Defence theory(16): Tricks to defending

Source: http://blog.jpmahjong.net/read.php/567.htm

In the previous articles, we discussed betaori techniques when someone has declared riichi or when defending. However, to truly defend is not just about defending when someone attacks. As the discards increases, the chances of the other players reaching tenpai increases. If you put in a little effort in some details, you can avoid dealing in.

A few commonly seen techniques are shown below.

Discard the good tile first\Keep the safe tile
For example, like the hand below.

A hand in iishanten, and the 4 sou is an useless tile.
Let's assume that you next draw a dead tile, nan. If the situation doesn't look dangerous, you'd normally discard 4 sou first. This is what it means by discarding the good tile first.

The theory behind discarding the good tile first, is that 4 sou has a higher chance of becoming a dangerous tile than nan in the future. Since 4 sou is useless to us, you should discard it before someone reaches tenpai, in order to get rid of the risky tile that would be in your hand when you reach tenpai.

The tactic about keeping the safe tile should not be abused.

If we change 4 sou to 7 sou, the situation is very different. If you draw nan this time, you should definitely discard it.
Previously when I was still playing in high level games in 東風莊, I would discard 7 sou and keep nan with these type of hands. But in modern day mahjong theories, this is a conservative method. In the image above, there are 20 tiles that can be added into the hand. But if we were to discard 7 sou for a safe tile, the number of effective tiles would be 16. This a significant 20% loss when compared to the former.

The first condition when keeping the safe tile is to not sacrifice tile efficiency. With the hands below, a safe tile would be discarded if drawn.

At the beginning of the hand, 4 sou should not be discarded, it has a chance of becoming a ryanmen.

 As 2356 pin has effective tiles repeating, it cannot be considered as a good shape. At the beginning of the hand, 7 sou should be kept, but can be tricky after the middle discard rounds.

Going into betaori early
Mahjong is played by four players, and each player has a 25% chance of winning. Among them are hands that have very little chance of winning, for example:

This is a sanshanten hand with many bad shapes, and there is no way to increase the speed via calling tiles. If your hand remains this way even after the 5th or 6th, unless you make godly draws, your chances of winning is close to zero. Since you can't win, you should try to minimize your losses. Hence going into betaori is the best choice here.

Discard tiles that have been discarded by other players(kamicha would be the best) in the same discard round, and keep the tiles(like word tiles) that have a high degree of safety. If someone declares riichi later, you can remain in a safe state and reduce dealing in to the minimum. This method also protects you from dealing in to the damaten of other players.

Normally, a sanshanten hand in the 6th discard round and a ryanshanten hand in the 12th discard round has zero chance of winning. With this sort of tiles, you can either call tiles and try to win or go into betaori. Just like in poker, it's not possible to have a good hand every single round. Attempting to win when you have a bad hand will only cause you to incur more losses.

The two theories originate from the principle below.

Discard the dangerous tile before someone reaches tenpai.

Discard safe tiles sequentially after someone reaches tenpai .

This is the true essence of reducing the chances of dealing in.

Defence theory(15): Defence theory against two or more players(Continued)

Source: http://blog.jpmahjong.net/read.php/565.htm

This article continues the previous discussion about complicated situations.

At the beginning of a hand in a phoenix hanchan, the point spread is rather special. Toimen only has 1500 points, while I only lag behind shimocha by 2000 points in third place.

[First image in http://blog.jpmahjong.net/read.php/565.htm]

Originally, I had hoped to declare riichi with this hand. It's a good thing even if I knock out toimen and end up as second place. However, it was frightening when kamicha called 3 sets of tiles early in the game. Even though I knew that if kamicha had tsumo, I would end as third place, but because my hand was not in tenpai, I had to go into betaori.

Shimocha seeing this situation, naturally goes into betaori. The one you have to take special note of is toimen who is in fourth place. I followed kamicha's discard and discarded 7 wan, which toimen pon on, indicating that toimen is not in betaori. As I mentioned previously, attacking during situations where betaori is needed, is a sign of being in tenpai. Therefore not only do you have to defend against kamicha, you have to be careful of toimen.

I discarded the genbutsu of kamicha 7 sou, I draw 9 sou next. What should I discard?
[Second image in http://blog.jpmahjong.net/read.php/565.htm]

As both players have exposed melds, we can analyze it a little bit.

Let's talk about kamicha first. Kamicha called three sets of tiles early in the game, and has been discarding what he drew since the sixth discard. Obviously, he's in tenpai and his hand is not only worth 3900 points. (If kamicha's hand only has 3900 points, with this kind of points situation, there's no need to go into betaori.) Honitsu and toitoihou is the most likely here. In other words, kamicha has at least a mangan, even a 18000 points dealer haneman is possible, that's why it's frightening. As the discards of kamicha is too neat, we are unable to find out which suit he's waiting on. However since kamicha discarded quite a lot of pinzu early, the chances of a pinzu honitsu is the lowest. Also, in order to avoid toitoihou, you must avoid discarding any live tiles. Thus, the 8 and 9 sou in your hand is the most dangerous.

As for toimen, the hand is most likely a tanyao. Since dora was discarded, the value of the hand shouldn't be too high. It would be 3900 points even with aka-dora, the danger is low but you shouldn't be careless. Since you're in betaori, you might as well avoid dealing-in, although the main focus is on kamicha.

