1,828
edits
mNo edit summary |
(update some strategy points) |
||
Line 10: | Line 10: | ||
[[Image:Indefensible riichi.png|thumb|right|250px|Riichi does not let you respond to [http://tenhou.net/0/?log=2012111209gm-0009-7447-x33034e7e5279&tw=2&ts=4 any conditions that change in the future].]] <!--Aligns with the disadvantages section better here--> | [[Image:Indefensible riichi.png|thumb|right|250px|Riichi does not let you respond to [http://tenhou.net/0/?log=2012111209gm-0009-7447-x33034e7e5279&tw=2&ts=4 any conditions that change in the future].]] <!--Aligns with the disadvantages section better here--> | ||
* Riichi is a yaku. It grants 1 additional han and allows any closed hand to meet the yaku requirement. | * Riichi is a yaku. It grants 1 additional han and allows any closed hand to meet the yaku requirement. | ||
* You can score additional han through [[ippatsu]] and/or [[ura dora]] | * You can score additional han through [[ippatsu]] and/or [[ura dora]]. In addition, you are more likely to [[tsumo]] with a riichi, increasing the chance for [[menzen tsumo]]. When considering these bonuses, a riichi is worth an average of around 1.5 han. | ||
** Each [[kan]] formed by any player provides a kan uradora to the riichi, in addition to the regular kan dora. | ** Each [[kan]] formed by any player provides a kan uradora to the riichi, in addition to the regular kan dora. | ||
* Other players may defend against a riichi call, giving up their hands to avoid dealing in. | * Other players may defend against a riichi call, giving up their hands to avoid dealing in. | ||
Line 24: | Line 24: | ||
== Considerations for riichi == | == Considerations for riichi == | ||
Once again, riichi is a discretionary play. | Once again, riichi is a discretionary play. Every time you call riichi, some risk is involved, so you should consider various factors before making the call. At the same time, riichi is strong, so there are many cases where you should use it. Whenever you should or shouldn't riichi depends on the gamestate. | ||
As an alternative, you can go [[damaten]], which is simply a closed tenpai that hasn't called riichi. Another alternative is to [[betaori|give up the hand]] entirely. | As an alternative, you can go [[damaten]], which is simply a closed tenpai that hasn't called riichi. Another alternative is to [[betaori|give up the hand]] entirely. | ||
Line 45: | Line 45: | ||
Since most of the risks of riichi only apply if the hand fails to win, having a good wait means that riichi is more desirable. Bad shape waits may be more pressured to keep the possibility of failure in mind, as well as the possibility of changing the hand into having a better wait. | Since most of the risks of riichi only apply if the hand fails to win, having a good wait means that riichi is more desirable. Bad shape waits may be more pressured to keep the possibility of failure in mind, as well as the possibility of changing the hand into having a better wait. | ||
If you are in [[furiten]], you should be more wary about declaring riichi. A furiten | If you are in [[furiten]], you should be more wary about declaring riichi. | ||
* A furiten two-sided wait is around the level of a non-furiten one-sided wait. It's a bit less likely to win, but the average win score is a bit higher due to guaranteed [[menzen tsumo]]. | |||
* A furiten three-sided wait (>=9 tiles left) is relatively strong. Don't be afraid to riichi with it. | |||
== Score == | == Score == | ||
Riichi gives 1 han, and possibly more via ippatsu and ura dora. Since every han doubles your score until you reach [[mangan]], these are all valuable. However, there are rare cases where score does not matter. | Riichi gives 1 han, and possibly more via ippatsu and ura dora. Since every han doubles your score until you reach [[mangan]], these are all valuable; riichi triples your average score. However, there are rare cases where score does not matter. | ||
*When riichi wouldn't actually change your hand value. For example, [[haneman]] is scored with 6 or 7 han. If you have a 6 han hand, riichi's +1 han would still result in a haneman. Getting [[mentsumo]], ippatsu, and/or ura dora will increase it up to baiman, but these are unreliable, and the difference between haneman and baiman is too small to be worth the risk. | *When riichi wouldn't actually change your hand value. For example, [[haneman]] is scored with 6 or 7 han. If you have a 6 han hand, riichi's +1 han would still result in a haneman. Getting [[mentsumo]], ippatsu, and/or ura dora will increase it up to baiman, but these are unreliable, and the difference between haneman and baiman is too small to be worth the risk. | ||
Line 54: | Line 56: | ||
*When you have a valuable hand. If you have a guaranteed mangan or higher hand without needing riichi, the extra score may not be worth the loss of win rate. | *When you have a valuable hand. If you have a guaranteed mangan or higher hand without needing riichi, the extra score may not be worth the loss of win rate. | ||
That being said, the extra points from riichi often do matter. | That being said, the extra points from riichi often do matter. Mahjong is a game about scoring the most points, after all. | ||
=== Winning chance === | === Winning chance === | ||
Pressing players to defend will lower your chance of winning. If a hand has a yaku other than riichi, declaring riichi will generally lower your winrate by about | Pressing players to defend will lower your chance of winning. If a hand has a yaku other than riichi, declaring riichi will generally lower your winrate by about x4/5. Since riichi more-than-doubles your score until reaching mangan (1 han from riichi doubles score, then potential extra han from ippatsu/uradora), when you are below mangan, riichi will generally has a higher expected value. | ||
If you don't care about points, but you do care about winning, you should dama. As mentioned above, if you are in the lead by a large amount, dama. | If you don't care about points, but you do care about winning, you should dama. As mentioned above, if you are in the lead by a large amount, dama. |
edits