Tenhou Secrets

Hi. xkime here. So, you have learned all yaku, played for a few years, you are very confident about your tile efficiency skills, and whenever you defend you rarely deal into anyone, but somehow you just cannot rank up at Tenhou. I hope that these little hints will help you.

Note, if you are still below average, you might want to check this link first. Beginners go here.

Tips for Tenhou

What kind of game is Tenhou?

To understand how to beat the system, you must first learn how the system works. In other words, to win in mahjong you must be really aware of what the rules of the game are, and what is considered “winning.” Obviously, getting second place in a game is pretty neat if you are in a tournament where the top 2 players move forwards, but it is just dreadful if only the guy on top will advance towards the finals. Tenhou’s golden rule for ranking up is very straightforward: do not get last place. It is as simple as that.

Why is last place so bad?

If you are at any rank higher than 3rd Dan, you probably know how the distribution of points works for each placement. You can find it in Tenhou’s Manual. In short, as your rank goes up, the reward for first place is relatively medium sized, for second it is a small sized reward, third place gets nothing, and last place gets a huge punishment. Bigger as you rank up. To get the most out of the points you get from the table, you should play in a table according to your skill level as it is displayed in Tenhou. Sure, a higher level table means stronger players, but the rewards are also worth of it. Not only the point rewards, but the experience rewards as well.

L0000 (the free lobby) has 4 rooms for players of different levels. They are as follows:

Type Description Requirements
一般 (ippan) General All players
上級 (joukyu) Dan 1 Kyu or greater & Premium
特上 (tokujou) Upperdan 4 Dan R1800, or greater
鳳凰 (houou) Phoenix/Gigadan 7 Dan R2000, or greater

Of course, you are aiming for either Phoenix Room shorter-term, or Tenhou Rank in a far longer term.

Whenever you feel like playing, play ranked and with your account. Don’t fall in love with your stats.

It is alright to take a break once in a while and play friendly matches, and we all love starting new accounts because it feels like you are “cleansing yourself of all sins” and that “now this shows your REAL skills!” but the truth is, if you add a lot more of pressure into a single point, you will get much better results. Just keep playing ranked. Thousands of games. Never fall in love with your stat screen, for two reasons. 1, you might avoid playing because of the risk of losing. 2, they are probably worse than you think anyway. If you want to fall in love with your stats, wait until you are 10th Dan or so. Also, in order not to falter in your way and motivate yourself to play, you might want to read Pechorin’s suggestions.

 Last Place Avoidance

I already explained how avoiding last place in tenhou is vital towards ranking up or, in some cases, towards not ranking down. Leave aside any pride, all your “heroism,” and “positive thinking,” and begin being more logical and taking into account the worst possible scenarios. If you riichi nomi with a kanchan wait, and then the dealer pursuits you and you deal into a 12000 points hand, you deserve it. I can’t stress that enough.

The basic of Last Place Avoidance is ORI. If possible, never deal into the player who is last. Let him sink.

Let us remember a little bit how point exchanging work in mahjong. If someone gets a tsumo, everyone sinks below that person. Okay, cool. However, if you discard someone’s winner, you sink below everyone. Which of the two is more likely connected to getting last place? You will have to make up not only for the point difference that you originally had, but now for the point difference that the player who was fourth set on you as well. If your opponents are any good, once you fall to a solid last place, you are likely to remain last unless you find some magic sand. As you get deeper into Tenhou, the amounts of magical sand decrease. Go back into your games and see how many times you ended up last because of dealing into the last place player. You might be surprised.

 Dama is important in South Round! (Hanchan)

By the time you are into the south round, everyone’s positions are usually more or less already decided. First place usually has somewhere between 30 and 40k, and the player in last is usually between 10 and 20k. In most of your games, you will be around the middle. And now I want to talk to you about situations where you riichi looking for the hope of first place. There is nothing wrong with that, if your hand merits it, but you must look at the contents (and I also mean value) of your hand, the current score situation, and of course, your wait. That is because the player in last (and sometimes the other two as well, depending on their hand) is likely to come at you. Especially if he is the dealer. And of course, after you riichi everyone with marginal hands will already have pulled out. This means that if your wait isn’t very good, and/or your hand isn’t very expensive (pinfu nomi, tan yao nomi), you must dama.