After some deliberation, I discarded 8 pin.

Although 8 pin is a live tile, it's also the suji-pai of toimen, giving it some degree of safety.
Kamicha discarded 7 and 9 pin earlier on, if kamicha had 7889 pin on the early stages, he wouldn't have discard 7 and 9 pin in a situation where honitsu is confirmed. Hence, we can conclude that kamicha discarded 7 and 9 pin while dismantling a taatsu. Since kamicha discarded 5 pin, a ryanmen wait on 58 pin is eliminated, thus 8 pin has a huge chance of passing.

Game Replay (Click on 全局牌譜 in http://blog.jpmahjong.net/read.php/565.htm)
L0000 鳳南喰赤 平均R2070.2
1位 くるる(七段/R2010/+46) 3位 zeroRX(八段/R2137/-24) 4位 麦わらぼうし(七段/R2037/-36) 2位 kickchi(七段/R2097/+14) 

These sort of safe tile judging in betaori is not easy, as it involves a lot of analysis and theory. Also you'll need to pay attention to other players, to make the correct decision. These two articles can be said to be an example to show everyone what to focus on when finding safe tiles. If there are any more good examples in the future, they'll be posted to be shared with everyone.

The next article will be about some tricks to defending.
(To be continued)

Defence theory(14): Defence theory against two or more players

Source: http://blog.jpmahjong.net/read.php/561.htm

For the sake of simplicity, the previous theories assumed that only one player was attacking. However that is not the case in real games. Situations where one player declares riichi and another player pursues riichi or goes into kanzen shinko are not rare.

The author mentioned before that if your order of discarding during betaori is done properly, it will effectively raise the chances of having safe tiles to discard when many players declare riichi. However, there will still be times when you have no safe tiles to discard. It is at times like this, where making good judgement becomes even more important.

General direction
The aim of defending is to avoid dealing in, but when this doesn't work, avoid dealing in to big hands is the second best practice.

When facing riichi from two players and above, avoid dealing in as the value of the hand is hard to estimate. But if there are melded sets, we can estimate the value of the hand with the exposed tiles and discards. If we know that the value of the hand is small, we can ignore that player or deliberately deal in to him.

Finding safe tiles
Firstly, we need to find genbutsu that both players have, but there may not be any most of the time. The following is the usual way to discard:
1. Genbutsu of both players
2. Genbutsu of one player, suji-pai of the other
3. Suji-pai of both players
4. Suji-pai of one player, non suji-pai of the other
5. A middle tile that is a non suji-pai

However as safe tiles increase quickly(as any discard by either players become genbutsu for both players), therefore the method mentioned previously where you discard toitsu might not be useful here... Instead, appropriate use of tile reading and situation judging will be of great use here.

[First image in http://blog.jpmahjong.net/read.php/561.htm]

Two players declared riichi, what is the safest tile in your hand?

The safest tile is 1 pin. Four 2 pin and two 1 pin have already been discarded, therefore 1 pin can only be a hell wait.

If there's no safe tile in the next round, personally I will discard 4 sou. Shimocha made an ankan with 3 sou and 5 sou was discarded 3 times, the chances of 4 sou becoming mentsu is the lowest. As for 8 wan and 3 pin, they are very dangerous as the tiles near them have not been discarded. For this round of discards, if the 4 sou passes, you should be able to make it to ryukyoku.

After discarding 1 pin, you should be able to make it to ryokyoku without making any dangerous discards. Toimen is waiting on 36 wan and shimocha is waiting on 7 wan.
[Second image in http://blog.jpmahjong.net/read.php/561.htm]

The next article will talk about more complicated examples.

Defence theory(13): Detailed explanation of each type of melded sets(Part 2)

Source: http://blog.jpmahjong.net/read.php/558.htm

The melded sets mentioned in the previous article, all belong to "sakidzuke katachi". What it means is that the melded sets are related to yaku. Sometimes, the first and second melded sets of the opponent may not be related to yaku. The aim of calling those tiles is to get rid of the bad shape of the hand. This type of attack is referred to as "atodzuke katachi", for example:

[First image in http://blog.jpmahjong.net/read.php/558.htm]

or

[Second image in http://blog.jpmahjong.net/read.php/558.htm]

Normally, "atodzuke katachi" melded sets have a higher value than "sakidzuke katachi", but as the possibility of the hand is heavily restricted, it's not hard to find out what tiles are dangerous. This article will discuss how to defend against this type of attack.

These two are commonly seen atodzuke yaku:
1) Yakuhai
2) Sanshoku, Itsuu, Honitsu

Among them, yakuhai has the highest probability of appearing. So once melded sets appear, the first thing you should check for are live yakuhai tiles. (capable of making a koutsu.) If you have such tiles, there are two choices you can take:
The first is to discard it immediately, hoping that the opponent is not in tenpai.
The second is to keep it until the end of the hand.

If the opponent starts to discard the tiles he draw, then the chances of being near tenpai is high. The live yakuhai tiles now are dangerous tiles, and should not be discarded.
If you've made sure that yakuhai is not possible, then the only possibility now is sanshoku and itsuu. Most of the time, you can rely on the discard pond to estimate the tiles related to yaku, and avoid discarding them. As for the value of the hand, you can estimate using the amount of dora you haven't seen. If the dora is related to yaku(for example the dora is yaochuhai and the hand is hon chanta), you need to be more careful.

