• Discard order & logic

    Yes, even though all your cards are the same 'age' at the beginning, discarding from the right at the beginning of the game (assuming you discard from the right during the game) is logical, not conventional.

    Game #1 ends - You have been discarding from the right during last game and placing your new cards on the left. Your teammates have noticed this and knew when to warn you not to discard certain critical cards.

    Game #2 starts - You are about to discard your first card of the game. Which one would be safer to discard?
    Actually, you are not here to be a traitor or something, this is a cooperative game, you don't want to trick your teammates. Therefore, the most sensible, most likely, most predictable slot your teammates will suppose you will discard from is the same slot you discarded from during the last game. So why not discard from there? Because the game allows no communication, therefore you should be as predictable as possible. Logic wins.

    Yeah, mate, but you complain about leftism being conventional. That's pretty much the same, bro.

    Actually, it's pretty different.

    Discard order is about safety, while conventional leftism may be dangerous in some situations because it is about getting cards played. A wrong play results in a strike, which in turn may result in a loss. If you apply conventional leftism to starting hands, you miss out on a lot of good clueing opportunities.

    « No conventional leftismSuppressing reflexes and questioning your knowledge »

  • Commentaires

    Mardi 9 Juin 2015 à 19:16

    Point#1 I adjust (or at least i try :)) my playing style all the time when playing on bga but that does not mean that playing that "adjusted" way is logically best thing to do.

    Having a discard order from the beginning is better there's no doubt about it.

    Of course discarding from right is conventional at the beginning, but that's not the point. Every convention derive from logical. Problems arise when you forget the logic behind the convention.

    Jeudi 11 Juin 2015 à 19:30

    Having a discard order from the beginning is just necessary. Therefore it is logical to think it should be done, in addition to being natural. If you don't do it, it can very likely go bad. Def’ly not conventional.

    Playing conventional leftism (CL) is not necessary. If you don't do it, it's not that bad, there are a lot of ways to make do without it. If the game was awfully hard, e.g. you start with only 2 clues in the pool and you have to play 2 or more cards per clue, then CL would more likely be experience-logical, because it would be impossible to do well without CL.

    Point #1 – "what about players you never played with before?" --> play one game with them and see which side they discard from. Or just save time asking them.

    Point #2 – logical leftism should not be preferred because it is more efficient, it should be preferred (by me at least) because it is more elegant and more intellectually exciting. And it happens to be efficient too, yeah.

    "If you see someone playing conventional leftism last game you should probably assume he will play so next game, too." à Yeah, and you might hate it because there are a lot of good clues you can't give for that reason.

    "The logical move would be adjust your playstyle accordingly when playing with that player and play conventional leftism." à If I ever hang around with a gang of guys whistling at and being disrespectful to girls, I won't adjust and do the same, no. If I hang around with smokers, I won't start smoking, no. Now, if I play with a CL player, I just get bored after two moves, breathe deeply and politely finish the game. I'll play another one with him if he's willing to take the challenge of LL.

    “You are about to ponder your first clue marking multiple cards on your starting hand. Which one would be safer to play?" You just can't know. The only sure thing is that the card closer to the chop is more likely to be important (same reason as why you play 1s from the right).


    Back to our topic. Logical deductions are those that scan all possibilities. This is what you do when playing sudoku or other logic puzzles. You can't just assume stuff. You can’t just go with “it’s more likely that this box contains a 9 than a 7 so I’ll put a 9 in  it” or "Paolo has an Italian name, so he must be the Italian guy in the puzzle". Logic is a strict thing – “is there any chance that it is not what I think?” – if yes, hold off; if not, go for it. I'll make an article about these interesting puzzles someday.

    Playing with CL means ignoring a whole lot of situations where it is counter-productive – is there any chance that it is not…?. Playing with CL means not learning from one's experience (and not liking the LL challenge but this is another topic). Playing with CL is not logical. I still haven’t found any convincing arguments to prove the opposite and for now I maintain that if you think CL is logical, your mind is perverted ;).

    Vendredi 12 Juin 2015 à 01:07

    "Just save time asking them" was just about saving time only for the very point of discard order. Of course you can play a game and observe but given the variety of people I play with IRL, forgetting which side they chop, I would waste a lot of time, which I'd rather spend having fun playing good games.

    Re assuming, I was extrapolating. I make a difference between "can't assume" and "can't just assume". Of course you can assume sometimes, about stuff that is clearly assumable, e.g. a single-card clue, a smart beginner* will soon find out it means the card should be played (therefore logical). Same with the useless clue technique (for it is logical, again) – it is clear that the useless clue means he should play the other card. That same smart beginner may also find out about game-start discard order and logical leftism when receiving a clue (still logical). However, in a CL situation, they can scratch their brains for hours, they won't find the logical thread leading them to knowing to a 100% conclusion (90-95% if you prefer) about which card is playable. This is where logical assumption sets reaches its limits this was my point.