No point in declaring riichi. None at all. Obviously, this is also true for the player across. You don’t riichi to “win harder,” and don’t be tempted by bonuses like ippatsu and 5 ura dora for an amazing comeback. Leave those for friendly matches. Needless to say, if a riichi comes in from the people below you, at least try to fold unless you are sure it is impossible. And, of course, if you are the player in last for quite a bit and you need points, you can go back and use the normal criteria for riichi (basically, anything with an extra yaku or dora up to 4 han, riichi).

When there is no risk of falling to last place, go for first!

Second place is not such a big deal in Tenhou. It is more of a consolation prize than anything. And of course, first place is pretty beneficial, while last place is just a dreadful pit full of poisonous snakes. This means that, whenever you can aim for first place without the risk of falling to last, you should go for first place! We can say the line for “risk zone” is direct mangan distance (don’t forget to count your riichi stick if you decide to riichi!).

I have very few logs in this computer and I couldn’t find a better example from a ranked game, but the point remains. Just looking at the the score (quite a great distance from last), wind seat (East!) and the quite decent starting hand we received, it is obvious that you will have to push this hand to the very end, unless it is pretty obvious that North’s winning condition (score-wise) is cleared. Take into account the other players’ reactions as well. I am not too sure about that South tanki; it was likely that someone had a pair, and I was expecting them to deal it immediately as a “safe” tile, and this would be effective under a situation where everyone is closer to each other. However, in this situation, no one has much motivation to fold. Except maybe for West, but he doesn’t look like the one holding South. In the end, South did have a pair of Souths, but held onto them and obviously (and logically) pushed to the end. Drew the fourth one. Lucky!

If possible, don’t go into All Last as the last place player.

I mean, if possible. If you won a mangan in South 3, but you are still in last place for all last, that is alright. But when you are 2000 points away from third place in South 3, don’t go and make a 1000 points trash hand to start All Last as the player in last, as it is high risk and low return. The opposite is also true, do not fall to last place right before all last! Even if the difference is 1000 points, and all you need to do to get out of last place is win any hand, just try to find a way out when the point difference is small. Nevertheless, mind you, if the point difference is wide, of course, you want to make it narrower. Just don’t do things like the following:


I understand what he was thinking. If I deal into anyone in all last, or someone gets a huge tsumo, he will be saved. Also, there are a lot of threats at the table. However, it is still possible for him to defend. It is not like it will hurt him if the oya gets a renchan. And also, there is a possibility that -I- will deal in! If he was planning on attacking with that hand, maybe he was better off keeping it concealed, at least until he adds a dora to his hand. orz

I don’t see a lot of meaning in that win. When you enter All Last in last place, you are virtually obliged to win a hand. And your theoretical odds of winning the next hand is under 25% if everyone plays the same (if you account for ryuukyoku, even lower). Of course, if you go under the pretense of pushing everything, you might rise it to some… 30% chances of winning the hand? So, 70% of the time you will be getting last place. In some alternate universe, even discarding the dealer’s winner and trying for a mangan next hand was a better idea. orz

If you care to know the result:

I would recommend folding in his place. The only other possible catastrophe if he folds is if North and West exchange points, or if someone gets a very cheap tsumo, which seems a bit unlikely. He can also see that after the riichi I pon’d 4p and pushed 2p. That should have been a cue. Perhaps he was afraid my hand was around mangan class and that I would win it.

And last place avoidance means just that. Going for first place when it is possible, and stepping it out and evading the negative point penalty when it is not. Then make logic decisions based on those grounds. It is pretty hard to acquire the right balance for that. The better you are at balancing it out with your mahjong, the higher your rank will be, of course.

Don’t go for useless renchan

You might want to use common sense there, but don’t push for a 1500 hand when you have a good lead as dealer near the end game when you could have passed the button safely. You are just asking to either get ron’d that hand, or get tsumo’d for a mangan/haneman the next one. Fukuchi Makoto, when I played him in one of my trips to Japanese mahjong parlors, was the current dealer and actually declared himself noten in a ryuukyoku when he was truly tenpai, to put me (second place player) into a tougher situation for a comeback, having one fewer round. And yes, it would have been pretty pointless for him to declare himself tenpai only for the 1500 points of noten payments that he would have gotten. If someone reached and his winning tile came out, he would have won on it, of course.

