何をuld you 切る?

UPDATE:

Due to being busylazy, we haven’t published the answers to the quiz until now. We apologize for the inconvenience to our dear readers. orz
You can find the answers to the quiz now at the end of this article.
Thanks to Ben for providing us with the answers!

As some of you probably know, the ECR (European Riichi Championship) Tournament, which my humble self attended to, was held for the second time after 2 years during the last weekend in Hannover, Germany.
Expect a more detailed write-up of my experiences and antics during those 3 days once I un-lazy a bit more.

On Friday evening, prior to the actual games on Saturday and Sunday, a 3 hour strategy workshop was being held, which is the actual focus of this post.

The workshop was planned and executed by none other than Takunori Kajimoto himself, member of the Mahjong Museum Committee, Pro-Mahjong TV-commentator and practically a Mahjong-Legend in general. Besides him, as the main speaker and voice (translator) of Kaji was -the among gaijins no less legendary- Mahjong Researcher Benjamin Boas. Both of them also served as head-referees for the tournament. Another Japanese Pro and actual player in the tournament, Mr. Y.T. also helped out here.

The venue was the upper floor of the jazz club, in a large, kind of theater-stage like room, while the actual games were played in the basement of mentioned club. About ~50 poeple in total attended the workshop.

The workshop consisted of a general discussion of several topics you can see written down on the easel below, a discard quiz, a showmatch with Kajimoto (and explanations/Q&A after each hand), as well as a closing Q&A session for general strategy-related questions.

The other 3 players of the show match were poeple that had the highest number of correct answers in the quiz.

As you can see, it turned out to be quite the Ponpalace.

While I have to say that I didn’t exactly learn anything mind-blowing new from thr whole thing, some discussions that came up and some points were certainly interesting and got me to consider some stuff in a different light. Unfortunately, there was no summary handout or something about the topics that were being discussed.

If you’re looking for expert advice, why don’t you check out TRA’s Tips or mou’s advanced strategies?

Kaji also signed strategy books that were being sold (I also got myself one) and mahjong set cases if you asked him, here’s a case of a Mahjong-mate of mine.

Lorizean, another denizen of #mahjong got himself a small Akagi Washizu-Mahjong set too.

So without further ado, let me come to the main point of this post: The Quiz. Everyone got a sheet of paper with the hands below and was to mark what to discard and if you riichi in the last 3 hands. It took many, including me, quite some time to think about it before finishing. Ben and Kaji also said the quiz was supposed to be quite difficult, as these questions were directed at Pros.

Now what do you think? Drop your answers and reasoning in a comment below. I will reveal the answers in about a week. I will also try to get in touch with Ben about actual explanations of the answers.

Note:
They didn’t specify whether the quiz was meant to be considered with kuitan ari or nashi (EMA) rules. I actually suppose it’s with kuitan ari, since they said these questions were aimed at actual Pros in Japan, where open tanyao is the norm. Also, note that the Dora: tile is the actual dora and not the indicator.

