situational something

When playing non-tile-related games it’s polite or even just normal for a guy to say ‘gg’ after the end of a game. Good game. Perhaps it was a long drawn-out struggle, a close match going down to the wire, or perhaps the newcomer was totally outclassed and outmatched. Either way, the better player won, and the worse guy learned something, or at least is taking his defeat like a man. I’ll be the first to admit though, that after losing a game of mahjong, ‘gg’ won’t be the first thing that comes to mind, only taking precedence after “what the fuck” and “I can’t believe this shit” or alternately “yeah whatever”.

Many times when we get together to play mahjong you’ll see people raging over newbies doing this or that – usually centering around making cheap yakuhai/open tan yao hands. Is it wrong then to assemble these hands, and does making them make me a noob? Why is it then that similar hands play a huge part in high-level play as well? Does ponning a haku make me a high-level player, then? Obviously not. The short answer is this: how well you play has nothing to do with whether you win.

…in the short term, that is. Say what you want about magic sand and the heart of the tiles, but at the end of the day mahjong is an EV game, and you have to make +EV decisions. There, I said it. Thing is, mahjong is fundamentally different from many simpler games, in that doing the same thing over and over again may not be a recipe for success – unlike most sports, where you just have to keep shootan dem hoops or scoran dem goals and hope the other guy scores less. Mahjong is a game of chance where it’s perfectly natural to lose. In fact, before you’ve even started, there’s an overwhelming chance you have already lost the game. 75% is not an easy feat to overcome.

But, you might say, there’s this dude that wins matches all the fucking time! I never win when playing against him no matter how hard I try! It can’t be just luck on his side! He just rapes people like… a thing that rapes! A lot!

Mahjong is a 25% game. Everyone gets a fair quarter of the pie.

So stop dropping your fucking pie on the fucking floor.

25% is the best-case scenario. If there are holes or weaknesses in your game, you’ve already given up a good portion of your chance to win, because you won’t be able to take full advantage of the chances when fortune smiles on you. Sure, a person can run hot or horrible for quite some time, but if you just can’t seem to win no matter what, it’s time to go inside your own game for a look. What’s that? I said playing skill has nothing to do with winning? Well, the opposite of that is obviously not true – you can be terribad and consequently never win a single match ever. He’s just scooping up your pie.

But let’s get back to whatever I was trying to talk about before you distracted me. Thing about mahjong is that winning the same 2000-point hand can make all the difference in some situations, and in other circumstances you’re doing nobody a favor. Not even yourself. The difference is situational something. I’ll get back to you on that, when I think of a nice word – situational strategy sounds fine but it’s not really a strategy here, while situational judgment sounds a little too severe, don’t you think?

What is this situational something I am talking about here? It’s what you should have done before tossing your riichi stick out there, or after drawing a funny tile and suddenly your 5+5 seconds seem to melt by in an instant. It’s assimilating all the relevant information and factors present up to this turn, and making a judgment call on where to go from here. If possible, it is best to take some time for quiet reflection, as demonstrated by Kaiji below. Of course, if you’re playing on tenhou you can’t plead a bathroom break, so you better make the most of your 5 extra seconds.


However, it usually still means making the best you can of woefully incomplete information. For me, I generally don’t try to ‘read’ hands, which may be surprising to you, so the bits of information I have to mess around with are those macro factors that are more apparent, the placing/score situation, the round/match situation, my own hand, opponents’ potential hands, etc.

And this is what helps me win, yes?

Sadly, not really. The decision-making process is just a tool for making the best plans possible given the situation at hand. Whether you actually win or even whether it’s actually accurate is irrelevant. You can make all the best decisions possible and lose just as easily.

If you’re a relative beginner it’s tough to separate playing well with winning, the two things that seem to go hand in hand. Many of my best games are when I drop to 3rd or 4th place but keep my cool and play a solid game, taking chances and making all the right calls, and finally, after lots of spirited struggling… nothing much happens. Well, what did you expect?

Still, mahjong is a game of losing. That’s why it’s so much more important to learn to hang on to the games where you should win. Play to win, but expect to lose – that’s how you’ll achieve satori. And don’t forget to read your SICP.

