Recommend
11 
 Thumb up
 Hide
10 Posts

Hanabi» Forums » Strategy

Subject: Play Hanabi Better: Methods to get that perfect game 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
edit: since there are multiple entries for hanabi, I cross posted this. please read the other version, as the thread is more active and has some good follow up info:
http://boardgamegeek.com/thread/804762/the-elusive-25-point-...
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Yoff Lag
msg tools
mbmbmbmbmb
Nice tips here.

With my game group,the card aging is becoming one of our most powerfull helpto play the game.
On the other hand, i advise them to sort their cards with the same back sideaand to roll the card when the have an hint on what it is. And even to pull the card 2cm up to tell they plan to play it or 2cm down when they plantto discard it.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David
Switzerland
Buchs
St. Gallen
flag msg tools
Avatar
mbmbmbmbmb
A lot of great tips here. In time I hope to impart some of them on my group because some are incredibly eager to give hints. So much that I often avoid discarding for tokens to prevent them from just burning it again.

One thing that I'm very skeptical about is the whole idea that if I'm getting a tip about some cards and none of them are known to be bad I should just play them. While it would certainly be a neat way to impart additional information about a card I'd expect that to severely limit the options for giving hints.

This means I can only hint at a number if:
a) all colors of that value are playable
b) no colors of that value are playable
c) the recipient of the tip has information from previous rounds to determine which cards are not safe to play
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Quote:
One thing that I'm very skeptical about is the whole idea that if I'm getting a tip about some cards and none of them are known to be bad I should just play them. While it would certainly be a neat way to impart additional information about a card I'd expect that to severely limit the options for giving hints.


If you were told about multiple ones, would you play them? What if someone else had already played one or two (or been told about them). If so, why change this policy after the ones are out?

You seem to think that the habit of playing cards when told about them means that you should *only* give a clue when you know all the plays are safe, but I don't think that's the case at all. Really you just need to be able to predict how they will interpret the clue, be relatively confident that the *first* one they are going to play will work, and that you will have a few turns to either make the others work as well, clue them to other plays, or warn them that things are what they might think.

I can understand that hesitation, but with the way that the numbers in the game work out, you simply have to impart multiple information with clues. That's only going to happen if people assume that clues about multiple cards mean to play them. I will say that a multi-part number clue tends to be more solid then a multi-part color clue, but either way, you have to trust that if you are going to play them, the players will fix it before it causes problems, otherwise you are waiting for too many clues. Keep in mind also that if you are waiting for situations where you can perfectly give clues, you are constantly going to run into this brick wall of people that have some playeable cards that are hard to clue. Playing this way gives you a method to handle those situations, assuming you can predict how they will play multiple cards. In my group, we think it makes the most logical sense to play new cards first, since those may have been the cards that inspired the clue, and if it had been older cards, we possibly would have gotten that clue earlier.

A quick analysis on a 4 player game: 50 cards, 16 cards in starting hand, 25 cards that we want to play, resulting in only 9 cards left in the deck for things like discards before the deck will run out. This functionally means that ever card passed the 9th that you discard will correlate to a need to play a card on the last round. If we have a perfect situation where everyone can play something and in the right order on the last round, we can raise the discard number by at most 4. This results 13 total discards, plus the 8 starting tokens means you only have 21 clues to give in order to play 25 cards. More realistically, you have 18-20 clues to give. This is all assuming your ultimate goal is a perfect game.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David
Switzerland
Buchs
St. Gallen
flag msg tools
Avatar
mbmbmbmbmb
rhitmojo wrote:
You seem to think that the habit of playing cards when told about them means that you should *only* give a clue when you know all the plays are safe, but I don't think that's the case at all. Really you just need to be able to predict how they will interpret the clue, be relatively confident that the *first* one they are going to play will work, and that you will have a few turns to either make the others work as well, clue them to other plays, or warn them that things are what they might think.