Hand 1

East 1st round, being south seat – Turn 5

3m4m6m8m8m5s7s7s2p4p4p7p8p 9p Dora: 6p

Hand 2

East 1st round, being south seat – Turn 5

1m1m3m5m4s4s6s3p4p6p7p2z2z 8p Dora: 4s

Hand 3

East 1st round, being south seat – Turn 5

7m8m3s5s5s7s8s3p5p5p7p7p9p 9s Dora: 2m

Hand 4

East 1st round, being south seat – Turn 5

1m1m3m4s4s6s1p2p3p3p5p7p7p 2p Dora: 8s

Hand 5

East 1st round, being east seat – Turn 8

2m3m4m4p5p6p7p2s3s7s8s8s9s 8p Dora: 6z

Hand 6

East 1st round, being east seat – Turn 8

1m1m2m4p4p5p6p7p6s7s7s8s9s 9p Dora: 7s

Hand 7

East 1st round, being east seat – Turn 8

3m5m6p6p7p9p9p7s8s9s9s1z1z 8p Dora: 9s

Hand 8

East 1st round, being east seat – Turn 8

2m4m6m2p3p4p5p5p3s4s6s6s6s 7s Dora: 6z

Hand 9

East 1st round, being east seat – Turn 8

2m4m6m3p4p6p7p8p5s6s7s8s8s 8s Dora: 6z

Hand 10

East 1st round, being east seat – Turn 8

1m1m6m7m8m5p5p7p1s1s1s7s8s 9s Dora: 6z
Riichi? Y/N

Hand 11

East 1st round, being east seat – Turn 8

2m2m4m3p4p4p4p5p1s1s1s2s3s 4s Dora: 6z
Riichi? Y/N

Hand 12

East 1st round, being east seat – Turn 8

2m3m4m5p6p7p8p4s4s5s5s6s7s 8s Dora: 6z
Riichi? Y/N

The Answers

Unfortunately, our main “osamuko translator” was too busy to work on the explanations which accompanied the document Ben sent, so here are just the raw answers.
You can find the original answers document with explanations here. ((´Д`;)Osamu orz)
Feel free to translate them if your 日本語 is good enough and share your translation with the osamuko folks. If you are struggling with some of the weird Mahjong-specific terms, check UmaiKeikis dictionary or drop a comment.

Actually, faithful Osamuko follower and bronze-member xKime already translated the whole thing in like 2 days. You get grab it here.
Thank’s a bunch! ( ´・ω・)つ

Hand 1: 7s

Hand 2: 5m

Hand 3: 7p

Hand 4: 3m

Hand 5: 8s

Hand 6: 2m

Hand 7: 6p

Hand 8: 7s

Hand 9: 2m

Hand 10: 7p and Riichi!

Hand 11: 4pwithout Riichi.

Hand 12: 4swithout Riichi.

30 thoughts on “何をuld you 切る?

  1. 1: get rid of that 5s, could be dangerous later
    2: 6s
    3: 7p, go for san shoku doujun and call everything needed
    4: 3m, chitoitsu
    5: 9s, go for tanyao
    6: 9p, go for pinfu
    7: 3m, 5m could be dangerous
    8: 7s of course, tanyao
    9: 2m should be safest, leave 8s for later
    10: 7p, ones are always a good wait, riichi
    11: 2m, no riichi, try a san shoku doujun
    12: 8m, no riichi, i would go for an iipeikou tanyao with tanki wait according to discards

    nächstes mal komm ich mit :3

  2. Some of these will probably be really bad decisions, but who knows – I decide differently under stress (probably even worse than this).

    hand 1:
    9p
    Aiming for tanyao (fast win -> get dealer), or toitoi/chiitoitsu if it comes, with the dora as a lucky bonus if it comes – if not, I will have to get rid of the 78p. It’s still the beginning of the game, so no stress.

    hand 2:
    6s
    Pon the south winds + doras, or keep the doras as pair and pon the 1m. Generally aim for a fast win again, with the 678p, 34p and the my seat wind pon.

    hand 3:
    5p
    Sanshoku doujun. Hopefully. The 5p leaves me a wait for 4p or 6p, will be using the 5s as a pair. Will try to keep it menzen as much as possible. If it gets hot, call stuff.

    hand 4:
    3m
    Hand is pretty shit, so chiitoitsu / toitoi. Probably ponpalace after some time, if it doesn’t go well. I get rid of the 3p, because the 6s can get a 7s for a dora wait – but only if I change the hand for the iipeikou (112233p)

    hand 5:
    9s
    Since I’m east, I’ll try to get as much as possible out of this – so riichi, pinfu, tanyao. Will need the 8s as pairs. Will keep the 7s first, because I could get 6s which would be a better tanyao wait than 23s

    hand 6:
    2m
    I’ll try to go for either pinfu or a shitty riichi – with the 3 dora it’s pretty cool. But I will need the 8p or a 3p / 5p. So I’ll keep the 1m as pair backup and hope shit doesn’t screw itself up.

    hand 7:
    3m
    I’ll ditch the mans, waiting for the east winds etc. Maybe even a toitoi if something comes up.

    hand 8:
    5p
    The 234 sanshoku looks promising, but I’ll keep my options open with the 246m just in case.

    hand 9:
    8s
    And keep waiting for those mans.

    hand 10:
    5p
    no riichi
    Need something better, so I’ll just keep waiting.

    hand 11:
    2m
    no riichi
    And hoping for the 2p

    hand 12:
    8s
    no riichi
    I want my iipeikou!

  3. I cheated, I used an EV calculator (#^_^) . I would probably answered like Sollniss for the first 4 hands.

    1) 2p
    2) Nan (South)
    3) 3s
    4) 3m
    5) 9s
    6) 9p
    7) 5m
    8) 7s
    9) 2m (or 6m, doesn’t matter), then 8s (riichi if first)
    10) if first, 5p and riichi; if second 8s and no riichi
    11) 2m, no
    12) 8p, no

  4. Instead of “stating the obvious choice” I will try to be radical here, and pull a few not-so-conventional ideas, expecting to get 0/12.