Using a simple example
[Third image in http://blog.jpmahjong.net/read.php/558.htm]

Kamicha calls 2 pin. You can tell that it's not a honitsu by looking at the discards. You can see four 4 sou and 9 pin, therefore the chances of itsuu and sanshoku is gone. The only possibilities left are hon chanta and yakuhai. From the discards, the possibility of the former is rather high.
The yakuhai that has not appeared yet are haku and hatsu, and since 9 wan is the dora, these three tiles are highly dangerous. Secondly, the tiles that can form shuntsu of 123 and 789, like 78 sou, 2378 wan etc are not safe.

From the standpoint of the top player, these tiles must not be discarded.

Defence theory(12): Detailed explanation of each type of melded sets(Part 1)

Source: http://blog.jpmahjong.net/read.php/552.htm

In the previous article, the author mentioned the important signals related to melded sets. This article will introduce in detail each type of melded sets and things to note.

As melded sets will expose information regarding the hand, if used appropriately, it will greatly help in making the proper judgement. Our aim is to use the melded sets to deduce the following.
1. Value of the hand
2. Area of waiting tiles(Or from another perspective, which tiles are safer.)

The following will explain in details the commonly seen types of melded sets.
1. Calling of middle tiles
This can be said to be the most commonly seen melded set in Japanese mahjong. Of course, the melded sets of the opponent are all middle tiles. The largest characteristic about calling middle tiles is that it's difficult to form with other yaku. (Among them, tanyao sanshoku has the highest chance of appearing, but even the chances of getting it is not high.) Hence, the amount of dora in the hand is the determining factor of the hand's value.

So once middle tiles are called, you first need to estimate the whereabouts of the dora. One hand of game has 7 dora(four indicated dora and 3 aka dora). The more dora you can see, the lower the chances that the opponent has a huge amount of dora. Also, if the attacker is really proactive at calling tiles(for example, calling 1 or 2 mentsu very early), then the chances of having a high value hand is higher. Sometimes when we know that there is a high chance that the opponent only have a 1000 or 2000 points hand, we can consider ignoring them.

Against calling of middle tiles, it's easier when it comes to defending, as any yaochuhai is a safe tile. However as the area of waiting tiles is narrow, over reliance on suji-pai and one chance tiles, especially dora soba, can be dangerous.

2. Somete
Like calling of middle tiles, this is also a commonly seen melded set. It's easy to judge if it's a honitsu or chinitsu melded set, as you can see it from your opponent's discards.

Somete has a narrow area of melded set and can be easily recognized, but it's return is usually higher. Somete usually have at least 3900 points, mangan or even haneman are not a rare sight.

For this reason, somete has defence theory like suji-pai etc. Unless it's a genbutsu or tiles incapable of becoming mentsu, or else it's a dangerous tile.

3. Toitoihou
Toitoihou attacks are a common sight in beginners' games, but in higher levels of mahjong games, the chances of appearing is lower. When the opponent often pon tiles, and the discards are made up of random middle tiles, then the chances of toitoihou is large.

The lethality of toitoihou is not low either, it's normally between 2500 to 8000 points. Additional han values are usually from yakuhai 9 out of 10 times. So if you note the yakuhai that have not yet appeared, you can probably estimate the value of the opponent's hand.

As toitoihou is always waiting on tanki or shanpon, there are no suji-pai. Live tiles, no chance tiles and yakuhai are all ultra dangerous. Tenpai is usually on easier waits, therefore middle tiles that have been discarded twice are relatively safer. And when we go into betaori, sometimes we would discard koutsu middle tiles.

4. Yakuhai
Yakuhai tiles are express tickets to winning. You'll get yaku as long as you can pon with it, and as it restricts tile shapes the least, it becomes easier to form hands.

Not only is it easy to combine it with other common yaku(toitoihou, honitsu), a dealer's double ton or yakuhai dora 2/3 are all powerful attacking methods in Japanese mahjong.

As the hand varies greatly, it's harder to deduce the value and waiting tile of the opponent's hand. Normally, the first thing you should do is to note the amount of dora tiles that have not yet appeared, and whether or not the opponent's discards show any characteristics of other yaku.

As for defending, once the signs of being in tenpai appears, treat it as a normal riichi as the waiting tile is not restricted. But as there are melded sets, we can obtain more information regarding what tiles are safe or dangerous:

a) Tiles that have not appeared, are all dangerous tiles. (Translator note: For those of you who can read Chinese, you might notice that I didn't translate this sentence fully. The reason is that it didn't make any sense. The author said that the tile discarded by the attacking player after making the final mentsu is a dangerous tile. Since the attacker cannot ron on a tile he discarded, how can that be a dangerous tile? I will leave it like this until I understand what he really meant.)
b) Tiles discarded by the kamicha of the attacking player but have not been called, are usually safer. The reason is that this proves that that tile is not need by the attacking player. (If it's a menzen, even if others discard a tile you want, you can't call on it.)

(As the current section is too long, the rest of the article will be moved to the next section.)
(To be continued.)

Defence theory(11): Defence against melded sets(1)

Source: http://blog.jpmahjong.net/read.php/550.htm

Starting from this hand, I'll begin discussing about defence against melded sets.