11 thoughts on “situational something

  1. You know what? it may sound really stupid, but I’ve always tought even chess is a game of luck, in a sense. There, you have 1/2 chanches of winning, and there’s no riichi ippatsu tsumo / tenhou involved. So, assuming two players play perfectly, nobody should win (well actually i don’t know, i’m making that up), so, why do most of chess games end with someone losing? well, it’s because the opponent makes a mistake. Unless you’re an Akagi-level player, it wasn’t you to force your opponent to make that mistake, so we could call it luck. The opponent got distracted for a moment and therefore missed. That means, the only “skill” included in the game is not “playing right”, but rather “not playing wrong”. If you play perfectly all the time (in mahjong that would mean reach tenpai in the fastest way and not dealing in everyone’s hand), then there’s two possibilities: either some player makes even a minuscule mistake and you win, or everybody goes on with a flawless game and the one to win is the one with the best draws.
    Just watch Akagi, and you’ll notice that in the most important matches, akagi can perfectly say when the opponent made a mistake and use that information to continue play perfectly (actually, since his game is godly and includes coincidence, we could say it is more than perfect) and just win.
    also what the fuck are a satori and a SICP

  2. I forgot a thing that is quite important, actually.
    Most of the time the biggest doubt when playing is: Should i go for the win? Should i play it safe because it looks like the other player has a big hand?
    Well, the only way to decide is to perfectly read the opponent hand, obviously that’s not something you can normally achieve, and that’s precisely why there’s not much pros that could survive 6 han chans with 20,000 points with a 1st place uma of 30,000. Well, that kind of makes it sound like it all goes down to reading hands, but you’d also have to consider the chance of someone getting a tsumo, or you dealing in someone’s hand after having riichi’d.
    Now, too, i’d not like to be misunderstood as a person who only makes shallow calculations when playing. I believe in the hearth of the tiles and in the magical sand, also I’m absolutely convinced that the chanches of dealing in someone decrease as i press the mouse harder.

  3. Osamu: Actually, I would say that even in sports, doing the same thing over and over again is not be a recipe for success. It it were, we would see the exact same plays over and over again in football and basketball, but we don’t. If one teams starts doing the same play over and over, the other team will adjust to defend that play, eventually forcing the first team to change the play (assuming, of course, that the two teams are fairly even matched; if one team has a height/weight/skill advantage, it becomes pretty hopeless for the second team).

    Taiga: If chess is a game where if two players play perfectly, nobody should win, doesn’t that that mean that it’s in fact not a game of luck? I thought true games of skill meant those games where even if everyone played perfectly, the game should end up in a draw.

    I’d also like to address the bigger point. Yes, in most chess games, if someone wins, it’s because of a mistake being made. But I think that in chess, and mahjong as well, there’s no such thing as playing a perfect game.

    I remember a chess book I read once which detailed the state of the current game ranked on a scale from 10 (checkmate for White) to 0 (checkmate for Black). The starting state was 5.5 for all games (since white is slightly favored at the start). One of the games basically consisted of black mirroring white’s every move, and the scale stayed completely in the 5 to 5.5 range, ending with the most boring draw possible. I have to think that that would be what a theoretical perfect game for both sides looks like, i.e. playing not to lose. In grandmaster games there are always gambits going on, even if they might be only minor or rote gambits. And it’s even more pronounced in games like Chinese chess, where there are massive gambits and uneven trades, and positioning is far more important than piece counts.

    In the case of mahjong, your two “perfect game” criteria, reaching tenpai as fast as possible and not dealing in everyone’s hand are, in the strictest sense, mutually exclusive. Unless you get extremely favorable draws (i.e. going into tenpai, then magically drawing tiles that match the last tile that was thrown), you can’t do both at the same time. And even when you’ve reached tenpai, what then? Declare riichi? Stay silent to trap others? Improve your chances by opening it (assuming you have a hand where this is possible)? Break up the tenpai to play defense?

    For instance, I just played a round yesterday where dealer had declared riichi, and had thrown a 8-sou early on. North had an open honiisou tenpai at the time with 2345678-sou in the concealed part of his hand, waiting on 258-sou. He draws a 7-sou right after the riichi, and he has two choices: toss the safe 8-sou and have a more difficult wait (he already had pon’d 1-sou), or toss the dangerous 7-sou and have a much easier wait. As it happened, he tossed the 7-sou and West tossed 8-sou three tiles later, giving North the win and landing him in second.

    In a statistical sense, North both increased the opportunity of winning and the risk of losing, which averages out to pretty much the same state that he was in previously. Which was the correct course of action in a perfect game? It’s impossible to say.

    I myself just wrote a nine-page strategy guide on Japanese mahjong detailing various strategies that I’ve used or seen on offense and defense. I find, however, that when I actually go play the game, I disregard my strategies half of the time, and I consider myself a very good player. I have to think that there isn’t a strategy guide on how to play the perfect game, because there isn’t one. Even pure statistical reasoning can let you down because opponents will spot that and change their strategy to wait on lower-probability tiles.

