runs out. So if you discard 11 cards over the course of the game, when the deck runs out and you each get one more turn, 2 of those turns need to be valid plays if you are going to get that perfect score. You better hope the right order of players are holding those cards and know which ones they are ;)

I also think this is worth pointing out because a group can fall into the trap of focusing on discards, not realizing that they are actually undermining their own efforts. Discard heavy games *feel* like they are under control, since you are learning a lot, you are getting to the relevant cards faster, making progress playing cards, etc. It’s like the game is providing you positive reinforcement for defeating yourself, laughing a dark and sinister laugh all the while; don’t give in to that temptation!


Pay Attention:
It can't be stated more simply, you have to pay close attention. You don't just need to know what cards you have, but to be effective, you need to know what the other players know about their own hands. Sure you can see their cards, but you also want to know things like:
"Are they going to play or discard on their turn?"
"How can my clue that I'm about to give combine with the knowledge they already have to reveal even more?"
etc.

The more you know about what everyone else knows, the more likely it is that you will be able to give good clues, and the better you will score. I will say that I've always had the house rule that the "memorization" aspect of the game isn't really what is fun for us, so saying things like "What do you know about your hand?" or "What do I know about my hand? This is my 3, right?" even stuff like "Oh no, you also know that's a 2, remember, Barry told it to you three turns ago" etc. are generally ok. You have to be very careful to not reveal new data in those times, and also not use such a reminder or inquiry as an indication of what action you want. (for example, it's not really in the spirit of the game to try and tell someone they should play a card by constantly asking them to tell you what they know about that card). If everyone is playing close attention, reminding someone safely of what they've been told shouldn't be difficult, and if you enjoy the memory aspect of this, feel free to not allow this type of interaction.

One last thing, it's also important to pay attention to clues that other people get because it may tell you something about your own hand, a topic I expand on later.


The more data you can impart the better:
Simply put, a clue that points out 2 playable cards is better than a clue that just points out 1 playable card. Watch for opportunities to tell people multiple bits of data whenever you can. Another way this can come up is when a player has a 5 and you are indicating a lower card of that color to play. Rather than saying "That’s a 3" You could say "This and that (pointing to the 3 as well as the card that they know is a 5) are red" Not only have you basically pointed out a card to play, but you've also let them know the color of their 5 in the process!


Token Management:
It is important to consider the implications of the token count. For example, if you take that last clue token, then the next player will be forced to play or discard, ask yourself if that is safe. Once the game gets to a small token count (2 or fewer tokens) I start to try and figure out what will happen in the next several turns before considering whether I should take a clue token, or even what I should use that clue token for. This is just another thing that everyone should be aware of, and also why paying attention is so important. It's hard to manage the clues if you have no idea what other players are going to do.


Negative Inferences:
This item is obvious to most people, but it's simply that when someone tells you that a card is a one, you also know that the rest of your cards *aren't*. Typically this doesn't matter (and is a huge pain in the but to track), but occasionally it can, especially if you start to learn other things about your hand, like what color certain cards are, etc. One of my later topics will offer some tips to be able to manage keeping track of your negative inferences later.


Process of elimination:
Remember that the Hanabi deck is a very structured, limited deck. As a result, the more information you can see (not just what's in other players hands, but also what has been played and discarded) the more you may be able to determine about your own hand. At the end of the game, when there are no cards left in the deck, you should know exactly what's in your hand, just not what order it's in. It's worth noting that you don't need to wait until the end game though for this type of information to be helpful. This can become useful in the midgame, particularly if a lot of 5's have been drawn.

Also, you don't need to figure out exactly what card you have in order to play it, you just need to figure out whether the set of cards you could have is playable. For example, suppose you know about a 5, and can see three other fives, meaning you know which two 5s you could possible have (the two you can't see). If both those possibilities are for stacks that are already at 4, then play your 5, it doesn't matter which one it is, it will work! This logic can occasionally apply to other values as well, depending on the game state.


Something weird just happened:
If at any time in the game something happens that feels very out of place or does not seem to fit with the way your group normally plays, then that is significant, and you should stop to ponder why it happened. This is true of clues as well, if someone gives you a clue, you should assume that they had a reason to do so. If the clue seems stupid or nonsensical (like you think they were telling you something you already knew) then you need to carefully ponder all the possible situations that may lead to them doing something so off the wall and unexpected. More often then not, there is a very good reason for why something weird just happened, but that reason is in your hand, so think about it, and see if you can deduce why it happened. This basic concept is actually fairly important to a lot of the more advanced topics later in this guide. Examples of things that are weird may include (but are certainly not limited to):
- someone playing a card that they shouldn't have had any reason to know was playeable
- someone discarding a card that they definitely knew was playable
- someone discarding a card even though they had another card they knew they could have played
- someone giving a clue to you about a discardable card when there were other, better clues to someone else they could have given


ADVANCED TOPICS
What type of clue was used:
There are two types of clues, number and color. Number clues are often stronger, because it is possible to give multiple plays with one clue. If I tell someone "Those are both ones" they can now think "Cool, I can play both of these ones!" but if I tell someone "Those are both red" they are still left thinking "well which one do I play first?" For this reason, clues should be numbers, ... unless they can't be. So if someone tells you "that's a red card" ask yourself "Why didn't they tell me it was a 3?" The answer is simple, you have other 3s in your hand, and they aren't playable. Keep this in mind! It may prove valuable later.

Another example to this would be situations that come up with 5s. Typically 5s are given as a number clue to tell people to keep them safe and not play or discard them. But if it's a 5 you actually want someone to play rather than save it may be possible to get them to do that by telling them the cards color.


Play if you can:
This one is simple, if you have a play, play it. Don't give clues when you have plays, save those tokens for players that don't know anything about their own hands. There is another compelling reason that you should play rather than give clues on your turn, and that is to avoid collisions. For example, if I am told at the start of the game that I have a one, I don't know what color that one is, so I shouldn't tell anyone else about their ones because they may have the same one as me.


Card Age:
Another huge tip is to keep track of your own card age. The way I do this is by always adding cards in a specific way. Whenever I get a new card, it goes in at the back of my splay, on the right hand side of my cards. That way I know that my oldest card is to the left, and my newest cards are to the right. This has a lot of little benefits to it that are worth while.