    1) 2p. There are many xyy/kan-toitsu taatsu already, and discarding 2p still leaves space for
    pinfu, chiitoitsu, or maybe tanyao with the dora if it were to come.
    It is easier to upgrade the other taatsu to ryanmen, too. If they became
    ankou, I wouldn’t complain either.

    2) 6s. Because even upong drawing a 5s, there’s no discarding the dora anyway. Upon
    a south pon, it is easier to discard 1m and have the 135 ryankan. Perhaps we can
    even use reverse method, and discard 1m first. But 6s is close to the dora, so it might
    get dangerous later on.

    3) 3s. As 5s is more likely to become atama rather than the taatsu around pinzu, which
    is likely to become san shoku (which is what should be aimed for, and also there are other
    interesting taatsu there. 3557(79) is more versatile than leaving a 35s kanchan and wasting the pinzu taatsu as atama, or discarding a pinzu and losing the multiple taatsu.

    4) 5p. If the hand were to go towards a pinfu with iipeikou or similar hand, 7p is likely to
    become the head. There’s no discarding 6s yet because drawing a dora would result in a
    ryankan. And this way, no pair is cut just yet, in case of chii toi tsu tenpai. The thing is,
    without drawing a dora or without riichi or tsumo, this chii toitsu is pretty much meaningless
    and I would take iipeikou and pinfu if I had the chance to go for it (even if it can fail and
    leave me with a kanchan), still leaving room for a wait on the dora if a draw a 7s, or
    incorporating the dora upon drawing it.

    5) 9s. Then 7s later. Towards mentanpin, with the jantou 8s and sanmenmachi in pinzu.

    6) 2m. It’s easier to get a ryanmen from 4p and go for pinfu with dora. 1m as atama is
    better for evolving the hand’s waits. Just drawing a 8p
    already “grants” a ryanmen and pinfu, and if not we can benefit from pretty much any pinzu.

    7) 9p. But I’m not sure. East koutsu and two dora are the priority, as this gives dealer mangan.
    Either the dora can be toitsu, or 6p can be toitsu. If it went really bad, it could end up
    as shanpon wait on east, too. So, 9p toitsu is not a great loss. 35m is not to be discarded because
    it is likely to backfire with quite a few tiles. 6p could go, too, and if it became a shanpon
    between 9p and dora or east, it could become a nice hikkake riichi. But I wouldn’t like
    to let other players have access to calling a middle tile.

    8) 5p. Let’s go back a little and improve the hand. Because I like mentanpin with a dora and sanmenmachi (either the pinzu one or the souzi one) more than a tan yao with a
    hikkake riichi, simply put. More tiles are availabe for calling too, if things went that bad. But
    I think it’s a matter of style. 7s is not so bad, given that we are east; it would be the fastest
    way, and the remaining ryanmen or hikkake aren’t bad either.

    9) 8s from within the hand. Tan yao with pinfu or hikkake are visible. I wouldn’t tsumokiri
    though, discard the same tile from within the hand and cause at least some “confusion”
    and fake/meaningless suji and/or information.

    10) 5p towards upgrading to junchan sanshiki and/or ryanmen. No riichi just yet. I wouldn’t
    like to engage into tsumo-kiri every tile with riichi nomi and an uneffective wait, seeing
    the opponents might soon reach tenpai too, and this hand is not so bad for defense.
    Declare tsumo if tsumo. The only backfire here would be losing the turn as dealer, but the
    biggest reward is jun chan, sanshiki, pinfu. Just two draws away.

    11) 2m without riichi. Wait the sanshoku or ryanmen. As before, an uneffective wait for riichi
    nomi is not worth having to tsumokiri every tile. Still, declare tsumo if it happens.

    12) 8p. Seeing that we are dealer and riichi tan yao (tsumo/bonus) is not that bad either,
    and it’s not so necessary to wait to upgrade to ryanmen, it’s not bad to take the chance
    at a swift attack here and riichi while being dealer with a good hand. It can backfire by drawing any other souzi tile, but I’m pretty sure ron’ing on a tanyao nomi or getting the tsumo would be the weaker point for this hand (assuming most people don’t let the ron/tsumo button escape).

    Without further info on each situation, this is as much as I can guess.