Having safe tiles is vital

Keep track of what tiles are safe against each player and make sure not to put yourself into a situation where even though you are not tenpai you are not able to fold either when a riichi shows up. It is also difficult to keep a balance between safe tiles and tile efficiency. It is important, many times, to get ride of dangerous tiles that you won’t use, early. Use your better judgment.

Before jumping to dangerous decisions, observe.

This kind of strikes against the last tip, but when you can’t really assess the situation as it currently is, don’t begin by discarding a dangerous tile that could possibly cost you the hand or even the match. Discard a safe tile, and observe and assess the situation. If your hand then comes together good and well, and you have assessed that it is not as bad of a discard, you may proceed. This, as well, is a problem of balance. But remember that the ultimate goal is to decrease the amount of times you deal in and improve your defense.

For tonpuusen, it is important to take the initiative, and defend if you lose it.

For players who prefer tonpuusen, remember that every match is a match with a score situation, and winning conditions. Much more than hanchan, because of the different in length. Dealing into a single mangan will, in most cases, cost you the game as it is. The basic strategy is attempt to take the initiative, hit tenpai before anyone else (lower hand values have a higher significance. 3900 point hands are ideal) and just be prepeared to defend if someone else seems about to outdo you. It is hit and retreat, and feels a lot like sword fighting. In any case, I don’t suggest you choose tonpuusen for ranking up until you have a very good knowledge of how to play a hanchan’s endgame.

 What to do when you are already in last place?

Attack! You are the reason why people may choose not to riichi, the person everyone is more likely to defend to. Because no one wants to be you! You are in a pretty tough spot, but if you look at it from the other side, you have all to win and nothing to lose. It doesn’t mean that you should be reckless, but your judgement of whether to attack or defend should start going more and more towards the attacking side as fewer hands remain. And when you are dealer, show no mercy, renchan is your first priority. Comebacks are much more easy from that position. Don’t lose track of your distance to third place, and how many points you need to escalate him. Especially in all last.  And this is imporant: don’t let third place renchan. If second is just as close, don’t let him either. Once everyone else got an oya mangan, you will be in a hole where not even Akagi himself will be able to rescue you from.

Persevere and Study

Never give up, it takes thousands of games to actually see any progress. As with everything, tenhou is also a matter of trial and error if you take it seriously. Study your games and why you lose/win, and try to understand why others do things the way they do. Especially the better players. You can look at many of the logs of ASAPIN, the first human being to ever achieve tenhou rank, and his quest for perfecting last place avoidance here.

38 thoughts on “Tenhou Secrets

  1. I am pretty sure I am forgetting about lots of other things. Like, how to proceed in West round, but let us leave that for a possible sequel.

  2. I was looking at the very first log of ASAPIN (2009-08-12 21:37) on the list, and I am wondering about something.

    In the honba of the 3rd East Round, on the 11th step, why did he choose to discard the 3s over the 2s? The 2s was a Genbutsu and it was also the last one, as the two others were already out. The 3s on the other hand was a live tile. If he decides to break up his ryan-men wait, why didn’t he discard the 2s first? A mistake?

    Just recently found your blog and must say it’s very entertaining and informative. Keep it up :)

    1. That wasn’t ASAPIN, it was うめ干しばばあ. ASAPIN was the dealer for that round. These logs aren’t all sorted in first person view; they all begin from the perspective of the first East player.

      And yeah, I can think of no other reason than a misclick. If his hand was any good, then you could say he was getting rid of it before someone /else/ got tenpai, but this is not worth of it and it is also too risky. Nice catch.

      1. Ah, I see. That makes sense then. I didn’t watch the full game and just assumed ASAPIN would be 私 in all the logs.
        It’s not a big deal then, but I thought I might have missed something, since I am a beginner and sometimes can’t really understand some of the discards the pro’s make.

        Thanks for your help!