One benefit is that it helps me remember certain things, in particular negative inferences. For example, if I was told that I had a single one, and then I played it, because I always manage the order of my cards in such a way, I know that my left 3 cards are not a one. All I have to remember is that I played a one. If I am then told about and play a 2, I can now say that my left 2 cards are not a 1 or 2, and my 3rd card is not a 2. While this may sound convoluted, in practice it's actually much simpler. When you are trying to remember things like that, rather than saying "what was I told and which cards were they" you can think "what did I play?" There is also a structure to your hand that may help in how you are thinking about things. For example, your right most card was the new card that you got for the last thing you played. The card next to it is the card you got for the card you played before that, etc.

Another huge benefit in tracking card age is in discard choice. In playing Hanabi, you will inevitably find yourself in a situation where you have to discard, and you don't have a definitive safe card to discard. In this situation it is best to discard your oldest card for one simple reason, that's the card that people have had the longest amount of time to warn you about, and is therefore the most likely candidate to not be a dangerous card to discard, like a 5.

And finally a nice subtle benefit that can be very powerful, when you are well aware of the age of your cards, this may make it easier for other people to give clues to you that otherwise would be too frustratingly vague. This is because in general, when you get a clue, especially if the game has stalled a bit, it's most likely your newer cards that inspired that clue. For example, let's say that the red one has been taking a long time to show up. Well into the game, you discard a card for a clue token and draw a new card. Shortly after, someone tells you that you have three ones. In this case, it's pretty clear that your newest card is the red one that everyone wants to get on the board. If it had been one of your older cards, they would have told you about it already. This can be very powerful! In this example, because of tracking card age, a single clue was able to tell me 1 play and 2 discards. Awesome.


Let things take care of themselves:
When you and all your partners are tracking card age, you have a fair amount of assurance that your partners aren't going to do rash, silly things, like discarding their brand new cards. To that point, don't bother telling someone about their terminal cards when they are the newest cards, you likely have several rounds of clues before that becomes necessary, and a lot can change in that time. Similarly, although it should seem obvious, I see players do this all the time, if someone has a full hand of crap, don't waste clues telling them so, just let the token management and discard process take care of itself. The worst thing that can happen is that they are forced to discard, but if their are holding a bunch of crap anyways, who cares. You don't need to tell them they can discard, they will assume that since no one ever tells them jack about their hand, that it must be safe to discard it.


Responsibility:
In Hanabi, players often find themselves in situations where it is necessary to prevent a player from discarding, typically because they have a card that will end the game if they discard it, but occasionally because they are about to discard a playable card that will be a pain in the butt because a lot of other things are waiting for that play and its replacement isn't in sight, but perhaps it's really difficult to clue them in on it due to their hand. Either way, all the other players are thinking "protect that player's hand" so who does it fall to? The natural assumption is that the player immediately before that player is ultimately responsible for that players potential discard situation, since they are the ones that may be taking a clue token and forcing a discard. This leads to the concept of responsibility, and is another area that trust can enter the game.

In general, you are responsible for the hand of the player that immediately follows you. This means that if you leave them without clue tokens and they don't have any good information in their hand, they should be able to trust that discarding their oldest card is a safe move. Once established this can help in a lot of subtle ways. For example, if the player before you is constantly discarding to keep you in clues (even if there are already multiple clues) then you know your hand is very dangerous. But even further, if you know something about your hand and you are a bit fuzzy about it, and the person before you forces you to play or discard by taking your last clue, they may be hoping to make you play the card you are a bit fuzzy about. If they were worried about you discarding it because it may have been dangerous, and the way you know about it doesn't make it obvious, they wouldn't have left you in a play or discard situation.

It should be noted that this does not mean you need to only consider the danger in the hand that follows yours. For example, they are also responsible for a hand, and if that hand is also dangerous, they may be inspired to discard earlier then you were expecting. For this reason, you need to constantly be vigilant about the clue token count and how it is going to affect the game.


Actions speak louder than words:
What happens if the player before you has a play, and decides to discard instead, in direct violation of the "default play" rule I just mentioned? This likely means that you have a very dangerous hand and that player didn't want to leave you in a forced play or discard situation. You should be aware of this when it happens. Just like weird clues can be informative, so can weird actions. By default, you should trust that there is a good reason for what other players do, so if they do something that doesn't make sense, rather than chalking it up to them being an idiot, ask yourself "what could be true about my hand that would have caused them to act that way?" This is particularly strong when you take card age into account. If all the players are tracking age, and suddenly you are getting strong warning indications about your own hand, then it's a safe bet that your oldest card (and most likely discard candidate) is game-ender. What's even better, is there may only be a limited number of game enders at the time, and you suddenly have a smaller subset of cards that such a card could be.

This is another good example of trust entering into the equation. It's hard to know whether the signal has been received, but it's powerful if you can figure that out. For example, a player may have multiple dangerous cards, like his two oldest cards. If I oddly discard before he plays two rounds in a row, or maybe not even consecutively, but twice through the course of the game without him playing or discarding anything, that would be a great way of showing him that both of the cards are dangerous, since I'm essentially trusting that he got the message the first time, and I'm sending the message again about what I now assume is his new "discard" candidate. At this level it is *very* easy for those messages to get muddled and misinterpreted, and that's part of the risk of doing things like this, but then again, sometimes they are heard loud and clear and it's great!


Cluing through the process of elimination:
If you are able to use the process of elimination to determine a card is playeable or that a card that could have been terminal is discardable, that can be a hint towards other players about the contents of their own hands, since their own hands is what enabled that process of elimination. For example, there are often occasions where you can determine information about one of your cards that you know is a 5 through what you can see about the other 5s. The nice thing that can work in our favor is that, barring an end of the game situation, it's relatively safe to assume that no one is going to play a 5 unless they are very confident that it will be a valid play. This can be used to impart information! For example, if I have a known 5, and another player has an unknown (to them) 5, I use their 5 to determine my 5 is safe to play. When I play that 5, that player should wonder why the heck I was willing to risk playing a 5 when it may have been dangerous.

This is another example of paying attention and doing a mental check when something odd happens. In this case, the mental check would consist of looking at everyone else's hand, and realizing that the only possible way to know that the 5 that just got played was safe was if the other 5 was in their own hand. This typically will be a situation that applies to 5's, but it may apply to other situations as well, like someone brazenly discarding what could have been a game ending 4, because they just saw that you had drawn the game ender and they knew it wasn't in their hand, etc.


Endgame deductions:
(contributed by SevenSpirits)
The deck running out often sneaks up on people. It's important to realize that once the last card is drawn, you know the contents of your whole hand, though not the order. Random plays (favoring the most recent draws) can be useful if you can afford the strike, if you know you have a certain playable card. Additionally, often you will have an incomplete piece of information, such as that a card is green, or that it is a 5. Near the end of the game, it may transpire that you have vision of every green card but one, or every 5 but one, and therefore learn the identity of the one that's in your own hand.