Everyone might be thinking that defence against melded sets are harder to deal with, as you don't know when the other side is in tenpai. However up to a certain degree, mentsu that are melded allows you to find out the situation of your opponent's hand. From there you can formulate better and more accurate offence and defence tactics.

Mentsu that are melded, at the very least allows you to know the following situations.

1. You can predict the yaku your opponent wants, estimate the value of the hand and then the area of waiting tiles from there. Sometimes when we know the opponent has a small value hand, we can consider not going into betaori.

2. You can know the order of mentsu forming and along with discards, you can find out the area of dangerous tiles more accurately.

Also, even though the player with melded sets cannot declare riichi, but based on the situation of the game, we can usually estimate whether the opposite side is in tenpai, this sort of tile reading is highly accurate. The following are commonly seen methods of finding out if someone is in tenpai.

1) Opponent's amount of melded sets and time
Normally, you know whether an opponent is in tenpai from the following:
a. 3 melded sets at any time
b. Several tiles change occurring after 2 melded sets
c. 1 melded set near the end of the hand
Unless the opponent is calling tiles randomly, otherwise these sort of behaviour is no different from declaring riichi.

2) Opponent's discards
The most important thing to note is whether or not the opponent has discarded tiles related to major yakus. Example:
a. Discarding manzu when making a manzu chinitsu.
b. Discarding a middle tile dora after calling middle tiles.
c. Discarding chun while it's a live tile, after having pon on haku and hatsu.

When situations like these occur, the opponent is probably in tenpai, at the very least in iishanten. Also, if the opponent continuously discards the tiles they draw or discard a very safe tile, this signifies that the opponent is in tenpai.

3) The action the opponent takes when other players attack
This sort of tile reading is useful when playing in high level games.
You can tell that an opponent with melded sets is in tenpai, when he takes the following actions even when another player takes an obvious offensive stance.
a. continue to call on tiles
b. continue to discard dangerous tiles

Of course, knowing whether or not the opponent is in tenpai is just the first step. The next step would be determining the value of the hand and the area the waiting tiles are in. This will be discussed in the next article.
(To be continued.)

Defence theory(10): What do you do when there's no safe tiles?

Source: http://blog.jpmahjong.net/read.php/549.htm

The betaori techniques mentioned before, were all assuming that you had enough safe tiles. Usually, when someone is attacking and you hold two or more safe tiles, there's a high chance of avoiding a deal-in. The reason is that once you have discarded two safe tiles, chances are there'll be more safe tiles. According to statistics, when your hand is closed, there's a 50% chance of drawing a safe tile. (Including the tiles discarded by the player in riichi or other players, and the tiles you draw and confirmed to be safe tiles.)

Even so, there are times when we have no safe tiles when someone is attacking. The following text will introduce a few commonly seen methods.

1) Dismantle anko/toitsu
This is the most commonly seen method, mainly based on the idea that "betting on a dangerous tile can buy a few safe tiles".

Sometimes, just when we've just dealt with all the word tiles, the opponent declares riichi.

(In riichi)

If your hand is a ryanshanten hand, it is rather far-fetched to attack a player who's already in tenpai, going to betaori here is the norm.

However there are no safe tiles here. In normal situations, I would recommend discarding the half suji-pai of 6 wan. Betting on a dangerous tile can buy a few safe tiles. If there are no safe tiles even after discarding finish all three 6 wan(the probability is very low), you can still rely on suji-pai of 3 wan. By the way, if the attacking player discards 9, and if you have 3 and 6 in your hand, discarding 6 is tesuji.

This idea remains useful even when there are safe tiles available.
When someone declares riichi, and your hand is like this:

The only suji-pai is 3s, and 1 pin is one chance, while the rest are not suji-pai. As the suji-pai is a dora soba, you might not feel comfortable in discarding it, therefore you can discard the riskier 1 wan. Once 1 wan passes, you won't have to worry about not having any safe tiles. You're already in a dangerous situation, and if you discard 3 sou first and don't get any safe tiles in the next discard, you'll still have to discard 1 wan. Compared to discarding 1 wan before, 3 sou becomes a pointless risky tile.

2) Kanzen Shinko
When you don't have any safe tiles, going into kanzen shinko is also a choice, especially when you have called any tiles. You have to understand, that the lesser the amount of tiles in your hand, the lower the chances of new tiles appearing for the next round of discards. Of course, whether or not to attack, is also dependent on the points situation.

[First image in http://blog.jpmahjong.net/read.php/549.htm]

Like the example above, there's only one suji-pai of 3 wan, it is logical to discard 8 wan or 7 sou to kanzen shinko. Even if you discard 3 wan, you'll probably need to discard more dangerous tiles in the future. Even though the chances of winning is low, but the value of the hand is high enough and you're the dealer. Summoning up your courage and attack seems to be the only choice here.

Defence theory(9): Important points of betaori(Part 2)

Source: http://blog.jpmahjong.net/read.php/547.htm

The aim of this article is to develop your perception towards betaori, while also introducing some high level techniques.

[First image in http://blog.jpmahjong.net/read.php/547.htm]

It's south round first hand in the above picture, kamicha declares riich on the 8th discard, what should Ton do?
First you should find out, is this a situation to go into betaori for?

Ton gained a considerable advantage from the east round, the important thing to do now is to reduce the amount of points loss in the south round. Even though the pinfu dora 2 is a good hand, but it's only ryanshanten, in a situation where you're leading, I believe going into betaori is a correct choice.