  5. Just to share my own opinions about the quiz…I am weak in tile efficiency orz
    1. 6m
    A sort of advanced tile efficiency theory here. 68m, 57s, 24s seems the same here, but if you count on the further reshape possibility, for mantsu, 5m does not give you a good shape since the 34m need this. This implies that 688m is the weakest taatsu.

    2. 6s
    since you will not be too happy when you draw 5s (because you need to discard the dora 4s)

    3. 7p
    siimilar argument as Q1…you need 8p for your 79p already

    4. 5p
    another similar argument as Q1

    5. 8s
    being east seat, keeping the best probability of tenpai is the key. and discarding 79s will not guarantee a tanyao here.

    6. 9p
    the key is to get rid of the bad shape of this dora hand. even though later get furiten by drawing 8p, the shape is still much better than 112m…here I would keep the tenpai probability at maximum.

    7. 9p
    smelled a sort of atotuke here. If your left player discard 4m, I will chi and tenpai. For atotuke, keeping the another pair being more difficult to get out is a common strategy (to minimize the change of being furiten)

    8. 7s
    I don’t see any reason to go back to 2-shan-ten for this hand, it’s already round 8

    9. 2m
    discarding 2m gives you the most efficient hand

    10. 7p reach
    not many tiles to improve the shapes here, instant reach hand for oya.
    It is easy to get 2600 all by ura.

    11. 4p
    Although instant reach is not a bad choice here

    12. 8s
    again, it is also good to reach for east seat.

  6. hand1 2p
    hand2 1m
    南を安牌として残したい。
    it will use minami anpai(if others reached)

    hand3 7p
    789の三色を作りたい。面子オーバー。
    make 789sansyoku. 355s 355p 79p.

    hand4 3m
    七対子イーシャンテンだが面子手を見たい。
    reach for tiitoi,but it will be iipeiko and pinfu.
    ex. 4467s 112233p 567p.
    11m44s122335577p. (5p aka)

    hand5 8s
    hand6 2m
    hand7 9p
    hand8 7s
    hand9 2m
    you can get 7m,aka5m

    hand10 7p reach
    hand11 as you like.
    hand12 8s? or reach.

  7. In my expert opinion:

    1: 4p now in case people hoarding the dora make a 47p wait later. Aim for sanshoku.

    2: Since there are several pairs already, we don’t need that extra 1m. Also, if we make a sanshoku, we’ll probably use one 4s but not both, so cut one early before people build a chance to call/ron it.

    3: Cut a 5p to wait for 46p, keep the extra 7p for now so we can make sanshoku.

    4: 6s, as it won’t be of much use in going for sanshoku.

    5: 7s, so we can focus on making shuntsu and getting sanshoku.

    6: 2m. Then if we draw 789m, cut a 4p and go for sanshoku. If we draw 567m, cut the 9p and go for sanshoku.

    7: Cut 9s, then cut 9p, then ron 9m for sanshoku.

    8: 7s, because it can’t be used in either sanshoku made by the anticipated 3m or 5m draw.

    9: 5s, then draw 78m for sanshoku.

    10: 7p, riichi, and ron someone playing “safe” 1m. Without sanshoku, what else is there to do?

    11: 2m and no riichi, hope for a 2p to come so we can make sanshoku.

    12: 8p and riichi to make up for the points we missed by not getting sanshoku.

  8. Here are my autopilot answers, because in a real game there is a time limit. This is just my excuse for not needing to think though.
    1: 8m
    2: 1m
    3: 5p
    4: 3m
    5: 8s
    6: 9p
    7: 9s
    8: 7s
    9: 2m
    10: y, 7p
    11: n, 4p
    12: n, 8s

    I await my failing grade -3-

  9. Arkhangel’sk what do you mean by EV calculator? Do you have a program that calculates the EV of each potential discard? If so, I’d love to see the details of its output.

  10. Yes, it computes the tree of all hands values up to a depth of four discards (3-shanten). The output is too long to published here… several pages. Many simplifications hypothesis are taken, so the end result is not the exact EV, but a fairly good approximation of it. Enough to notice that many 何を切る?quizzes don’t have clear answers. They are too close to call. Your guess is as good as mine. Often, the actual answer don’t really matter. What matters is the reasoning… if there is any given by the author of the quiz.

    For instance, for hand #1, 2p, 5s, 7s and 4p are all almost equivalent answers. Only a 1% difference between the EV of those four discards. TRA’s answer (8m) is the second worst answer possible, just slightly better than 7m. Puyo should also revise his analysis.