The Random Strange Play:
(contributed by SevenSpirits)
You get used to the idea that when someone does something that doesn't make sense to you, it's because there's something you don't know about your own hand. There is typically a direct connection; e.g. see the above techniques. But it's also possible to take advantage of this association without any real connection. Two strange and meaningless things you can do:
* Discard a card that is NOT the expected discard (e.g. it was not the oldest card in your hand that you don't have information about).
* Attempt to play a card that you expect won't play.
Both of these (the second one especially) are liable to induce feelings of confusion and doubt in the player immediately after you. The most likely result of this is for them to avoid the course of action that previously seemed obvious to them and/or to play more conservatively. However it is of course quite dangerous as you're doing something dumb to your own hand, and should be used sparingly.


End Game Considerations:
There are a few things in this game that don't really become relevant until you are near the end of the game. When you are doing well, your enemy is almost always going to be the deck running out of cards and the end of the game is going to come from not having enough turns to finish, rather than failing three times or somehow discarding a terminal card. There are a lot of things happening here, but mostly it all comes down to paying close attention to the order in which the remaining card plays are going to happen.

The first way in which this is relevant is whether to discard or give a clue. You may find yourself in a situation where clue tokens are no longer a problem, but the number of cards left in the deck is. In this case, it may make more sense for you to give a clue, even if that clue is totally useless and redundant, just because it doesn't cause you to pick up a card. This happens when closely managing who will end the game and trying to get that out as far as possible. For example, if it's my turn, and I see that the player next to me has the 5 and the player two away from me has the 4 that we need to win, and they know about them, then I would rather give a clue, so that we can get the game around both players again, since the player with the 5 will need to wait extra while the player with the 4 plays their card.

The one of the other considerations at the end of the game is which card to play if you have multiple plays available to you. In general, playing a 5 is always a great move, except at the end of the game when your other plays are needed to move those colors along. For example, if you knew you had a playable 5 and a playable 2 and it was very close to the end of the game (only one or two cards left in the deck) then it is most likely much better to play the 2, so that the other cards can follow, rather than playing the 5, which you will always be able to play on a later turn (should you get a later turn). If you wouldn't have gotten a later turn, then you wouldn't have gotten a perfect game anyways, since everyone would have stalled out waiting for your 2.


Handling Redundant Cards (avoiding failures):
This one will be short and sweet. There are sometimes situations that arise in which giving a person a clue to play something is exceedingly awkward due to the situation you are in, typically because they will, by most conventions, assume you are telling them to play something they are not. This happens often when they have repeated cards, or some cards which are playable and some which are not. Often if you just wait a few rounds and that player is forced to discard, this situation will clean itself up. For example, if a player has a Yellow3, Red2, Yellow1, Yellow3 and you want them to play one of their Yellow 3s (and obviously not the other), rather than trying to give them a vague clue which may cause them to play both 3s, you can simply wait, possibly even force them to discard the oldest yellow 3 by not leaving them any clue tokens, and then easily clue them about their situation.


EXPERT TOPICS
When you are playing with a group of people that are all paying close attention to everything that is going on, some truely beautiful things can happen. One of the best possible plays is when you can use a single clue to tell multiple people exactly what cards they have (or exactly what they should do anyways, like "Oh, I should wait, he will play that card, and then I should play this card" all with a single clue.) What really enables these is essentially the following: tracking card age, tracking what everyone should know, paying attention to exactly when playeable cards appear, trusting that everyone else is operating on a high level where these kinds of clues will work. (One way to get there is to simply try them and explain why afterwards, and people will start to get it).

Finesse Play
It's hard to construct an example fully, but imagine the following. In your hand you know that a certain card is red, but for some reason you don't know what number it is or whether it was playable or not (maybe you got the red clue a while ago and it helped with something else, but you still have that remaining red card in your hand). Let's say that the red pile has a 2 on it. The person before you goes and gives a clue to the person that will play after you, pointing out a red 4 and telling them "That's a 4". So from your point of view, that clue doesn't make sense, because the red 4 isn't playable ...... unless your red card is a 3. So what the player before you has tried to do is finesse telling you that your unknown red card is a 3 by telling another player that their card is a 4 and forcing your hand. Hope that made sense. Those situations are pretty rare because a lot of things need to be true to be able to make use of them, but when it does happen it is intensely satisfying for all involved.

The "Wait, why didn't that go horribly wrong?":
(Titled by SevenSpirits)
(original example)
Here is an example that ties a lot of these concepts together (paying attention, card age, cluing multiple people at one time, token management, etc.). In this particular situation, we had 4 stacks going strong and had stalled hard on the red 1, it was just buried in the deck and it was taking us a while to find it. The player before me (Bob) discards, and draws the red one. On my turn, I point to another players hand (Amy) and point out that she has a single 1, simultaneously letting Amy know she could discard that card while informing Bob that he had finally drawn the 1 we needed. There is a lot going on here, but it worked and it was beautiful, so I'll break it down.

- Amy was tracking card age and she knew that she had been holding that card for a while; if it had been the 1 we needed, then she would have been told about it earlier. She also could see that the 1 we needed was now in Bob’s hand. For many reasons, she knew that the 1 could be discarded.
- Bob was also tracking card age, but was also paying close attention and noticed that my clue to Amy was odd. Why focus on a single discard, and do so in such a dangerous manner (pointing out a single 1 when we need a single 1) and not only that, why did it WORK!? This is even more puzzling if there are other more valid clues that I could have given to someone, like dangerous cards or other playable cards. Bob correctly concludes that his new card must have been what enabled the safe clue to Amy.
- As a result I've used one clue to give Amy a discard (as well as some token management, not forcing her to do something since she would play before Bob) as well as indicate to Bob that he should play the card he just drew.

(SevenSpirit variations)
The red 3 is playable, but no one has it. Finally, Alice discards a card and draws it. Yay! Bob leaps into action. "Carol," he says, pointing at Carol's red 4. "This card is red." Carol cheerfully responds by discarding a different card.

Waitaminute, thinks Alice. First of all, why did Bob tell Carol to play that non-playable card? Is he crazy? And second, why didn't Carol play it?! Of course, she realizes there is an easy answer to the second question: She herself holds the red 3. Carol sees it and reasons that her own red card is not the 3. But the first question: why did Bob jump straight to telling Carol about the 4, before telling Alice about her 3? (Followed by Carol NOT telling Alice where her 3 was.) The answer to that is that the 3 is in the only slot that makes any sense: it was the card she just drew. So on her next turn, Alice plays the card she just drew, and then after that, Carol plays her 4 triumphantly. 2 plays for one clue, high fives all around.

