Hanabi : A beginner's guide to conventions and strategies

[Ankeszu]

@Zamiel

As for Cluing Safe Discards and Double Discarding - let's agree to disagree In my experience ambiguity you mentioned is usually contradicted by the flow of the game - who clues and how, when they clue and when they don't, what do they play. In most games it is easy to distinguish between a save or play - and anyway it shouldn't be a problem if a clue giver is aware that a chop clue will be considered as a save. "Each player is incrementally disempowered to perform finesses themselves" - that's what I felt about your way. (Like I said, your system is coherent. I just disagree that the benefits are higher in your case than in my way of playing. )

Double Cluing the Same Cards
(...) R4, R1, R2, R3, B5 (...) 2 clues to get 4 cards played

...Or clue red, then 4. It works exactly the same, within the rest of your framework, and - if instead of B5 there was some 4 we want to save - you can save it as well with it.
I'm not saying that this convention is totally useless, just... I still need to see an example of situation that could be covered only by this convention. The example you gave failed at this. And I really would like to see the person (not knowing the convention) figuring this during the game. Most (all?) of conventions I know/use, given enough time even when first time encountered, are possible to figure out.

ad. The Trash Finesse
OK, I see your point and I agree. (I play 5p games so rarely that the difference didn't occur to me And I admit I didn't use it in 4p - for double play of one person - either, it would be very situational.)

ad. Play Time
I am somehow curious and somehow afraid of playing with your group - I disagree with many things you mention (including 3 bluffs ), though it would be definately interesting experience. Maybe some day (I don't quite get that chat). Or maybe you will invite your group here?

@Hyphen-ated via Zamiel

Doing complicated chains of reasoning that relies on stuff like "he knows that she knows that I know that he knows this" is something almost completely unique to Hanabi in boardgames, as far as I know. It's like those logic puzzles with the 100 blue-eyed islanders, or the people in a line with colored hats. That's the stuff I really love about the game.

While I agree that these chains of reasoning make really cool feeling, I am not so sure if is it unique (though I am convinced Hanabi WAS a milestone in game design). As for newer games in a second I can think of "Overseers", "Abracada...what?" (here you can't see your own, um, "cards" too) or (2p though similar) "Hanamikoji", generally - all games that rely on information exchange and why someone did what he did. (Though, I admit the main difference between those and Hanabi - all mentioned games aren't cooperative.)

[Zamiel]

Double Cluing the Same Cards

Or clue red, then 4. It works exactly the same,

This is a point of confusion about our framework then, because It works quite differently. Recall that the example is:
R4, R1, R2, R3, B5 (from oldest to newest; the R4 here is on chop)

If reds is clued, and then 4, it is a signal that "the red 4 is playable right now". The person should believe that: "If the red 4 is to play right now, I should have red 1, red 2, and red 3."

This is a situation covered by first principle #6, "Left-Most Playable Principle", which states that you should play your cards left to right. (Note that we call this type of move is called a "Self-Prompt", which has its own section in the document.) So the player would proceed to play their cards left to right, and misplay red 3 as red 1. (In this situation left-most is synonymous with "newest", because we are seeing their hand flipped around from the opposite side, much how it would look in real life.)



Different Frameworks

I just disagree that the benefits are higher in your case than in my way of playing.

Well, one point of non-arbitrary comparison is the results themselves. What is your average score? Our framework is so good that I feel we have completely conquered the Rainbow variant. Both me and Hyphen-ated, one of the members of our group, have had streaks of eight 30's in a row at different points in time. (This is spread between 3, 4, and 5 player games. We don't play 2 player games.) In fact, I would go so far as to say that the only time we don't get 30 is if someone makes a big mistake, or we get an incredibly unlucky deal.

We've actually resorted to having to play harder variants in order to keep the game challenging. We invented two new variants, and these are the only ones we play now. (Unless there are new players around, and then we would play no variant.) For reference, these are the variants:

1) Mixed-color Suits

Suits:
1) Green (Blue / Yellow)
2) Magenta (Blue / Red)
3) Navy (Blue / Black)
4) Orange (Yellow / Red)
5) Tan (Yellow / Black)
6) Burgundy (Red / Black)

Clue Colors:
1) Blue
2) Yellow
3) Red
4) Black

2) Mixed + Multi-color Suits

Suits:
1) Teal (Blue / Green)
2) Lime (Green / Yellow)
3) Orange (Yellow / Red)
4) Burgundy (Red / Purple)
5) Indigo (Purple / Blue)
6) Rainbow

Clue Colors:
1) Blue
2) Green
3) Yellow
4) Red
5) Purple

(This is also part of the reason why we can't use BGA to play. We use our own server.)



Arbitrary Conventions

And I really would like to see the person (not knowing the convention) figuring this during the game.