Once you know you need to go into betaori, the next step would be finding the safe tiles. The genbutsu of kamicha are currently 4 pin and chun. While 1 wan is the safest tile outside of genbutsu. (1 wan is a no chance tile, don't forget the 2 wan that was pon.) I'll take the opportunity to mention, a lot of beginners believe that just because kamicha discard 1 and 3 sou, 2 sou would be a safe tile. This is a misconception.

Then, the genbutsu of 4 pin and chun, which should be discarded first? I will first discard 4 pin, there are two reasons for this.
1) Currently, 4 pin is an absolutely safe tile. (Kamicha has discarded 4 pin before, therefore he/she cannot win on the 4 pin.)
2) Another reason for discarding 4 pin, is expecting shimocha to call on it and continue to attack. From this current situation, kamicha who is in third is in riichi, shimocha who is in fourth won't go into betaori easily. Deliberately letting shimocha call your tile, is to cause a situation where both players attack each other, increasing the chances of dealing in for both players. In other words, decreasing your chances of dealing in. This is a practical high level technique.

A lot of readers may feel that discarding chun and maintaining the taatsu is still commendable. However everyone has to understand this, to reach tenpai with this hand, you need to gamble on discarding 2 sou and 9 sou, both of which are not suji-pai. With this situation in points, this is too much of a gamble.

Betaori, is not just about finding a few genbutsu and discarding them. One of the important points of an excellent betaori technique, is to take stock of the situation, list out an accurate discard order, and genuinely decreasing your chances of dealing in to the lowest.

I hope that everyone can continue to improve their betaori.

(To be continued.)

Defence theory(8): Important points of betaori(Part 1)

Source: http://blog.jpmahjong.net/read.php/544.htm

The previous article spoke specifically about betaori's timing, the next two articles will go into detailed explanations about the important points of betaori and techniques.

Again, betaori is totally giving up your chances of winning, reducing your chances of dealing in to the lowest. Hence at this stage, especially those who are new to betaori, don't be afraid to dismantle good mentsu or taatsu, or attempt to reach tenpai or even try to win.

The text below will start to explain the related points.

1) Betaori starts with the safest tile
If the reader paid attention to the previous articles, then you should know that this is not the first time I've said this. As this is an important principle in betaori, I'll patiently repeat it once more.

[First image in http://blog.jpmahjong.net/read.php/544.htm]

Shimocha declares riichi, it's obvious that this situation requires betaori. A hand in suushanten, if it's because [I am the dealer] and recklessly try to attack, you'll only lose even more badly.
When you're in betaori, you better prepare to sort out the safe tiles in your hand in order.

Which are the safest tiles now?
The safest tile is of course ton, the genbutsu of shimocha. As haku is already pon by toimen, therefore it's a 100% safe tile. At the same time, don't forget that 7 pin was pon by shimocha, therefore it's also a genbutsu.
What's the fourth safest after that? A lot of people think that it's suji of 9 wan... The correct answer is 9 pin. Toimen pon the 7 pin, and the last 7 pin is in your hand, in other words, 9 pin is a no chance tile. There is also two other 9 pin discarded, 9 pin can only be a tanki wait, it is the fourth safest tile. 9 wan is a suji-pai, but there are no other 9 wan discarded, therefore it's degree of safety is lower.

Therefore the betaori discard order in this situation is: Ton/haku -> 7 pin -> 9 pin -> 9 wan -> 9 wan

The most common mistake by beginners, is over reliance on suji-pai, discarding 9 wan thinking that it's relatively safe. This is a bad habit, because it is not rare to deal in over careless mistakes. The thing to note is, after discarding genbutsu for 2 turns, more genbutsu could appear, and you may not even need to discard 9 wan. Don't underestimate such triviality. During a few hundred games you play, there could be a few hundred hands you need to go into betaori for, and a stricter betaori could cause you to deal in a few hands lesser.

2) Try to hold on to safe tiles common to multiple players
On the other hand, once your hand holds several genbutsu, and you decide to go into betaori, I will first discard genbutsu of number tiles. The reason is that if another player starts to attack later, then your word tiles could be used to defend against both players. The danger about holding genbutsu of number tiles is, once another player declares riichi, this safe tiles will turn into dangerous tiles.
 
[Second image in http://blog.jpmahjong.net/read.php/544.htm]

The situation is currently in betaori. The 8 pin discarded by kamicha is a new genbutsu. You should discard 8 pin here, and keep the xia, which is relatively safe to the other players. In the real game, toimen discarded 3 sou and declared purse riichi. This xia became an important safe tile.

Let's cite another difficult example.
[Third image in http://blog.jpmahjong.net/read.php/544.htm]

Shimocha is in riichi, you need to go into betaori here too. Although 8 wan and 9 pin are both genbutsu, but you should discard 8 wan here. The reason is that toimen discarded 6 pin, 9 pin is a relatively safe suji-pai. If toimen declares pursue riichi afterwards, you still have a safe tile to discard.
There are people who assume that there is no difference between 8 wan and 9 pin, this is a casual thought. Betaori is not just discarding the genbutsu of the player who declared riichi. A person who is excellent in betaori, will take careful note of the order he discards his tiles. Remember, your enemy is not only the one who declared riichi.