Extension 1: This works even with more players... for example, say Daniel is in the game and also just drew a card on his last turn. But Daniel can see that Alice just drew the red 3, and Alice can see that Daniel didn't. So they both know it's Alice who was just given this free clue.

Extension 2: This also works if the players are in the other order. (For example, if it was Carol who told Bob about his red card.) Alice can still see that something is up before Bob even goes - the clue is basically a lie! It's telling Bob to play a card that's not playable, and the only way this doesn't end in disaster is, again, if Alice has that 3 and can guess which of her cards it is.

The "This is going to work, wait what? You bastard!":
(contributed by SevenSpirits)
This is a meta-technique based on the previous one. It will probably end in disaster if you aren't used to that one. Consider the last situation discussed above, where Carol tells Bob about his red card, which is a 4, while Alice just drew a playable red 3. But instead, Alice just drew a playable green 2, and Carol (laughing maniacally in her own mind) gives Bob the exact same clue anyway. The following then occurs:
* Alice plays the card she believes to be a red 3. It's not, but it plays anyway. Realistically speaking she's going to stare a bit in disbelief and confusion.
* Bob observes that Alice reacted to the clue as though the red card in his hand were a 4, and definitely not a 3. So he saves his red card for later. The end result is that two plays were again clued using only one token, although in this case one of them will have to wait until later to pan out. This is most useful if the red 4 is in danger of being discarded, or the just-drawn card by Alice would be hard to clue normally.

The Free Discard:
(contributed by SevenSpirits)
This one also builds off the idea that just-drawn cards are most likely to be the playable ones. Alice is told quite unambiguously that she has a playable card (e.g. she has been told it's a black 4). But she discards it instead. Why? Because someone else has the black 4 in their hand, in the just-drawn position. That person looks around and sees that they are the only one who could have a black 4 in their just-drawn position, so they play it!

It should be noted that this kind of crap can backfire rather easily. However, I believe that if you are playing with the 6th suit it's pretty much necessary to take risks on this level and hope you are understood.


Conclusion:
After that giant wall of text I want to reiterate what I mentioned at the start of the post; discovering what communicating styles and strategies work best for you is one great part of this game. It would be easy to read this guide and think of Hanabi as this cold, sterile experience, but it's not at all. Probably one of my favorite things about Hanabi is that it involves trust. You have to trust that someone told you something so that you would play it, or trust that a twisty course of logic that makes a clue makes sense is actually true; that your fellow player thought of that twisty logic as well, and gave the clue trusting that you would also see that logic.

If I had to take one thing away as being the most powerful suggestion, it would probably be card age tracking. I encourage comments and discussion, did I get anything wrong? Any good insights that I haven’t captured that you would like to share?
  • [+] Dice rolls
Bart de Vos
Netherlands
Groningen
flag msg tools
Avatar
mbmbmbmbmb
Fantastic strategy article, well done!
After playing the game twice, I thought I knew most of the subtle hint giving tricks, but you've certainly opened my eyes a bit wider!

I agree with you that this is a fantastic game! The only niggle I have with this game is that similar levels of experience can be vital in order to enjoy the game fully. This relates especially to trying to pull off more complex clues. As you implied, the more experienced people are the less likely their seemingly silly actions are in fact 'just idiotic' in stead of 'actually ingenious'.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Sean McCarthy
United States
Seattle
Washington
flag msg tools
designer
badge
Avatar
mbmbmbmbmb
That is a really good summary of a lot of important stuff about Hanabi. Really well done and pretty darn comprehensive. So much so that I feel like if I think of anything you didn't mention, I should add it! Here's something (I guess this all qualifies as advanced):

Endgame deductions:
The deck running out often sneaks up on people. It's important to realize that once the last card is drawn, you know the contents of your whole hand, though not the order. Random plays (favoring the most recent draws) can be useful if you can afford the strike, if you know you have a certain playable card. Additionally, often you will have an incomplete piece of information, such as that a card is green, or that it is a 5. Near the end of the game, it may transpire that you have vision of every green card but one, or every 5 but one, and therefore learn the identity of the one that's in your own hand.

Also, here are a couple of my favorite techniques that we discovered during play (mostly expanding on your "cluing multiple people at one time" section):

The "Wait, why didn't that go horribly wrong?":
The red 3 is playable, but no one has it. Finally, Alice discards a card and draws it. Yay! Bob leaps into action. "Carol," he says, pointing at Carol's red 4. "This card is red." Carol cheerfully responds by discarding a different card.

Waitaminute, thinks Alice. First of all, why did Bob tell Carol to play that non-playable card? Is he crazy? And second, why didn't Carol play it?! Of course, she realizes there is an easy answer to the second question: She herself holds the red 3. Carol sees it and reasons that her own red card is not the 3. But the first question: why did Bob jump straight to telling Carol about the 4, before telling Alice about her 3? (Followed by Carol NOT telling Alice where her 3 was.) The answer to that is that the 3 is in the only slot that makes any sense: it was the card she just drew. So on her next turn, Alice plays the card she just drew, and then after that, Carol plays her 4 triumphantly. 2 plays for one clue, high fives all around.

Extension 1: This works even with more players... for example, say Daniel is in the game and also just drew a card on his last turn. But Daniel can see that Alice just drew the red 3, and Alice can see that Daniel didn't. So they both know it's Alice who was just given this free clue.

Extension 2: This also works if the players are in the other order. (For example, if it was Carol who told Bob about his red card.) Alice can still see that something is up before Bob even goes - the clue is basically a lie! It's telling Bob to play a card that's not playable, and the only way this doesn't end in disaster is, again, if Alice has that 3 and can guess which of her cards it is.

The "This is going to work, wait what? You bastard!":
This is a meta-technique based on the previous one. It will probably end in disaster if you aren't used to that one. Consider the last situation discussed above, where Carol tells Bob about his red card, which is a 4, while Alice just drew a playable red 3. But instead, Alice just drew a playable green 2, and Carol (laughing maniacally in her own mind) gives Bob the exact same clue anyway. The following then occurs:
* Alice plays the card she believes to be a red 3. It's not, but it plays anyway. Realistically speaking she's going to stare a bit in disbelief and confusion.
* Bob observes that Alice reacted to the clue as though the red card in his hand were a 4, and definitely not a 3. So he saves his red card for later. The end result is that two plays were again clued using only one token, although in this case one of them will have to wait until later to pan out. This is most useful if the red 4 is in danger of being discarded, or the just-drawn card by Alice would be hard to clue normally.