That's an interesting wrinkle of discussion. While we would like our framework to be as intuitive as possible, this concern is actually not of the utmost importance to my group. We will brief each other on our strategies before the game starts, so this won't be an issue. What is most important to us? Scoring 30 in the maximum number of games that we can, without cheating. (We don't speak whatsoever during the game and don't use any timing tells.)

It seems to me that playing with a set of conventions that are strictly "naturally figure-outable" during the course of the game is an arbitrary restriction that you place on yourself that prevents you from reaching your true potential. We don't limit ourselves to this restriction, and we want to take the game to the limits. Hanabi is an optimization problem, and if some conventions are complicated enough to require briefing each other about them before the game starts, then that is not inherently a bad thing, especially if it gives you more power while actually playing the game. It allows you to put more "tools" in your toolbelt (to use the same metaphor that I used in my previous post), so that you are better equipped to optimize each turn.

In the past, I've argued with players who feel that creating arbitrary conventions like this is an abomination. This quote from Sande24 really shines light on the fundamental issue here:

Hanabi is a co-op game, so it works a bit differently from other games. If you play the game without any prior knowledge, you'd play with some really basic logic. That is fine. But when you play another time with the same group, you inevitably start using some kind of group-strategy to get a better score. The clues will get more complex because you know how the other players react to your clues. You WILL have a pre-defined strategy of some kind in any game like this. So if you want to get better in the game, it is not enough that only you play with some kind of strategy, it has to be set up by the group. Otherwise the ceiling of the logic is quite low.

I think playing a strategy game without a strategy is quite boring too. If you can't really know how the other player would react to your clues, it would make the game much more frustrating. In the end you would have to resort to very basic clues which makes you sacrifice 2-5 points by the end of the game. Having a convention offers you more flexibility in the terms of giving clues to other players and makes the game much more interesting to me.

Also, co-op games where you are not allowed to communicate verbally like Bridge also have these conventions which are set before the game. Hanabi is very much like that.

[Zamiel]

Maybe some day (I don't quite get that chat). Or maybe you will invite your group here?

That's too bad that you don't want to play. Perhaps hearing about these new variants has tickled your interest?

The "chat room" is a Discord server. Discord is a voice and text chat application that allows anyone to create a hosted chat room. We use one such server as the primary hub of communication to see if anyone is around to play a game. You can think of it like Skype, but you don't have to download anything, you can just leave it open in a tab. If you leave it open, we can get together at some point when other people are around and get a 3-5 person game going. (We don't ever play 2 player.)

Using some sort of voice-application is essential because after every single game we all get together in a voice chat and go over the review of the game, turn by turn, to analyze each other's play for mistakes and talk about strategy. This is how we improve! Is that something that you guys do as well?

Or, if Discord isn't to your liking, feel free to PM me to schedule a time to play when your group is around and playing games and I can check out a few games with your group.

[Liallan]

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

That's exactly how I'd prefer to play it, when I've been drinking.

BTW, my dad was an engineer. It's where I seem to get all my traits, though my dad was never a gamer and well, not social. He would probably have preferred discuss how to make the computer figure it out.

[Liallan]

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

Hello Liallan. That's fair - I've edited my post to say that the guide is explicitly meant for intermediate to advanced players.

Well, I'm glad you recognized what I was meaning, but I was talking about beroberokun's post. I saw that you had posted something, but I didn't make any attempt to read it because I was fairly sure I wouldn't understand it either. (Besides, I agree it does sort of take away the fun of trying to learn stuff on your own.) But I'm glad you were able to appreciate what I was getting at.

I have noticed that mostly what this post has done is get experienced people talking about the game and their own strategies. It's mostly over my head.

[Ankeszu]

R4, R1, R2, R3, B5 (from oldest to newest; the R4 here is on chop)

Oh, so the leftmost is the oldest one here? I guess that is the part I misunderstood the example.
(It's hard imagining that kind of situation - in my way it most likely would be clued far earlier, but let's say I see the point. Is that a real game example or imagined one? And if real game, can you please provide me a story of that game? It has to be interesting.)

ad. Different Frameworks
I have no idea what's my typical score . BGA tells 26,5 - including my very early games and with no distinction between Multi/5c games. And ofc it doesn't mention the rest of the team. What would be your typical score with unknown player without sharing strategy, especially if half of your games were 5cards-2p?

When playing with very good team, I believe I had more than 8x30 in a row - in 2p, in 3p, in 4p and in the old days maybe also 5p, before 5p-games started to bore me. It's easily checkable here... If someone has enough time to check

Few random checks: (it would be hard to have a line of 8+ games with no 2p as I like to play it; and I guess it would be hard to find line of 10 games which I play only with fully trusted experienced team... xD)
27Apr-1May 2016 I had a winning streak of 12x30 (3p, 2p and 4p) even though I don't call all of my teammates "best".
17-18 January this year: 8x30 (4p, 3p, one 2p) with good teams.
April-May and March this year: lines of 7x30 (2p+3p, 2p+3p+4p, 3p+4p) - also not necessarly with fully trusted players.