    I'm sorry if you felt upset by the end, it wasn't meant to be as bad as an ad hominem, I was just wink-referring to a point I'd already made in another post.

    And I do have an idea to compare the efficiency of both techniques (I even might not be as bored by CL if it's an experiment as if it was a real game), which is to play preset deals on BGA, one game CL, one game LL. Let me know if you want to try sometime. I think qpenguin will be busy in the near future so we may want to wait a few days.


    *the recurring character of the smart beginner/NASA guy is a "pure" fast-learning mind, immaculate of any learned Hanabi habits/techniques, with only its own resources to think and work out things and playing his first few games being unable to talk to his teammates (who might be smart beginners too)

    Vendredi 12 Juin 2015 à 21:42
    I think that the real point is that LL can solve some situation that CL can't. Of course it has cost in terms of time and clue. If you like the logical co-op aspects of the game you probably prefer LL. If you're not so "logical dependant" you prob go with CL. Is there a best way of playing? Don't know. You could prove that one is better calculating all the possible combination where LL is needed vs all the possible combination where CL save you time and/or clue. If you're going to do I'll be interested in the result. In the meantime I keep playing hanabi :D
    Vendredi 12 Juin 2015 à 22:56

    I'm afraid you've missed my point about the comparison of CL/LL and discard order in terms of logic.

    Anyway I've explained at length why I think discard order is logical and CL is not and why they can be treated differently, so I won't get into this again, it's all above.

    Regarding the comparison experiment, I'm quite sure someone who is good at LL will be good à CL too because CL is easier, "more relaxing" – just play leftmost when clued multiple cards (with very rare exceptions). So we can give it a try, with the same players of course.

    If there are any mistakes in one or another game, we'll be able to tell how it affected the result.

    Samedi 13 Juin 2015 à 17:19

    No beri, your point is clear. And I pretty much agree with you. What I'm trying to say is that other player may not like the logical aspect of the game as much as they like other aspects: like everybody trusting each other in the name of leftism, or the "thrilling of finesse" (even when it's not logically understandable). That's all.



    Dimanche 14 Juin 2015 à 01:15

    Sorry Frollo my comment was directed to Isdariel. On that at least, I was unclear ;)

    Mardi 23 Juin 2015 à 13:32

    I agree with Isdariel when it comes to draw the line. As for a person which doesn't like conventions, it's fun to read that you want to use pre-game info during the game.

    -> "If you see someone playing conventional leftism last game you should probably assume he will play so next game, too." Yeah, and you might hate it because there are a lot of good clues you can't give for that reason. <-
    And there are a lot of  other good clues that may be done because of CL. Play and save with one clue, not few.

    I'm not sure if I use CL as you understand it... so often playing right card of multiple clued cards on chop, when its surely playable and cant cause 3rd strike. Logic is behind both of these styles, and again, I agree with Isdariel - there is more logic in the most effective use of a clue. IMO LL may be good only at the beginning, when needed cards are on players' hands but not in order (but sometimes CL + finesse may handle it!); in later game, when the clues were rationally saved by using CL and we have them enough (rather than in LL when we need more clues to get the cards played) its easier to save crucial cards and clue the newest to play (or finesse) in CL.


    I wish you come back to saving clues with color. Especially with wasting clues in LL, I see it more dangerous to clue color on chop.

    Also I would love if you get back to anticipated saves in less than 4-players games.

    copy here the main part of the comment I gave you on BGA:

    -> I partially disagree with savings with colour. It depends on the situation at the table and on other players' hands. Sometimes, when players are really busy to get all crucial cards saved, there is no time for cluing playable but noncrucial card. It also can be understood as finesse - when 'finessed' player see you could tell 3 to save, but instead you clued yellow... I give color save only when its necessary - i.e. person got four 4s (also useless) at hand and we want only to save chop. Player with yellow clue cant be sure, but then, he will discard and see "here there was a four! so clued card is propably 4y".

    Also there is an anticipated save case you haven't written of. Happens mostly when playing 3player game. Early or middle game with not much clues, but it's quite obvious no 5 is playable and won't be playable soon. Player A clues B with xy5z. B discards z. C is surprised, tries to play as finessed or discards his own chop... which is y. And A may easy clue y on B hand.
    Shortly - to predict that there will be a case you need to clue 2 cards in a row (and your coplayers cant see that!), when 5 would be the rightmost of saves. So clue 5 earlier, when noone can think it's playable or finesse (or dont care about 'finesse' bomb, sometimes you can bluff this way). If A discard rather than clue, B propably discards too, C clues or discard too. And there A needs to protect both 5 at chop y next to the chop. Doesnt matter there are more clues... C didnt know he should help A to clue, so discarded.
    And there is a problem. How to know - being B player - which case is that? is xy5z clue protecting y or z?
    Logically, it should be mostly version 5z saving - the same story with third player discarding card turn ago and it happened to be on another players chop. But there is easier to manipulate the game the way that saves are only about chop cards, excluding 5s as valuable rare cards; easier to predict and clue 5 earlier.
    I'm curious what do you think about this way of thinking. <-




    Mardi 23 Juin 2015 à 14:06

    I don't want to use pre-game info, read me again, again and again.