The Free Discard:
This one also builds off the idea that just-drawn cards are most likely to be the playable ones. Alice is told quite unambiguously that she has a playable card (e.g. she has been told it's a black 4). But she discards it instead. Why? Because someone else has the black 4 in their hand, in the just-drawn position. That person looks around and sees that they are the only one who could have a black 4 in their just-drawn position, so they play it!

It should be noted that this kind of crap can backfire rather easily. However, I believe that if you are playing with the 6th suit it's pretty much necessary to take risks on this level and hope you are understood.

The Random Strange Play:
You get used to the idea that when someone does something that doesn't make sense to you, it's because there's something you don't know about your own hand. There is typically a direct connection; e.g. see the above techniques. But it's also possible to take advantage of this association without any real connection. Two strange and meaningless things you can do:
* Discard a card that is NOT the expected discard (e.g. it was not the oldest card in your hand that you don't have information about).
* Attempt to play a card that you expect won't play.
Both of these (the second one especially) are liable to induce feelings of confusion and doubt in the player immediately after you. The most likely result of this is for them to avoid the course of action that previously seemed obvious to them and/or to play more conservatively. However it is of course quite dangerous as you're doing something dumb to your own hand, and should be used sparingly.
23 
 Thumb up
1.25
 tip
 Hide
  • [+] Dice rolls
Conan Meriadoc
France
flag msg tools
Avatar
mbmbmbmbmb
Nice piece of advice!

This pretty much sums up the experience that my playing group has gathered, and our playstyle is similar, but you explained everything far better and in more detail than I would have been able to. I picked up a few pointers along the way, especially on weird plays and non-clues.

The section on card age and card order I feel is particularly important, if you want to avoid taking unnecessary risks. There's a lot less randomness involved if you "stick to the plan" and can expect others to do the same.

Keep up the good work!
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Martin G
United Kingdom
Bristol
flag msg tools
badge
Don't fall in love with me yet, we only recently met
Avatar
mbmbmbmbmb
Superb article (and follow-up by Sean). I think this bit really gets to the heart of what's so awesome about Hanabi:

Quote:
Probably one of my favorite things about Hanabi is that it involves trust. You have to trust that someone told you something so that you would play it, or trust that a twisty course of logic that makes a clue makes sense is actually true; that your fellow player thought of that twisty logic as well, and gave the clue trusting that you would also see that logic.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Thanks for adding some great points! I can't believe I left out the end game deductions point, I think that may even belong in the "basic" category, and people often forget that they can figure out whether their four is dangerous, or maybe that they have a 5 and although they don't know it's exact color, they know that either option for what it could be is playable, etc.

Also....

SevenSpirits wrote:

The "This is going to work, wait what? You bastard!":
This is a meta-technique based on the previous one. It will probably end in disaster if you aren't used to that one. Consider the last situation discussed above, where Carol tells Bob about his red card, which is a 4, while Alice just drew a playable red 3. But instead, Alice just drew a playable green 2, and Carol (laughing maniacally in her own mind) gives Bob the exact same clue anyway. The following then occurs:
* Alice plays the card she believes to be a red 3. It's not, but it plays anyway. Realistically speaking she's going to stare a bit in disbelief and confusion.
* Bob observes that Alice reacted to the clue as though the red card in his hand were a 4, and definitely not a 3. So he saves his red card for later. The end result is that two plays were again clued using only one token, although in this case one of them will have to wait until later to pan out. This is most useful if the red 4 is in danger of being discarded, or the just-drawn card by Alice would be hard to clue normally.


Had never even considered this possibility and I *LOVE* it!!! Can't way to look for these opportunities and discuss this with my local group. Thanks!

Edit: Also, wanted to point out that I added your contributions into the original article, crediting you where they appear. I figured it would be nice to have it all organized and in one place, please let me know if you object.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
I actually wrote this originally late at night right before attending a game conference (geekway to the west, 2012). I played a lot of Hanabi at that conference and was reminded of a few things here and there that I could add to this article. In addition I incorporated all of Sean's excellent observations, as well as restructured and reorganized things a bit. For those that don't want to re-read that entire thing, but may be interested in what has been added, it is the following:

Under Basic topics:
Added the following:
Quote:
Process of elimination:
Remember that the Hanabi deck is a very structured, limited deck. As a result, the more information you can see (not just what's in other players hands, but also what has been played and discarded) the more you may be able to determine about your own hand. At the end of the game, when there are no cards left in the deck, you should know exactly what's in your hand, just not what order it's in. It's worth noting that you don't need to wait until the end game though for this type of information to be helpful. This can become useful in the midgame, particularly if a lot of 5's have been drawn.

Also, you don't need to figure out exactly what card you have in order to play it, you just need to figure out whether the set of cards you could have is playable. For example, suppose you know about a 5, and can see three other fives, meaning you know which two 5s you could possible have (the two you can't see). If both those possibilities are for stacks that are already at 4, then play your 5, it doesn't matter which one it is, it will work! This logic can occasionally apply to other values as well, depending on the game state.


Under Advanced Topics:
Added the following:
Quote:
Let things take care of themselves:
When you and all your partners are tracking card age, you have a fair amount of assurance that your partners aren't going to do rash, silly things, like discarding their brand new cards. To that point, don't bother telling someone about their terminal cards when they are the newest cards, you likely have several rounds of clues before that becomes necessary, and a lot can change in that time. Similarly, although it should seem obvious, I see players do this all the time, if someone has a full hand of crap, don't waste clues telling them so, just let the token management and discard process take care of itself. The worst thing that can happen is that they are forced to discard, but if their are holding a bunch of crap anyways, who cares. You don't need to tell them they can discard, they will assume that since no one ever tells them jack about their hand, that it must be safe to discard it.


Cluing through the process of elimination:
If you are able to use the process of elimination to determine a card is playeable or that a card that could have been terminal is discardable, that can be a hint towards other players about the contents of their own hands, since their own hands is what enabled that process of elimination. For example, there are often occasions where you can determine information about one of your cards that you know is a 5 through what you can see about the other 5s. The nice thing that can work in our favor is that, barring an end of the game situation, it's relatively safe to assume that no one is going to play a 5 unless they are very confident that it will be a valid play. This can be used to impart information! For example, if I have a known 5, and another player has an unknown (to them) 5, I use their 5 to determine my 5 is safe to play. When I play that 5, that player should wonder why the heck I was willing to risk playing a 5 when it may have been dangerous.