Defence theory(7): Requirement for betaori

Source: http://blog.jpmahjong.net/read.php/542.htm

We've prepared the absolute requirement for betaori. For the next 2 to 3 articles, we'll be explaining how to betaori using actual games.
This article will be explaining when we should go into betaori. (To a certain degree, this is a discussion about offence and defence for beginners.) Look at the game below.
[First image in http://blog.jpmahjong.net/read.php/542.htm]

Kamicha declared riichi at the fifth discard, what should you do?
Discard west here, just for this discard. Someone who discards anything else, might not have the qualifications to discard on the level of "Special". (Translator note: My guess is that the author is referring to a ranking called "Special".)
The hand is currently in yonshanteen, while kamicha is in tenpai, the situation is like you're 50m behind someone in a 100m race, don't even think of trying to catch up. Additionally, avoiding a deal-in while trying to reach tenpai is a difficult problem.

Betaori is the only way for this sort of tiles.
[Oh, let's discard 1 pin and see what's going to happen next.] Thoughts like this, should be avoided at all cost. You must discard the safest tile, xia here.
Remember, when in betaori, start with the safest tile.
Usually, if you want to pursue riichi or challenge someone in tenpai, you need to have a iishanten or better with a good shape. For anything ryanshanten or below, unless there are no safe tiles or the situation is special, please do not hesitate going into betaori.

In Japan, some high quality tactics discussion recommend beginners to employ a simple offence and defence baseline.

Once an opponent declares riichi, if your hand is already in tenpai, then attack immediately(or pursue riichi), or else go into betaori.

Do not underestimate this method of judgement, if everyone can execute this perfectly, going up to 3 dan, or even entering "special" might not be a problem.

Shanten number, is more important than hand value

The largest factor in deciding whether or not to attack, is your own shanten number, instead the hand value is a secondary factor. Because, if you win, your opponent's chance of win disappears at the same time. A hand worth 3900 points in tenpai, is better to attack with than an iishanten hand worth 8000 points.

[Second image in http://blog.jpmahjong.net/read.php/542.htm]

The hand in the image above has a high chance of becoming a mentanpin dora 1 mangan hand. But if shimocha declares riichi, and your hand is still in ryanshanten, you should immediately go into betaori.

In this discard, discarding the live tiles xia or haku is very dangerous. You should discard the two genbutsu, and go into betaori. Most beginners find that it's very unnatural to go into betaori here, but having the courage to go into betaori with such good tiles, is an important step to improving.

Defence theory(6): Tile risk table

Source: http://blog.jpmahjong.net/read.php/541.htm

Theories regarding kabe and suji have already been covered, next will be a summarization of the different degree of safety of tiles, making it into a risk table.

The risk ordering above, should be memorized by beginners, because the ordering above decides what tile to discard first when in betaori, which heavily affects deal-in rate.

Risk factors
The table above refers to normal situations, you should consider the following in a real game:
1) dora
As there will be more points from waiting on the dora, waits on the dora can be predicted. When the dora is chunchan, the deal-in rate is 1.5-1.6 times of the normal rate, while when the dora is a yaochuhai, the deal-in rate increases to 1.8-1.9 times. Also, the tiles near the dora (a.k.a dora soba) have a 10% increase in deal-in rate.
Hence, the degree of safety of the dora tile, yaochuhai's grade has to drop 3 grades, chunchan has to drop 2 grades. While the dora soba has to drop 1 grade.


2) One chance tiles
The degree of safety for one chance tiles is roughly between non suji-pai and suji-pai.

3) The outer tiles of early discarding of 28,37
If the attacker discarded 28, 37 number tiles early, the outer tiles will be safer than usual. (For example if the player who declared riichi discarded 8 sou, then 9 sou is relatively safe.) Deal-in rate is 30% lower than usual.

Also, there are many discard reading books about ura suji, aida yon ken, matagi suji on the market. You might be wondering, why hasn't there been any talk about discard reading.
The reason is because according to game record statistics from 東風莊's super high level tables, this so called discard reading, do not affect the risk of related tiles. So when you're in betaori, it's better to ignore these "skills".

Defence theory(5): Kabe

Source: http://blog.jpmahjong.net/read.php/540.htm

Assuming your opponent is waiting on 58 wan, his hand must have 67 wan each.

If you can see(including your hand, discards and called tiles) four 7 wan, then there's no way his hand is holding 67 wan. 8 wan can only be a shanpon or tanki wait, making it safer.

This 7 wan is known in Japanese mahjong as [kabe], 8 wan is known as no chance tile.
The degree of safety of a no chance tile is equivalent to that of a word tile, therefore the more 8 wan you can see, the safer it is. (If three can be seen, the last one is 100% safe.)

Also, it is known as [usu kabe] if three 7 wan is seen, 8 wan is known as one chance tile. The risk of one chance is between a suji-pai and non suji-pai. The important thing to note is, as the game continues, the degree of safety for one chance gets lower and lower, as the chances of drawing the last 7 wan gets higher and higher.

Fusion of kabe and suji
Effective use of kabe and suji can help determine more safe tiles, listed below is a few common examples:

a) If 4 and 7 are both kabe, then 5 and 6 are the same as word tiles, they can only be tanki or shanpon waits.
b) If 9 has been discarded, 6 is half suji-pai. If 4 is a kabe, because 45 cannot form with 6, the degree of safety for 6 is equivalent to double suji-pai. It would be even better if 5 is a kabe, as 6 is on the same level as suji of 1 and 9.