  11. @ron5

    Tenhou/2 doesn’t compute EV. It only shows you the maximum waits count. independently of hand value.

  12. I would like access to this godly calculator.

    Also, I feel a little better about my choice for the first discard.

    Eh, furthermore…

    >TRA’s answer (8m) is the second worst answer possible, just slightly better than 7m.

    7m is a terrible choice. You should never cut a tile you don’t have. That’s a chombo, I think.

  13. Sorry about that confusion. In my region of the Universe : p=s, s=b, m=c

    8m is ok.

    The two worst are 7p and 8p.

  14. pin: Sapek, illegitimate love child of Spock. (…sapèque: money coin with a hole in it)
    sou: Bambou, the thing never translated right by anyone
    man: Caractères, asian squiggly things.

  15. I could probably translate the explanations, but the pictures of the tiles are missing when I view the posted doc so it would be pointless and nonsensical. If there is another copy with the tiles or if there was some way I could view the tiles then I would do it. It would be nice to read the reasoning for me, because I’m a total beginner.

  16. Nice quiz, my answers is as follows:
    1. 2p [WRONG]
    2. 5m
    3. 7p
    4. 3m
    5. 8s
    6. 9p [WRONG]
    7. 6p
    8. 7s
    9. 2m
    10. 7p riichi
    11. 2m no riichi [WRONG]
    12. 4s no riichi

    I was wrong in the 1st (lol), 6th, and 11th hand…

    based on the answer osamu’ gives I can clearly understand why was my 1st hand is not an efficient cut, yet I found the official answer in the 6th and 11th hands is quite “intriguing”, oh maybe after your official translator back you can make another blogpost to explain the reasoning…

    Still; in the 11th hand, people like puyo and TRA guess it right, so maybe they can explain?

    Thanks for the quiz!

  17. Thanks Osamu.(´・ω・`)

    I also remember one Question having 2 possible correct answers. Might have been question 2 because it doesn’t have a proper 答え line unlike all the other answers.

  18. Question 2 indeed has two possible answers. “5m” and “6s.”

    Also, the reasoning for 5m is actually quite good. I didn’t take into account that there were red 5’s involved either.

    If it’s aka-nashi, 6s all the way.

  19. I have inspected the answers, and the “red 5’s consideration” seems to be an important side in quite a few problems, not just that one. If only I had considered there were red 5’s involved in the equation, I would have answered a bit differently in some cases, such as the fourth or fifth problem. But, oh well, I just assumed, but it would have been good to know.

    Still, this quiz is amazing. The problems seem more constructed than casual, and there is a theme behind every each one of theme. The answers alone are some kind of “guide to intermediate theory.”

    I wish I could get more of this kind.

  20. I had free time, so I translated the answers for your viewing pleasure.

    From these answers I can see that some of my answers weren’t completely “idiotic,” but I just wasn’t considering red 5’s, or was selecting the “radical” or “subordinate” answer/reasoning. I like to think I’m getting closer.

    http://www.mediafire.com/?resqmccq87028km

    It’s a PDF. If someone prefers a .txt, a .html, a .doc or anything similar, contact me directly.

    It’s a dynamic translation. I guess could have been more literal, but it would have been pointless. I mean, if you can understand terms like “tenpane” you probably don’t need a translation at all. I have trouble expressing myself here and there, but the reasonings behind the answers are all there. English and Japanese aren’t my native languages, so yeah.

    Enjoy.

  21. xKime, good job. I started a translation but got bogged down, espeically on question 5 which was a bit hard for me to follow. I like how you arranged the tiles and and such, it’s easier to follow than the raw Japanese where everything is run together.

    One quibble, in question 1 he states that 3 pairs is the weakest shape, not the strongest.

  22. Some questions were just a pain, because the narrator hits around the bushes a few times. And yeah, I arranged them because the raw confused me so much it would have been impossible to work that way.

    Ah, you are right. “つまり、メンゼン手は3ヘッドが一番弱い形となる” wwwww
    Sorry, orz.

    Fixed it: http://www.mediafire.com/?57bkm6enk225he7

    I can draw the conclussion the writer is a red 5’s freak, and I like that.

  23. “I can draw the conclussion the writer is a red 5’s freak, and I like that.”

    If you’re playing with red 5s, being a ‘freak’ makes you win :P

Leave a Reply

Your email address will not be published.