This is another example of paying attention and doing a mental check when something odd happens. In this case, the mental check would consist of looking at everyone else's hand, and realizing that the only possible way to know that the 5 that just got played was safe was if the other 5 was in their own hand. This typically will be a situation that applies to 5's, but it may apply to other situations as well, like someone brazenly discarding what could have been a game ending 4, because they just saw that you had drawn the game ender and they knew it wasn't in their hand, etc.


also the "Actions speak louder than words" section was expanded upon (first paragraph expanded, second paragraph added)
Quote:
Actions speak louder than words:
What happens if the player before you has a play, and decides to discard instead, in direct violation of the "default play" rule I just mentioned? This likely means that you have a very dangerous hand and that player didn't want to leave you in a forced play or discard situation. You should be aware of this when it happens. Just like weird clues can be informative, so can weird actions. By default, you should trust that there is a good reason for what other players do, so if they do something that doesn't make sense, rather than chalking it up to them being an idiot, ask yourself "what could be true about my hand that would have caused them to act that way?" This is particularly strong when you take card age into account. If all the players are tracking age, and suddenly you are getting strong warning indications about your own hand, then it's a safe bet that your oldest card (and most likely discard candidate) is game-ender. What's even better, is there may only be a limited number of game enders at the time, and you suddenly have a smaller subset of cards that such a card could be.

This is another good example of trust entering into the equation. It's hard to know whether the signal has been received, but it's powerful if you can figure that out. For example, a player may have multiple dangerous cards, like his two oldest cards. If I oddly discard before he plays two rounds in a row, or maybe not even consecutively, but twice through the course of the game without him playing or discarding anything, that would be a great way of showing him that both of the cards are dangerous, since I'm essentially trusting that he got the message the first time, and I'm sending the message again about what I now assume is his new "discard" candidate. At this level it is *very* easy for those messages to get muddled and misinterpreted, and that's part of the risk of doing things like this, but then again, sometimes they are heard loud and clear and it's great!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
While playing Hanabi at Geekway the concept of determining cards through process of elimination came up, but in particular, this situation arose:

There are two stacks at 4, no stacks complete. 4 5's were out in various players hands (for simplicity sake, let's say one in each players hand). There's a lot of potential inferences that can happen here...

If 3 of the 5's that are in player's hands are unplayable, then the player with the playeable 5 can determine this to be the case, and play their 5. When this happens, every player should be able to determine from that action what the identity of their 5 is. (If they had one of the playeable 5's, then the other player would not have been comfortable playing their own 5.)

But here's another potential situation that I would be interesting in hearing people's thoughts on: If we assume that everyone playing strongly prefers to play a 5 on their turn as soon as they possibly can since it is both a play and a clue token generator, then can we read information into the fact that none are willing to play? This is going to happen because both of the playable 5s are in player's hands. If you've had a few rounds with no one willing to play their 5, then you should be able to assume that there are two playeable 5s. If you only see one of them, then you have the other one, and should play it. I guess the real trick here is how long do you wait until you decide to read information from the fact that people haven't played their fives (and this really only works if you know everyone playing is going to be constantly checking to see if their 5 is playable.)

There are some wrinkles with this thinking if one player is holding multiple 5's, but it still applies to other players in that game who only hold one 5. Imagine something like this happens but one player (who is holding a playable 5) is holding 2 5s (one playable, the other not playable). If we trust that that player has inferred one of his 5's is playable, might as well clue him in to the color of the nonplayable 5, effectively telling them the identity of both, right? (gotta be careful, I'd only do that if the other 5 was the only card of that color so as not to have them misinterpret the offcard that also got clued)

Other situations may have similar implications, but they would all revolve around 5's and the process of elimination. For example, if only 1 is playeable, and 4 are in players hands, and absolutely no one is willing to play, and there is one in each hand, then you assume that the playeable card is still in the deck, at which case, every knows the identity of their 5s!
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Scott Burns
United States
Centennial
CO
flag msg tools
badge
Avatar
mbmbmbmbmb
Hanabi is the perfect example of complex play emerging from simple rules. It's a masterpiece. d10-1d10-0
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Played a stupid amount of Hanabi this past weekend; something like 16 - 18 hours of it, with two 25 point games(one of which used a fail token, and the other with no fail tokens). While playing, I realized a few more things that were worth adding to the guide. The items added to the guide are found in the quote block.

Under Basic Topics:
Quote:
Something weird just happened:
If at any time in the game something happens that feels very out of place or does not seem to fit with the way your group normally plays, then that is significant, and you should stop to ponder why it happened. This is true of clues as well, if someone gives you a clue, you should assume that they had a reason to do so. If the clue seems stupid or nonsensical (like you think they were telling you something you already knew) then you need to carefully ponder all the possible situations that may lead to them doing something so off the wall and unexpected. More often then not, there is a very good reason for why something weird just happened, but that reason is in your hand, so think about it, and see if you can deduce why it happened. This basic concept is actually fairly important to a lot of the more advanced topics later in this guide. Examples of things that are weird may include (but are certainly not limited to):
- someone playing a card that they shouldn't have had any reason to know was playeable
- someone discarding a card that they definitely knew was playable
- someone discarding a card even though they had another card they knew they could have played
- someone giving a clue to you about a discardable card when there were other, better clues to someone else they could have given



Under Advanced Topics:
Quote:
Responsibility:
In Hanabi, players often find themselves in situations where it is necessary to prevent a player from discarding, typically because they have a card that will end the game if they discard it, but occasionally because they are about to discard a playable card that will be a pain in the butt because a lot of other things are waiting for that play and its replacement isn't in sight, but perhaps it's really difficult to clue them in on it due to their hand. Either way, all the other players are thinking "protect that player's hand" so who does it fall to? The natural assumption is that the player immediately before that player is ultimately responsible for that players potential discard situation, since they are the ones that may be taking a clue token and forcing a discard. This leads to the concept of responsibility, and is another area that trust can enter the game.

In general, you are responsible for the hand of the player that immediately follows you. This means that if you leave them without clue tokens and they don't have any good information in their hand, they should be able to trust that discarding their oldest card is a safe move. Once established this can help in a lot of subtle ways. For example, if the player before you is constantly discarding to keep you in clues (even if there are already multiple clues) then you know your hand is very dangerous. But even further, if you know something about your hand and you are a bit fuzzy about it, and the person before you forces you to play or discard by taking your last clue, they may be hoping to make you play the card you are a bit fuzzy about. If they were worried about you discarding it because it may have been dangerous, and the way you know about it doesn't make it obvious, they wouldn't have left you in a play or discard situation.