Kabe and suji have roughly been discussed. The next time will be introducing tile risk table.
(To be continued)

Defence theory(4): Suji-pai

Source: http://blog.jpmahjong.net/read.php/539.htm

Next I'll introduce two very important concepts in Japanese mahjong: suji-pai and kabe.

In Japanese mahjong, 60% of tenpai in riichi are ryanmen waits. Therefore, if it is certain that some tiles cannot be ron by ryanmen tenpai, they are considered to be safe.

Suji-pai(Degree of safety: S~B)

We'll use the previous riichi image as an example. If the player who declared riichi discards 5 pin, then if you discard 8 pin, then the opponent can't ron on your discard even if he holds 67 pin and is waiting on 58 pin.(According to the furiten rule.) To ron on 8 pin, the only possibility is shanpon and kanchan, therefore the probability of dealing in is lower. Similarly, 2 pin, 9 pin, 9 sou are safer due to the fact that they're suji-pai.

If 4 is discarded, 1 and 7 are the suji-pai.
If 5 is discarded, 2 and 8 are the suji-pai.
If 6 is discarded, 3 and 9 are the suji-pai.

We normally remember these three groups as 1-4-7, 2-5-8, 3-6-9.
It is worth noting that, even if 1 sou is discarded, 4 sou is not a suji-pai, as he could still be waiting on 4-7 sou.

By the same token,
If 1 and 7 is discarded, 4 is a suji-pai.
If 2 and 8 is discarded, 5 is a suji-pai.
If 3 and 9 is discarded, 6 is a suji-pai.

The 456 here is called double suji-pai. If only one of the suji is discarded, it is called half suji-pai.

Difference between degree of safety for different suji-pai
There are many suji-pai introduced above, but their degree of safety are all different.

For example, the suji-pai of 9 pin, as the chances of ryanmen is eliminated, then it can only be tanki or shanpon tenpai, the chances of safety is similar to word tiles. If you see two of the 9 pin already discarded, 9 pin can only be hell wait, then the degree of safety is similar to that of tanki word tiles. If you can see three of the tiles, then it's 100% safe.
Hence, suji-pai of 1 and 9 can be said to be the safest among all suji-pai.

For suji-pai of 2 and 8, not only can it be shanpon and tanki, it can also be kanchan for 13 and 79. Therefore the degree of safety is lower.
For suji-pai of 3 and 7, there can be penchan of 12/89 and kanchan of 24/68. Therefore, the risk is much higher. The kanchan could even be part of a tanyao hand, increasing the degree of harm to you.
As for the double suji-pai of 456, as there can only be kanchan, the degree of safety is similar to suji-pai of 28.

In summary, the degree of safety for suji-pai in order is:

Suji-pai of 19 >> Suji-pai of 28 == Double suji-pai of 456 > Suji-pai of 37

The next round will be about another important concept: Kabe.
(To be continued)

Defence theory(3): Genbutsu

Source: http://blog.jpmahjong.net/read.php/537.htm

From this article onwards, I'll slowly explain the technique of betaori. For the sake of convenience, let's assume only one person is in tenpai.

Firstly, the most important principle in betaori is:

According to how safe a tile is, discard in order starting with the safest tile

Note, once you decide to betaori, you should discard a tile based on how safe it is, even if it means discarding your mentsu or shuntsu.

On an average, betaori happens 30%-40% of the time, therefore a person's betaori skill will directly affect their score.
Today, we'll introduce a few highly safe tiles.

Genbutsu(Degree of safety: SS)
Reusing the previous example:

The tiles here discarded by the player who declared riichi, including 56 pin, 7 wan, 6 sou etc, are known as genbutsu. They're 100% safe to the player who declared riichi. Even if the riichi is waiting for the tiles that the player discarded himself, because of the furiten rule, the player who declared riichi won't be able to win on your tile.

Similarly, any tile discarded after the riichi is genbutsu.

The fourth word tile(Degree of safety: SS)
Of course, if there is any word tile that has been discarded three times, it should be 100% safe to discard unless it's kokushi musou. As kokushi musou has unique discards, and is easily detected by others, added to the fact that the probability of kokushi musou is low, it's degree of safety is on the same level as genbutsu.

Tiles discarded by the player before you in the same turn(Degree of safety: SS)
Due to the furiten rule, tiles discarded in same turn by the player before you is also 100% safe.

Basically, these three are the only absolutely safe tiles, I'll continue to introduce other safe tiles based on the degree of safety.

Exhausted word tiles(Degree of safety: S)
Word tiles that have been discarded twice, even if there's a chance of a hell tanki wait, are relatively safe tile. Firstly, there is little chance of a player who declared riichi to be on a hell tanki wait, because they won't be able to win if the last tile is in the dead wall. Also a word tile waiting on tanki is easy to change, and most would change to a better tenpai before declaring riichi.

(To be continued)

Defence theory(2): Overview on defending in Japanese mahjong

Source: http://blog.jpmahjong.net/read.php/536.htm

Starting from this article, I'll begin to talk about Japanese mahjong's defence theory. I will first introduce common defence tactics in Japanese mahjong.

The goal of defending is rather simple, just don't discard the tile your opponent is waiting on. To achieve this, everyone might have thought: [As long as you know the tile your opponent is waiting on, you won't deal in.] Alright, now I'll invite everyone to guess, what is the player in the picture below waiting on?