  3. Nice article there as always :) Btw I think this strategy is somewhat universal to all types of mahjong sites, so it’s a bit misleading to name it Tenhou Secrets.
    Oh and before i forget i have some questions.
    -Do you plan to make a pdf from all of the tips and stuff on the site? It would be mutch appriciated, and ppl could just read it on an ebook reader or print it out.
    -Also I’m somewhat fascinated by occult mahjong. Do you plant to (or anyone else on the site) write something about occult or you are too strictly digital to even consider that ? :)

    1. It is not misleading, really. And it doesn’t apply to all sites. It is just the huge gap between third and fourth that tenhou has that makes playing for non-last an important aspect, even if it means renouncing some chances for second in many occasions. Servers like Janryuumon have a different point system, where all four placings are balanced (during Janryuumon 2, it was 200, 100, -100, -200) and it becomes “rentai mahjong” rather than “last kaihi” mahjong.

      I have plans on making a book with content from all the articles I wrote, and some new (original) material and publishing it on Kindle for e-book readers.

      Occult mahjong doesn’t even exist in reality. It is just an excuse to justify things that happened in a match without much of a reason. No occult theory can be used while you are playing. Only after you are done.

      1. A book certainly sounds good, because riichi mahjong material is just hard to come by (at least materials with english language).
        As for blaming occult things like the flow of the game after you lost is just cheap. The whole mentality of occult mahjong has deep flaws, but things like “i get the most points out of wins when im behinds everyone” mentality helped me a lot for pulling though harsh rounds where i was beyond 10K or more and even win.
        There is some kind of flow of luck even its just in peoples heads and not in the tiles. Guess i have to ask around in japan for the general beliefs beyond the normal “nagare” stuff. I just find it somewhat fascinating.
        And last but not least good luck with the book and keep us posted on the release date.

        1. Thanks! I will.

          I don’t mean it in the sense of blaming it on luck when you lose.
          Most “occult theories” are based on “kekka ron” or “result-biased theory.” A theory that is “proven” true only by the immediate or direct result in one single event, or a small group of events, and where the person is expected to believe in it for it to be true. Mahjong movies, anime and manga encourage people to trust their heart to superstitions, but the reality is that, just like with God and agnosticism, even if you could prove such a thing exists, you have no use for it in a game and thus it is irrelevant.

          Moreover, there is not a single “occult” theory that statistics haven’t been able to debunk. In fact, they are sub-standard in comparison to digital playing. And that is explained by their lack of reasoning.

          I could throw at you a bunch of “occult” one-liners, but it is up to you whether to use them or believe in them.

          -When you haven’t been having good draws, in a bad day, it is easier to go for chanta hands.
          -You are likely to draw a pair of 7, 8 or 9, if you have a pair of 4, 5 or 6 respectively.
          -If you win off the “lucky” player at the table, you will steal his luck.
          -If you don’t make use of your “luck” at your lucky moments, or if someone wins with one of your discards, you will lose it until you steal it again from someone.
          -When the “lucky” player declares a riichi, call any tile that turn. He is likely to get ippatsu tsumo.
          -In a bad day, the tile you draw immediately after a riichi is usually the opponent’s winning tile.
          -Luck levels in mahjong aren’t relative. They are absolute.
          -You are much more likely to get ippatsu tsumo the round after you win a hand.
          -If you call tiles in a lucky day, you will be depriving yourself of great draws, wasting your luck away.
          -If you concentrate, you can “feel” when others are tenpai just by the thickness of the air.

          Also, Tetsuya wrote a book about occult mahjong. Kojima’s book have a lot of occult talking in them as well.

  4. I love reading your articles. I’ve been playing Janryuumon as of late. Is either program better than the other? Just curious.

    Also, I love that background….where can we get it?

    Last but not least. Other than Tenhou, is ther any other Mahjong program you reccomend?

    Again, thanks for the articles…..love to read them!

    1. Hey. You are welcome.

      Janryuumon is, originally, beginner oriented (notice also the similarity with “Jannyuumon,” introduction to mahjong). And now that they took out the higher room more than ever, there isn’t a very competitive circle in it. In other words, people are playing for fun, and the vast majority isn’t very good. The system for ranking up isn’t really complicated either. It is ironic that leveling up is a lot harder at kyu levels than at dan levels.