It should be noted that this does not mean you need to only consider the danger in the hand that follows yours. For example, they are also responsible for a hand, and if that hand is also dangerous, they may be inspired to discard earlier then you were expecting. For this reason, you need to constantly be vigilant about the clue token count and how it is going to affect the game.


End Game Considerations:
There are a few things in this game that don't really become relevant until you are near the end of the game. When you are doing well, your enemy is almost always going to be the deck running out of cards and the end of the game is going to come from not having enough turns to finish, rather than failing three times or somehow discarding a terminal card. There are a lot of things happening here, but mostly it all comes down to paying close attention to the order in which the remaining card plays are going to happen.

The first way in which this is relevant is whether to discard or give a clue. You may find yourself in a situation where clue tokens are no longer a problem, but the number of cards left in the deck is. In this case, it may make more sense for you to give a clue, even if that clue is totally useless and redundant, just because it doesn't cause you to pick up a card. This happens when closely managing who will end the game and trying to get that out as far as possible. For example, if it's my turn, and I see that the player next to me has the 5 and the player two away from me has the 4 that we need to win, and they know about them, then I would rather give a clue, so that we can get the game around both players again, since the player with the 5 will need to wait extra while the player with the 4 plays their card.

The one of the other considerations at the end of the game is which card to play if you have multiple plays available to you. In general, playing a 5 is always a great move, except at the end of the game when your other plays are needed to move those colors along. For example, if you knew you had a playable 5 and a playable 2 and it was very close to the end of the game (only one or two cards left in the deck) then it is most likely much better to play the 2, so that the other cards can follow, rather than playing the 5, which you will always be able to play on a later turn (should you get a later turn). If you wouldn't have gotten a later turn, then you wouldn't have gotten a perfect game anyways, since everyone would have stalled out waiting for your 2.


Handling Redundant Cards (avoiding failures):
This one will be short and sweet. There are sometimes situations that arise in which giving a person a clue to play something is exceedingly awkward due to the situation you are in, typically because they will, by most conventions, assume you are telling them to play something they are not. This happens often when they have repeated cards, or some cards which are playable and some which are not. Often if you just wait a few rounds and that player is forced to discard, this situation will clean itself up. For example, if a player has a Yellow3, Red2, Yellow1, Yellow3 and you want them to play one of their Yellow 3s (and obviously not the other), rather than trying to give them a vague clue which may cause them to play both 3s, you can simply wait, possibly even force them to discard the oldest yellow 3 by not leaving them any clue tokens, and then easily clue them about their situation.


I also noticed that a Hanabi vocabulary that had started to develop. Whenever someone did one of the expert moves (cluing multiple people in at the same time) we would call that a "Finesse." For example, if I tell Larry about his red card, which immediately clues Bob that he should play his newest card to make my red clue work, then I "finessed Bob through Larry." We also called SevenSpirits brilliant "trick them into playing through a finesse" type of play a "Bluff" If we were in the same situation as before, but Bob doesn't have the red card that makes it work, and instead has another playable card that is hard to discard, and I give the same clue, causing Bob to play his new card, thinking it was red, then I "Bluffed Bob off of Larry" and Larry needs to realize that a bluff has been "bounced off him." Added to this would be any card that would end the game being called a "terminal" and someone's newest card being "on deck," these terms became useful when discussing how games went after a hand, "Did you realize you could have finessed his 3 through Bob?" "I was trying to bluff Mary off Adam, but it didn't work" "Adam's oldest card was terminal, so I had to discard" "Since Larry had the 3 on deck, I knew I could ... " etc. In an odd way, this really added to the fun for us. I think the other term we had was "force" for example, if a player has a hand full of plays, but knows that one of them is a 2 (maybe the rest are 1s) and I give a clue about someone else's 3, then I've forced their 2 (I may have don't that for timing reasons, or simply to give the other player a useful clue because there weren't any other.)
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mike Olson
United States
South Bend
Indiana
flag msg tools
Avatar
rhitmojo wrote:
someone's newest card being "on deck"


It's funny... after all that time playing with you this weekend I was not using words the same way. I was using "on deck" to mean a card in a player's hand that they knew was playable, regardless of its position, and "play position" to mean their newest card. I don't think it led to any serious miscommunications, though.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Clyde W
United States
Washington
Dist of Columbia
flag msg tools
Red Team
badge
#YOLO
Avatar
mbmbmbmbmb
Quote:
It is very tempting to fall into the trap of telling people about discards, discarding tons of cards, and getting a lot of clue tokens, which allows you to easily identify and play cards. The problem is that invariably, you will run out of cards and you will be out of plays *well* before your perfect score. The size of the deck is every bit as limiting to a perfect game as the number of clue tokens, and needs to be carefully managed. This is another reason why it's always best to target card plays when possible, and not card discards.
I'm confused by this...(forgive me, the rules aren't in front of me) Are you implying that if you discard a card, you get a clue token?!

I've never played this way before. Also I am confused by your entire article...we've always played where you discard cards face down, so you never get to know what they were, but a few things you write here contradict that concept.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
clydeiii wrote:
Quote:
It is very tempting to fall into the trap of telling people about discards, discarding tons of cards, and getting a lot of clue tokens, which allows you to easily identify and play cards. The problem is that invariably, you will run out of cards and you will be out of plays *well* before your perfect score. The size of the deck is every bit as limiting to a perfect game as the number of clue tokens, and needs to be carefully managed. This is another reason why it's always best to target card plays when possible, and not card discards.
I'm confused by this...(forgive me, the rules aren't in front of me) Are you implying that if you discard a card, you get a clue token?!

I've never played this way before. Also I am confused by your entire article...we've always played where you discard cards face down, so you never get to know what they were, but a few things you write here contradict that concept.


Yes, when you discard (rather then play) you gain a clue token that anyone may use. We also play that all cards are discarded face up (and failed plays are kept face up) so that we are able to track which cards are left, as well as know which cards, if discarded, could result in a game that is not possible to play all the way to 25.

From the rules:
Quote:
Discarding a card allows you to return a blue token to the lid
of the box. You discard a card from your hand and place it,
face up, in a discard pile.