Though as I play mostly here, I don't use other variants except for those that can be done within BGA system and 'telling rules'.

ad. Arbitrary Conventions
I like exploring this game, and I like doing it on my own. Figuring the puzzle by myself is the main fun. When I was teaching people this game, the only thing I was telling first-time-players (except for rulebook rules) was suggestion to discard from the same side each time - that's a style you learn on the very first games anyway. Everything else is up to them to find out - and fell in love with this game like me. Everything comes with experience, and if you can't figure anything by experience but you must be told about it... Where's the fun? How is following someone's footsteps (or, more accurately, watching someone's pictures from the trip - far less than having a trip by yourself, missing the feeling and the continuity of thinking) helping my own development, instead discovering the way by myself?

I play with many people. I admit I started to minimize the number of inexperienced player in my Hanabi games, though there is still pretty much group I can play with - most Masters on this site - and when I first meet anyone in a game, I don't expect to share strategies. I expect thinking, not repeating algorithm I have not made on my own; and if I see my algorithm fails, I rethink it.

On BGG Hanabi forum there is a thread about totally arbitrary convention. What do you think of those?
https://boardgamegeek.com/thread/155116 ... unt-conven

As for the quote you provided, IMO Bridge is not coop game - it's team game. And while rules for aucton differ between groups, I think about it more like about variants than conventions. Have you heard about a group that doesn't use any bidding system at all? I didn't. (I admit I am not very experienced in bridge. Though I do feel that from the good understanding of the rules and goals, it is possible to figure a lot of the bidding system elements.)

I won't give you quotes, I'm just curious of examples - of the feelings when you figured out something not obvious, the eureka shining in other player's eyes, of understanding without having to name every strategy/logic behind your move... I don't see your gamestyle providing it.

This is how we improve! Is that something that you guys do as well?

Often (= if there were some unclear issues during the game; if all players agree that something was effect of distraction&unfocus, not much to discuss about). Though I guess it depends on other players. Some people leave the table immediately or don't want to discuss because of the need to replay and refresh memory. And sometimes there is language barrier you can't avoid. It's pretty hard to talk via Google Translate when both of your primary languages are ... let's say, not well supported by GT (I remember I tried to talk with some Asian, I think Japanese, but the only thing from his/her words that GT translated clearly was "I don't understand", and the other player didn't speak english at all.). Even if both non-english native speakers theoretically speak english, it doesn't mean they can easily talk about tough situations without annoying and time-consuming constant assuring they understand each other.

[RobertBr]

Some of the tone of these posts are what make playing Hanabi online no fun at all. A lot of player treat convetions like they are strategies - they are not. There are lots of possible conventions you could adopt. For example a very effective convention with beginners is 'unambiguous singles' which says that if some-one gives you a clue that identifies only one card they are indicating the card is definitely safe to play, where-as multiple clues may not be safe. This sets up a lot of thinking on the part of the players about how best to play their cards which is very valuable at learning the game. It is also not a system a lot of conventions by intermediate players use (they often treat multiple clues as instructions to play), but there are lots of possible ways you could set up a multi-card convention. Yet most online players think treat their particular convention like it is divine writ on optimal play and will not share the convention when opening a game, instead insisting its 'the' convention.

[Liallan]

Some of the tone of these posts are what make playing Hanabi online no fun at all. A lot of player treat convetions like they are strategies - they are not. There are lots of possible conventions you could adopt. For example a very effective convention with beginners is 'unambiguous singles' which says that if some-one gives you a clue that identifies only one card they are indicating the card is definitely safe to play, where-as multiple clues may not be safe. This sets up a lot of thinking on the part of the players about how best to play their cards which is very valuable at learning the game. It is also not a system a lot of conventions by intermediate players use (they often treat multiple clues as instructions to play), but there are lots of possible ways you could set up a multi-card convention. Yet most online players think treat their particular convention like it is divine writ on optimal play and will not share the convention when opening a game, instead insisting its 'the' convention.

Amen, although I sort of already said this in essense:

Every time I see conversations about this game, it's like everyone assumes that everyone else should all be playing that same way. I would find that a little boring. When I played with my group, I realized you have to know your group.

[Cynthia 56]

J'écris en français car je lis l'anglais bien mieux que je ne l'écris.
Je suis d'accord avec les personnes qui fustigent ces stratégies : bluff, finesse...

Depuis que j'ai appris les règles de ce jeu, j'ai joué avec des groupes super sympas et des personnes très agressives qui s'énervent dès que, par malheur, on jette une autre carte que celle qu'ils voulaient nous pointer grâce à la stratégie Y ou Z.

Pour ma part, ça me gâche mon plaisir et je fais de moins en moins de partie de Hanabi. C'est dommage.

[Liallan]

I feel like they need to put two options on this game: Competitive Camp and Casual Camp. Same game, but that would show with the options. Save a lot of headache.