      Tenhou’s higher levels are very competitive, and the players are very good. But I wouldn’t call it “fun” in a wider meaning of the word. There is some sort of “community” for the higher levels, but between you and me, most of them aren’t very friendly, as opposed to Janryuumon.

      So, if you ask me which is better, it depends on what you want to get out of the game.

      Here you go http://i.imgur.com/J6zxo.jpg

      If I had the necessary money, lots of it, I would play in Marujan.

      You’re welcome!

  5. Oi, money isn’t in abundance for me atm. Its one reason I started with Janryuumon, its free…..unlike the ipad version which you have to buy points to play with real people.

    I’m attempting to learn Japanese, its been very slow going but I love the language and the culture. Anime, Manga, and Drama’s I just can’t get enough of. Then thanks to the anime Saki I feel in love with Mahjong. Its very fun and can provide many levels of depth depening on how far you want to go. I’m hopeing to become fluent so that I can enjoy more of these things with out the loss that sometimes occurs with translation.

    Thanks a buch for the image, now to go figure out how to apply it to Tenhou! :)

    See you around :)

    1. The above post was suppose to be a reply to your last message but I apparently didn’t hit ‘reply’ sorry about that.

    2. Yeah, me neither, that’s why I play in free servers. You can see my progress at my blog, as it is linked in my comment information here.

      “Mahjong” Japanese language is a completely different universe from normal Japanese. It has its own vocabulary, too. Most translators are at loss when there are mahjong scenes. Even those for official translations and broadcasting of anime. Detective Conan’s mahjong episode was just unintelligible.

      How are you trying to attempting to learn it? Contact me by skype (x_kime), I might be able to help you.

      You need to download and use the windows client, and select it in the settings. Download, log in, and use are completely free, but you have to be a paying tenhou member to actually play with it.

      See ya!

      1. Don’t mind my terrible grammar in my last two comments. My brain is not with me at the moment.

        I have been doing some major cleaning at my house for the last two days and I am exhausted.

        1. Thanks again and no worries. I totally understand what being exhausted can do to ones abilities. :)

          I’ve never used skype but if it requires a mic I don’t have one. Its chaos at this house and I doubt you would be able to hear me much over the background noise.

          As for learning, I have been useing Jpod 101, I have the 1st edition of Genki. I happened by an old copy of Pimselur too. I seem to grasp somethings but others tend not to stick. I wouldn’t be opposed to using a chat program or email even.

          One thing I’ve been looking for is a try translation of the rules for Richii Mahjong. Some sources seem shadey to me (such as the tutorial for the Janryuumon for the ipad). Im not even sure if Underwater- Mahjong’s subs are accurate. But I could swear that I read somewhere they were.

          Again, thank you for taking the time to reply back with me and I hope to hear from you again. :)

          1. “I totally understand what being exhausted can do to ones abilities. :)”

            And we’re back to the topic. At a high level, you should not play while tired, sick or otherwise incapable of thinking well. Or you could try it once and observe how your awareness goes down. Being tired doesn’t only worsen your skill, it also makes your reactions to losing far more damaging to your mindset.

            I’d also like to state that when your rank goes up, the importance between second and third place increases greatly. When aiming to get past the 1800 threshold, the difference between -2 rate and +2 rate is very meaningful. And once you’re in the upperdan lobby, the +30 second place prize is quite good for estabilishing an upward trend.

          2. You’re welcome.

            As for skype, I don’t meanit as a way to use a microphone or video, but as an instant messaging application common to almost everyone. At least, it is the one I am currently using for personal communication. So, I would suggest you should try to contact me there on the issue.

            I have written a book on that in Spanish, and I am currently trying my best to rewrite it into English, but I don’t have that much time lately, moving and all. Also, I am not even sure it will sell copies, as the Spanish version only ever sold one copy.

            But it feature a solid, vast collection of rules.

          3. Thanks again, got Skype and sent you a friend request. Hopefully I did everything correctly.

  6. @Iapetus

    It is true on not playing while tired, or sick, but then again, if you are tired you are not likely to want to play in the first place. However, on your next comment:

    >I’d also like to state that […] for estabilishing an upward trend.