Also from the rules:
Quote:
You can consult the discard pile at any time.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Martin G
United Kingdom
Bristol
flag msg tools
badge
Don't fall in love with me yet, we only recently met
Avatar
mbmbmbmbmb
clydeiii wrote:
I'm confused by this...(forgive me, the rules aren't in front of me) Are you implying that if you discard a card, you get a clue token?!

I've never played this way before. Also I am confused by your entire article...we've always played where you discard cards face down, so you never get to know what they were, but a few things you write here contradict that concept.

So you've been playing that you just get 8 clue tokens total for the whole game? Wow, that would be *tough*!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Clyde W
United States
Washington
Dist of Columbia
flag msg tools
Red Team
badge
#YOLO
Avatar
mbmbmbmbmb
qwertymartin wrote:
clydeiii wrote:
I'm confused by this...(forgive me, the rules aren't in front of me) Are you implying that if you discard a card, you get a clue token?!

I've never played this way before. Also I am confused by your entire article...we've always played where you discard cards face down, so you never get to know what they were, but a few things you write here contradict that concept.

So you've been playing that you just get 8 clue tokens total for the whole game? Wow, that would be *tough*!
Ha, yes, indeed, we have. No wonder I've thought the game was pretty much impossible to get 25 points with.

If you play it this way, you'll find you discard pretty much half the deck, and ONLY give clues when you can get three or four cards out onto the table with the clue. Ie, it makes the game quite dependent on getting a lucky hand.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Martin G
United Kingdom
Bristol
flag msg tools
badge
Don't fall in love with me yet, we only recently met
Avatar
mbmbmbmbmb
I think you're going to really enjoy the actual game
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Teik Chooi Oh
United Kingdom
Chorley
Lancashire
flag msg tools
badge
Avatar
mbmbmbmbmb
We just started Hanabi over the weekend, played 4 games in total. Achieved 25 with 'joker variant' but without, we could not.

Glad to read that we sussed out many of the tips here. Can't use many of the 'expert' tips here unless our group (which being a convention type thing I was at, would not be the same again) have similar 'group think'. Did really like the 'B draws playable Red 3, A tells C that C has a Red, which is a 4' hence informing B her newly drawn card is the red 3. Will try that!

I can only contribute 2 'tips', one obvious & one which am sure your expert groups know but beginners may appreciate. First is that 'use the discard pile'. By realising which duplicates are discarded, you can work out if certain numbers more vital than others since they are remaining copy left required. Which brings me to 2nd tip, if you have a card (eg white 4) which you know is a 4 but not white & there is a white 4 in discard pile. If you are 'forced' to play or discard (ie no info tokens left), & '4' is your only choice, do NOT take risk of discarding if we will 'lose' (ie you want 25 & nothing less) as just play and if its wrong, take the red/lightning (in our version) tokens. I consider those 'red' potential mistakes like extra lives.

Basically, if in doubt a card is playable or discardable, and you know if discarded, you won't complete, just play it & suffer potential 'mistake' than to discard & hope for the best. This is used (as you should always) with observing the entire discard
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Edward
United States
New York City
flag msg tools
badge
Avatar
mbmbmbmbmb
You realize that if you misplay, then you both take the red token AND discard the card? You don't get to keep it.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Martin G
United Kingdom
Bristol
flag msg tools
badge
Don't fall in love with me yet, we only recently met
Avatar
mbmbmbmbmb
theory wrote:
You realize that if you misplay, then you both take the red token AND discard the card? You don't get to keep it.

I think he meant that if you have a card that you think might be a (playable) white 4 or might be some other (unplayable) 4, then rather than risk discarding the second white 4, play it instead. If it is the white 4, great. If it's not, oh well, you still have another copy of this 4 somewhere.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
qwertymartin wrote:
theory wrote:
You realize that if you misplay, then you both take the red token AND discard the card? You don't get to keep it.

I think he meant that if you have a card that you think might be a (playable) white 4 or might be some other (unplayable) 4, then rather than risk discarding the second white 4, play it instead. If it is the white 4, great. If it's not, oh well, you still have another copy of this 4 somewhere.


If your goal is "perfect score or bust" than I can't endorse this approach, the fails are particularly annoying because they chew away a card but get you nothing for it. Since the biggest limitation is often the number of cards more than anything else, this is particularly a problem.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Teik Chooi Oh
United Kingdom
Chorley
Lancashire
flag msg tools
badge
Avatar
mbmbmbmbmb
Our goal was a perfect 25 hence why better to play and suffer potential red token and discard (if it was wrong card, in which a copy exists) rather than discard the remaining copy and ruining the game if you knew that chance existed.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
chooi wrote:
Our goal was a perfect 25 hence why better to play and suffer potential red token and discard (if it was wrong card, in which a copy exists) rather than discard the remaining copy and ruining the game if you knew that chance existed.


But those aren't your only options, why not discard a different card? Or use your turn to give a clue. There are times when we give a clue not because it was useful, but because we weren't comfortable playing or discarding and just need to get past our own hand.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Teik Chooi Oh
United Kingdom
Chorley
Lancashire
flag msg tools
badge
Avatar
mbmbmbmbmb
Sorry, to clarify, I meant in situations where you are not sure if other cards are safe or not & your only information is the 1 card's number, then you can (or in my view should) play if potential to lose 25-point game.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
chooi wrote:
Sorry, to clarify, I meant in situations where you are not sure if other cards are safe or not & your only information is the 1 card's number, then you can (or in my view should) play if potential to lose 25-point game.


Right, that's what I thought you were saying. "I know this card is a 4, so I either play it, or I discard" But what I was saying is why not just keep the 4 for now, and discard another card? Or give a clue? Wait until a situation where you have more confidence that the 4 you have is safe to play (or discard) and then you can do so without taking a risk.

Are you saying you may be worried about *all* your cards not being safe to discard? In that case, if the other players have forced your hand by leaving you with no clue tokens, then there probably are safe discards (or they've played poorly) and if there are clue tokens available, than just give a clue.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Teik Chooi Oh
United Kingdom
Chorley
Lancashire
flag msg tools
badge
Avatar
mbmbmbmbmb
situation I can recall was player knowing he held 3 other vitals (think were all 5s) so only 1 safe play or discard in which we had no time to give info. He decided to discard, 'losing' our 25-point attempt whilst I pointed out that since white 4 was in discard & playable, & I think he knew it was a 4, he should have played it since we still had '3 lives'.

basically, my 'tip' is, if forced to play or discard a card which if discarded could prove vital & prevent perfect score, you should play it!
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
1 , 2  Next »   | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.