    Given that, possibly a “perfect strategy,” or whatever the top players of Japanese mahjong in the world use, is to take the best pure strategies and mix them in an unreadably random fashion. depending on the situational, monetary, emotional, and yes, karmic circumstances.

    1. Perfundle: I’m not talking about perfect play here, but logical play given the circumstances, making use of imperfect information. I’m sure you’ll agree that with reasonable opponents of similar skill level you can still enjoy your 25% without needing to play around your opponents’ waits unerringly.

      And re:sports, I am not talking about strategy but the design of the game in general, scoring points is always good, and there is no time when scoring goals is a bad thing. However in mahjong it’s the decision to score points that matters, and not the actual execution.

  4. Perfundle: Of course, i myself believe a “perfect game” would not be possible, since it would require not only reaching tenpai with the smallest number of tiles possible, which would mean you’d know what you’re going to draw, but also you’d have to perfectly know what your opponents are waiting on so you can keep your tenpai and avoid dealing in.
    As the article says, the “perfect strategy” changes according to the situation, if we want to think of mahjong mathematically, we could say that in normal condition (start of the game) if you need to discard a dangerous tile to keep your tenpai, you’d just have to consider the potential of your hand and the one of your opponent as if they were numbers.
    If you are first, you might want to keep a 75% safe tile and deal a genbutsu instead even if that might mean breaking your tenpai, if you’re last you’d consider dealing a 15% safe tile because it’s your last chanche anyway. All of this while elegantly saying you don’t believe in numbers and it was your gut istinct to lead you to victory.
    >Taiga: If chess is a game where if two players play perfectly, nobody should win, doesn’t that that mean that it’s in fact not a game of luck? I thought true games of skill meant those games where even if everyone played perfectly, the game should end up in a draw.

    What i mean is that for someone to win, you need skill (not to lose) and luck (for the opponent to make a mistake). Of course it’s a different kind of luck than than the one you find in mahjong, but still…

  5. Oh, the perfect play comments were directed at Taiga; I know you never mentioned anything about perfect play. And actually, I enjoy my 75% of not losing more than my 25% of winning, since I only play mahjong on Tenhou, and there not losing is mostly what it’s about.

    Although…I’m not really sure what you’re arguing here, since your post sort of goes all over the place. It seems that the main thrust of your argument is that small wins aren’t bad. There’s also something about when to try for a small hand and also that sometimes karmic lightning strikes you (and not in a good way) no matter what you do. The post has the feel of responses to real-life mahjong rants by other people that you tried to tie together loosely.

  6. @Taiga

    Chess has not been solved. ie all perfect play ends in draw. But still about half of high level game do end in draws. Even playing against supercomputers high level player can force a draw.

  7. I agree with most of what’s being said here, but the ideas about chess I cannot agree with.

    Winning or losing in chess is not simply about making ‘mistakes’. It’s not entirely untrue, but to state it that way is an oversimplification. There are so many moves that you can make and so many permutations of possible games that you cannot easily label what is a ‘mistake’ and what isn’t.

    In a game like Mahjong or Poker it’s usually more clear whether a certain call is good or bad. The results are more contained. Moreover, it would be wrong to characterize chess as a game of ‘luck’. Making errors in chess is not about luck. It’s about forcing your opponent to make errors, which has very little to do with luck. Chess is a game of complete information, so there is probably less luck in a game of chess than pretty much anything else in this life!

  8. @Taiga & Sumio
    You guys might be interested to know that when chess masters analyze positions that aren’t a clear win, they use the term “greater winning chances” in reference to the person who has an advantage. This term is all over the place in opening theory books, which are supposed to teach you “perfect” play.

    The same term, “winning chances,” was also until very recently used in arbitrating draws. I don’t know if that means luck is a factor in a chess game, but chance certainly is. If your opponent makes a stupid blunder that causes you to win a lost game, what can you call yourself other than lucky? And if you’re playing in a tournament that seeds you advantageously (particularly in quads), did you get 2 byes because of skill? :P

    Chess hasn’t been solved but supercomputers unfortunately have reached the point where they trounce high level humans. Not every game, of course, but I don’t believe a human has won a high profile human vs machine tournament in a few years.
    BTW, it’s very possible that perfect play will end in a win for white.

  9. Sure, I’m not saying there is NO luck/chance in chess. Just that there is less luck in it that just about any other game or situation of life. There is ultimately a fundamental difference between games of complete information (like chess) and games of incomplete information (like poker).

    Also, the mahjong statistics quoted in this article are misleading. Your hypothetical chances of winning are 25% only if you are playing against opponents who all have a skill level that is identical to yours. In reality you will go into games where your expected winning percentage is considerably higher or lower.

Leave a Reply

Your email address will not be published.