Thumb up
1 Posts

Riichi Mahjong» Forums » Strategy

Subject: Riichi Mahjong Strategy rss

Your Tags: Add tags
Popular Tags: [View All]
msg tools
Riichi Mahjong is the name given to the Japanese ruleset of Mahjong, a game in the rummy family, similar to Rummy, Gin Rummy, Michigan Rummy, etc. As a game, its largest selling point is popularity. You can always find a game online at, and 雀荘 (Mahjong Parlors) are ubiquitous throughout Japan, featuring 雀卓 (Mahjong Tables) which perform automatic shuffling of tiles and game preparation, making it quite easy to play many games in a short span of time. As a warning however, most Mahjong Parlors are unfortunately uninviting to women. Even most of the ones specifically advertised as "Women Friendly" are not places most people would feel comfortable without a personal invitation and group to go with. Furthermore, if you go to a Mahjong Parlor without the requisite four (or three if playing Western Japan style) players required to play, your only option will likely be "on the books". I haven't been to a Mahjong Parlor since gambling was legalized in Japan, but I imagine the system has not changed terribly. "The books" provide a method of keeping track of score and often implies a gentlemen's agreement of settling the difference via a ___ yen per point "unofficial" gambling exchange of money. The house doesn't really profit from this unless tables are short players (at which point employees join to fill empty spots), but does take an hourly fee. Finally, Mahjong Parlors cannot legally operate 24 hours, so they often have lock-ins at night and block windows / doors. Most do, illegally, still operate.

The shortcomings of Mahjong are similar to its rummy cousins. The probabilities, once understood, do not change, at which point decisions become automatic and uninteresting. Perhaps most unfortunately, despite elements of interaction (ability to see all opponents' discarded tiles at any time, ability to 鳴く "call" and steal tiles to 上がる "go out" more quickly often at a slight loss of score, and a critical decision of when to 降りる "switch to defensive" for fear of being the person who gives another player their winning tile), statistical analysis of the vast game data available online reveal rather fixed strategies where other players' decisions fit nicely into a matrix of probability management rather than unique occasional decisions. Despite online's popularity, face-to-face, where poker-like skills of skilled mentalists reading other players may provide confidence override such probabilities, may be the only real sustainable avenue.

In 2004, high-ranked player とつげき published 科学する麻雀 containing hordes of statistical data from tenhou, and, some rather blunt answers to many of Riichi Mahjong's "difficult" decisions. Conventional wisdom remains:

* straights (waiting on 8 tiles) are better than "of-a-kind" waiting on 2
* don't keep more than two "of-a-kind"s if you have other valid mentsu (tiles that might become melds)
* aim for pinfu
* don't call unless you have a yaku and probably want to be nishanten (two away from tenpai) or better unless you are oya
* pay attention to other people's open "of-a-kind"s (especially) kans, because they dramatically change the probability of your own straights
* if you have two or more pairs of special tiles you should aim for konissotsu (same suit) and can call twice
* call more than twice and you are in heavy danger of giving up a winning tile to someone else (you have very little choice on what to discard). do so only if you are prepared for all or nothing

but more interesting than these micro probability decisions (most of which can be proven mathematically rather than heuristically) is the macro wisdom:

* まわし打ち "playing defense and offense" is almost always bad. if someone calls reach or the oya has two open melds and you aren't at tenpai yourself (or maybe isshanten if your hand is worth a lot and you have good odds), you should switch to defense. you only need two safe tiles at the point of switching to defense to in most cases completely avoid being the one who gives the winning tile.
* 追いかけリーチ "calling reach after someone already called it" is almost always good. this extends largely from another point (reach is almost always good), but it's even better because you know for certain at least one other player isn't going to play defense against you
* calling reach is almost always good. don't think how likely your tiles are to come or not, or if other people are aiming for the same things--just call it. the ONLY reason not to call reach is because (1) you already have a yaku and (2) you are waiting on a midtile such as a 35 waiting on 4, that could be improved into a 23 or a 56 waiting on 14 or 47 respectively and that would give you pinfu. And even then if someone drops your midtile you should ron/agaru on it.

There's more too. If you care, buy the book. But let's move onto the most important idea: how quickly to try to go out. If there is one *key* decision in Mahjong, this is the one. Do you call, possibly lose points, to go out sooner? Do you go out with a winning hand even though you are one or two tiles away from a winning hand that might score even more points?

The answer is almost always "as soon as possible". This is somewhat obvious if you are oya. Your score already has a multiplier and you get another turn as oya. And if you are not oya, then someone else is oya, and by the same logic they will be trying to go out as quickly as possible. Since you get no points if they go out before you, you want to go out before them, which means, again, as quickly as possible. About the only time you don't want to go out as quickly as possible is if (1) you have no more chances to be oya and (2) you are in third or fourth place. If you are playing for a trophy, then it applies to second place too, but usually you are playing for rank and 2nd place is still an improvement. Anyway, in these situations you want to get rid of any dangerous tiles that might help the oya, 1st place, or 2nd place get a fast yaku and go out quickly as soon as possible, and aim big.

It can be argued that much of the reason why this is true is because of Reach and because of Aka-Dora. And, yes, Kuitan makes it worse. So house rules to have no Kuitan or get rid of Aka-Dora, or to play Western Japan style, or Chinese style, could possibly decrease the probabilistic advantage of simply trying to go out as soon as possible. There's not really enough data on those modes to say for sure, because they aren't as popular. Similarly, as I mentioned before, in face-to-face, a person skilled in psychology, mentalism, reading other players may be able to ignore statistics entirely if they have a 99% certainty they know how close other players are to going out. However, it does remain that simplifying one of the "most interesting" decisions in Mahjong to a rule of thumb does make the game significantly less interesting for many players. Therefore, accept this information at your risk!
 Thumb up
  • [+] Dice rolls
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.