Last game we had a green 1 out and another player had three 1s in hand (red, red, white). I told him they were 1s since any of them were safe to play. So this is where your MO works. But then come the next rounds where with your MO you'd have to rush to make sure he doesn't play a duplicate instead of building towards new plays.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Kempeth wrote:
rhitmojo wrote:
You seem to think that the habit of playing cards when told about them means that you should *only* give a clue when you know all the plays are safe, but I don't think that's the case at all. Really you just need to be able to predict how they will interpret the clue, be relatively confident that the *first* one they are going to play will work, and that you will have a few turns to either make the others work as well, clue them to other plays, or warn them that things are what they might think.

Last game we had a green 1 out and another player had three 1s in hand (red, red, white). I told him they were 1s since any of them were safe to play. So this is where your MO works. But then come the next rounds where with your MO you'd have to rush to make sure he doesn't play a duplicate instead of building towards new plays.


But everything is so situational. Were there any other clues that could lead to plays that were available? Were the ones that he had in anyones else hand? If you have no other good clues to give (other ones on someone else, or maybe some playeable two, etc.) then why not give that clue? Even if you have to spend another clue fixing it, you've gotten 2 plays and a discard for 2 clues, vs. the alternative of what, ... not telling him anything and squandering a few rounds?

You have another 3 rounds to go in order to "fix" your clue (and it is a collective responsibility, not just yours, as everyone else can see what's going to happen), and if there are other plays you could be building on, then build on them instead of giving the "one" clue. Maybe just tell him about the white one by way of a color based clue, etc. If the other plays you may want to build on are predicated on his ones, you are going to have to tell him and get him to play them anyways. Either way, unless there are other players with good clues to receive, it's likely the best move to tell him about his ones.

Sorry for the constant edits, I always reply and then think of like ten other things I wanted to say. Other things to consider, could you bluff or finesse one of his ones? If he is about to play his second one (and cause a fail) you could try discarding your second oldest card (sending a clear message that something is weird and something bad is about to happen) or you could just fix it with a followup clue. The fact that there is no easy, clear cut way to handle situations like this is one of the things I love about this game
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
David
Switzerland
Buchs
St. Gallen
flag msg tools
Avatar
mbmbmbmbmb
rhitmojo wrote:
Other things to consider, could you bluff or finesse one of his ones? If he is about to play his second one (and cause a fail) you could try discarding your second oldest card (sending a clear message that something is weird and something bad is about to happen) or you could just fix it with a followup clue. The fact that there is no easy, clear cut way to handle situations like this is one of the things I love about this game
I wish I could play at such a level in that group. Even with a lot of informal hinting we just barely made a score of 20. It took a few emergency explanations to avoid disaster too ("Just because there is one 3 on the table does not mean all your 4s are save to play." or "Just because there are no blue cards in play does not mean your blue card is safe to play.")

Also maybe things look different with four players. I only ever got to play with 3 which has less opportunities to hint someone before they're up again.

I might be able to try this method with the other group though...
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Adam Kunsemiller
United States
Indianapolis
IN
flag msg tools
badge
Mischievous Hobbes is Mischievous
Avatar
mbmbmbmbmb
Kempeth wrote:
rhitmojo wrote:
Other things to consider, could you bluff or finesse one of his ones? If he is about to play his second one (and cause a fail) you could try discarding your second oldest card (sending a clear message that something is weird and something bad is about to happen) or you could just fix it with a followup clue. The fact that there is no easy, clear cut way to handle situations like this is one of the things I love about this game
I wish I could play at such a level in that group. Even with a lot of informal hinting we just barely made a score of 20. It took a few emergency explanations to avoid disaster too ("Just because there is one 3 on the table does not mean all your 4s are save to play." or "Just because there are no blue cards in play does not mean your blue card is safe to play.")

Also maybe things look different with four players. I only ever got to play with 3 which has less opportunities to hint someone before they're up again.

I might be able to try this method with the other group though...