    Re your next points, when you say "I wish you come back to" do you mean you would like me to discuss these points again (or write new posts about them) or to question my views about them and change my mind?

    1) Re CL/LL, I'm really not sure LL wastes clues. On one hand it does. On the other hand, it allows moves CL doesn't. LL lets you give clues as soon as they are pure (all targeted cards are important, though not in order), while CL might force you to wait to avoid ambiguity and by the time you can clue safely, such things may have happened that you need more clues to make the first one clear. LL also allows long-term saves that may save a lot of time and some clues as well.

    In any case, even if CL was proven to be more efficient than LL, I just find CL highly inelegant and "easy". Hanabi is art and deserves better. Be this said once and for all (until someone performs the miracle of changing my mind).

    2) Re anticipated save, except for exceptional situations, a clear non-play clue on second chop is to save chop. If you receive a non-play clue on second chop and you can see no cards that need saving, your chop is likely the same as the player before or after you.

    In the case you mention, I let B discard, C discards then I clue B about second chop to save it and the chop 5.

    This is always how I have played anticipated save with more than 2 players and I've adopted it for 2-player as well now, since I find it more efficient, helps handle emergencies.

    Mercredi 24 Juin 2015 à 15:23

    I read you again and again and still I see the same.

    I mean I'd like you write about it. As you did.

    Re anticipated save. And how B should know it isnt a play clue, if there is any playable 4? Especially when the second chop isnt second from right, but lets say second from left card (if some other card was saved before). If there was small amount of clue tokens, and cards of A or C need some save clues soon, B may think its playable.
    Some cases it would be easier to save it on C's chop, even if it costs a bomb from B, but not always.

    Mercredi 24 Juin 2015 à 23:42

    If B hasn't just discarded a 4, he will conclude that if this 4 was playable, you would have clued it earlier. Then he must check if you had an opportunity to do so earlier. And I'm quite sure you did.

    Save it on C's chop, it really depends if C could be clued with a pure clue. If he could, it would certainly be best to save the card on C. But sometimes it's not that easy. Situation, always situation, considering all possible terrible situations vs. reflexes.

    "If there was small amount of clue tokens, and cards of A or C need some save clues soon, B may think its playable." Yeah, small amounts of tokens always make things complicated and you can't handle all situations. But if there was a clue to save the 5 first, let B discard, then clue the new chop, there is a clue to let B discard first then do anticipated save. Since there was a clue, B could see you had an opportunity to clue (the card targeted by the anticipated save clue) if it was playable.

    "Even if it cost a bomb from B" – ah the dreaded finesse craze, that makes a lot of smart saves impossible and causes many bombs. What if B has a danger card as newest? Finesse craze makes the situation unmanageable.

    Jeudi 18 Mai 2017 à 02:57
    Hello, someone on BGA told me to read this blog.
    I am slightly amused to read that a discard order is necessary at first.
    It is not.
    You could very well decide that your cards have an age, with your original cards all being Age_0, your first received card from the pile is Age_1, your second is Age_2,... and that you discard your oldest unclued card first (this fits with LL).
    Since there might be multiple unclued cards of Age_0, you can just discard one of them at random.
    It just forces everyone not to discard at first, as long as they have not had a chance to have someone warn them about their 5s or clue them some playable cards.
    Once one person has had a chance to have their hand clued and doesn't see anything urgent in others, they can discard one of their unclued cards of Age_0 at random... And they proceed with this randomized choice as long as there are unclued cards of Age_0 to be discarded.

    Maybe it should be included in the LL guide by the way as it's a way of playing that fits its philosophy since it assumes no conventional arbitrary order.

    PS: I have friends who do not accept the starting arbitrary order and use this randomized choice to discard their original cards.
    Works okay.
    Jeudi 18 Mai 2017 à 14:57

    Not discarding from the same spot at the beginning as you will do later in the game and as you did during most of the last game can only be negative, since all it does is adding randomness – i.e. risk – and, as you very rightly pointed out, forces the team to wait, i.e. lose time and most probably waste clues (i.e. subplay).

    So if it’s only negative, why do it? I find it obvious and therefore logical that something that is only negative should not be done in a coop game. No convention here.

    Suivre le flux RSS des commentaires

    Ajouter un commentaire

    Nom / Pseudo :

    E-mail (facultatif) :

    Site Web (facultatif) :

    Commentaire :