    When avoiding last place, that importance between second and third is trivial. The 1800 threshold is silly, and any good player will surpass it not only naturally but quite promptly without the need of thinking whether they are playing for second or for third. I have made quite a few subs, and all of them stay over R1800 before even hitting 3-4 dan. Trying to fish for two points of rate, giving into risk and falling to last place while trying is not only pathetic, but also substandard. While we are playing in tenhou and not jansou, let’s focus on not losing rank points; Rate will always come naturally when you play well (as will first and second places, or rentai). I have yet to meet a person that can achieve a certain rank, but can’t get past the Rate barrier, and if such a person existed, they probably reached that rank by variance and not skill. A 4 Dan with rate 1700 is not really 4D and will promptly fall back to third and never climb to 5th.

    On the second place +30, you are certainly over-estimating it, and when you play in upper dan for a bit longer you will realize that it is also trivial. It certainly makes you happy to end in second and get a little boost, but it is not worth losing your last place avoidance over. And it will become even more obvious once you reach 6th Dan. Even when I was at 5th, I found it quite trivial, but upon reaching 6th it became just ridiculous. The point difference between 1st and 2nd is 40 points, between 2nd and 3rd it is barely 30 points, but the difference between 3rd and 4th is 120 points. It doesn’t take a mathematician to see that the gap between 3rd and 2nd is the smallest, while the gap between 3rd and 4th is, by far, the widest. Manyfold.

    I’m not saying protecting a 2nd place is pointless, as while in-game it is one of the best positions to be (it guards you perfectly against falling to last), but if you take risks of falling to last because of protecting your second place, you obviously shouldn’t play in tenhou.

    1. This comment isn’t completely aimed at you personally, as I am pretty sure you were aware of this already, and that you didn’t mean it that way, but your comment seemed so misleading that I had to clarify it for anyone reading.

      1. Indeed, let’s not make things too complicated too quickly – simplicity gets one far. But if the reward/punishment ratio is 1:4, then taking your chances when the chance/risk ratio is over 4:1 is the right move. Mahjong is all about improving upon the trivial things that are least trivial.

        For new accounts made by good people, 1800 is trivial. But not everyone starts that way. My account had my entire mahjong history weighing down on it – I was at 1300 at one point. Debt from the kyu ages can hurt.

        1. That is why it its stated that, as far as there is no risk for last, you should play for first. Of course second place is a secondary effect of playing for first. Seldom will you find a +4:1 situation for second place, and seldom will you be able to calculate it accurately even if you happened to find it by chance. Simple is best.

          Those debts can be paid quickly, even in a day, by making a new account. I don’t know how hurting can be to go down like that; my first tenhou/online account never even went under R1500, fortunately; but I am pretty sure it is possible to get into joukyuu with R1800 in a day, and to tokujou in a few weeks.

          As to why I never went under the average in tenhou in my first try, hours of days of gamedesign’s flash game before actually playing online. w
          I didn’t want to be a hindrance to online players. :(

  7. Being a hindrance online is nowhere near the feeling when you beat your poor friends with a totally disgusting comeback win :)
    Also thoose occult one-liners are mostly funny concepts like stealing luck and such. Though calling a tile after someone “lucky” delcares a riichi is not that bad strategy as long as its not becoming a bad habit.
    And yea i agree that tiles wont magically shift in the wall but mahjong has this feeling that the whole game can “betray” you. Like throwing away a single tile because there is already 1 in the pond, just to get the same tile when you draw.
    I agree that statistics go a long way, but playing without any risk on pure strategy makes for some really stale games (there are a bunch of theese that anyone can find online and frankly i find it sleep enducing when someone plays safe every damn turn….). But then again im bad at playing so who cares..heh!

        1. “or less” would have still been technically wrong. It would mean 4dR1800 inclusive, when it is in fact exclusive.

          Like osamu said, “below” is the correct term.

          1. I’m a beginner at mahjong and can’t read Japanese at all but I’m pretty sure I play > 4 dan before….

            But then again sometimes I win without even knowing what yaku I have lol *beginer’s luck*

            Thank you for the answer :)

    1. Kinda late response but it’s tenhou.net ! I think it’s safe to say it’s the most popular place to play online riichi mahjong these days. You can play for free wIth the browser version at the URL or pay to play with the fancier windows client.

Leave a Reply

Your email address will not be published.