Heh, yeah, for sure, pretty much everything in the guide past the basic stuff is only going to work if you are playing with people that are all on the same page (or at least are all logical people that are paying attention and will learn, etc.). Also, I feel like Hanabi is really best with 4, hope you get a chance to play it that way!
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Todd Salerno
United States
California
flag msg tools
Avatar
mb
rhitmojo wrote:

A quick analysis on a 4 player game: 50 cards, 16 cards in starting hand, 25 cards that we want to play, resulting in only 9 cards left in the deck for things like discards before the deck will run out. This functionally means that ever card passed the 9th that you discard will correlate to a need to play a card on the last round. If we have a perfect situation where everyone can play something and in the right order on the last round, we can raise the discard number by at most 4. This results 13 total discards, plus the 8 starting tokens means you only have 21 clues to give in order to play 25 cards. More realistically, you have 18-20 clues to give. This is all assuming your ultimate goal is a perfect game.


You do get an extra clue back for finishing a suit, so there are potentially 4 more clues which can be earned through play (but this does not diminish your point about clue scarcity and the need to play the cards!).

Play. Give a clue. Discard. You have to choose wisely. It's a game of efficiency.

The more players there are in the game, the more urgent it is to get playable cards out of your hand and onto the table. You might need to skip a play to give a clue in a 2-player game, but absolutely forget about it in a 5-player game. By the time it gets back to you 5 turns later, a 3 got discarded that would have been playable if you'd only played your 2. Now that suit is hung up until the other 3 can be fished out of the deck. This means that if you keep drawing the cards that play, your game can be less interesting since everyone is telling you what to do. Resist the urge to give clues when you have cards to play.

You also have to resist giving a clue when no one is telling you about your cards. It might be that you have a hand of complete junk. But if your hand configuration makes it difficult to give you clues on your playable cards (especially with duplicates), it can be more efficient to wait until you get rid of a card or two. That only works if you take the hint and discard. Whenever I discard a playable card in my group there's no cause for alarm- it usually just means I had both of them. Letting me discard one first is more efficient than spending multiple clues or risking a strike if I were to play them both - the strike play would not return a clue like the discard does. But I have to discard early enough for this tactic to not screw up the timing of the game. Even when there are good clues to give on my turn, the right play might be to discard.

When giving a clue look to see if there’s a way to get 2 for 1. In my group, one of our favorite plays is to call a card (e.g. a 3) on the player two positions away which is not yet playable. The next player sees this and knows he needs to resolve the problem. The player who gave the clue isn't stupid, so there must be a way. The 2 must be in his hand, but where? It is resolved by playing his just-drawn card from the previous turn. Now the next player can safely play his 3 and all is well. 2 cards played for 1 clue given is a beautiful thing. The just-drawn card is frequently the one that plays, so sometimes you can give 1+ clues. It's better to say "these are your blue cards" and point to several cards, than "this is your 1" when that 1 is in the just-drawn position and playable. The first way is a little more ambiguous, but it imparts more information if you all follow the convention that the newest card will play or the clue won't be given. These leaps of bonus information transmission are needed to get ahead of the ever dwindling deck.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
James Rousselle
United States
Metairie
Louisiana
flag msg tools
Todd,

EXCELLENT IDEA ABOUT EFFICIENCY.

Unfortunately, it was buried inside a long post, so people may miss it.

Last night in a 3 player game, my RH (Right handed player) told my LH that one of his cards was a 3. The game had just started and there were several one's on the board, but no 2's. So either, I had the 2 that would allow the 3 to be played, or I had just drawn a 1.

So I played my newest unknown card. It was a playable 1.

Result: we scored a slight gain vs my RH telling me that I had a playable 1.

Here is another example from last night in the same 3 player game.
A green 4 was in the discard pile.
I was told I had 3 4's in my hand.
My LH had the remaining green 4.
I discarded my 4 that was closest to the discard position.
My LH can now deduce that he has the remaining green 4 somewhere in his hand. Otherwise, why would I risk discarding the remaining green 4?

We have other conventions. I may post them some day.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
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.