"Reading point" is impossible!
The numerous mahjong books available all have theories related to discard reading, where they attempt to read the hand of the opponent just based on the discards. In most mahjong mangas, there are scenes where the opponent's waits are accurately read.

However, reality is cruel. A situation where you can accurately read the waiting tiles(a.k.a "reading point") of an opponent just by reading the discards is impossible. The discard reading skill mentioned in mahjong strategy books could be considered impressive with just an accuracy of 30%. Part of the outdated theories have also been proven to be wrong. (Such as ura-suji, aida yon ken etc)

If the reader hopes to learn perfect defence in the future theory discussions, like the emperors of old who tried to seek immortality, you'll only be disappointed.

The essence of defence in Japanese mahjong, identifying safe tiles
There is no immortality drug in this world, but there are ways to live longer. There are no techniques that allow you to see through the wait of your opponent, but there are many ways to avoid dealing in, and they are not difficult to learn.

The amount of games I've played in MFC are around 4800, with a deal-in rate of 12.6% in over 20000 hands. (winning rate of 25.1%) Compared to other so called high level Yellow Dragon players who have the similar winning rate as me, my deal-in rate is on an average at least 2% lower. But in my 4 years of playing MFC, I've never tried to guess the waiting tiles of my opponent. This proves that in order to defend well, you don't need any supernatural powers.

In order to defend well in Japanese mahjong, all you need to do is to find the safe tiles. Like the example in the picture above, I believe most of the readers are unable to guess the waiting tiles. This hand has a shanpon wait of 3 sou and 8 pin. But what we can be sure of is, you definitely won't deal in by discarding 56 pin, 7 wan and 6 sou.

If you don't follow highly accurate(with at least 98%) safe tile theories, but follow those unorthodox theories with only 20-30% accuracy, it will only end badly for you.

Strategies when opponent is in tenpai
Against an opponent's riichi, the strategies that you can employ, are the following:

1) Betaori: Totally giving up any chance of winning, and at the same time decreasing the chances of dealing-in to the minimum. This is the most commonly used tactic. In this series of articles, more than half the text will be covering this area.

2) Uchimawashi: While discarding tiles that have a good chance to be safe, continue to proceed to tenpai. If the tiles that you draw continue to be not useful, you can consider going into betaori. But if you want to be able to perform well here, you need to have good judgment, this can be difficult for beginners to grasp, therefore this article will only be briefly discussing this.

3) Kanzen Shinko: You'll discard any tile without hesitation, in simple terms you're ignoring your opponent's tenpai. The important thing to note is, even if the chance of winning is low, it does not mean that you shouldn't go into kanzen shinko. It's fine as long as you attack reasonably. (Translator note: The original text by the author didn't make sense. Therefore, I assumed that it's a typo error and changed it to what I think he really meant.)

Of course there are strategies to deliberately deal in, but the amount of them is too low.

The biggest problem with beginners, is that they go into kanzen shinko and ignore uchimawashi too often, and rarely go into betaori. They feel that they'll go into last place if they don't win the current hand. However in Japanese mahjong, as the losses from dealing into riichi is not small(riichi has ippatsu, chance of ura dora, menzen, and many other yaku), if you know your chances of winning is small, you should retreat and minimize your losses.

The next article will go into details on how to find safe tiles.

Defence theory(1): Why do you want to defend?

Source: http://blog.jpmahjong.net/read.php/535.htm

Japanese mahjong can be said to be the mahjong that pays the most attention to defending. Regardless of the depth of tactics, or the ratio of defending, when it comes to defending it's the champion of all mahjong.

In fact, according to Japanese mahjong researcher とつげき東北's latest work, one of the three most important factors that affects the results of a mahjong player, is the rate of dealing in. (To be more accurate, it should be the dealing in rate outside of riichi and having two exposed melds.)

When beginners see the high level game records of Tenhou special, they will find the players too cowardly. Once someone declares riichi, the other players will dismantle their tiles and aim for a draw. But what I want to tell everyone is, this is an important process in Japanese mahjong. In a usual high level game record, 40% of the games end in a draw.

If you find this kind of mahjong boring, Japanese mahjong is not for you.

Since the frequency of abandoning the win is so high, knowing how to defend is very important in Japanese mahjong. But the main reason why beginners tend to have a higher deal-in rate, is not because they don't know how to aim for a draw, but because they refuse to. Therefore before I seriously discuss "how to" aim for a draw, I must first explain "why". Otherwise, if you're not interested in defending, no matter how much I discuss defence theory it'll just be a waste of time.

The payout of a person who dealt-in in Japanese mahjong
In Japanese mahjong, if you deal in, you'll have to pay the full amount of the hand's value. I believe that everyone knows this. For example, if a player wins with a hand of tanyao dora 3, you'll have to pay 7700 points. But if the player won with a tsumo, you only need to pay 2000 points, thus saving a lot of points.

Everyone please note, do not compare it with Chinese international standard's and Taiwan mahjong's payout, where all players have to pay when someone wins. If the payout for dealing in and tsumo is the same, it will not encourage players to betaori.

The special characteristics of position battle
In Cantonese mahjong, if you deal in this round and lose $100, you can offset the loss by winning $100 next round. As everyone is only concerned about their own wallet, no one would care how much the other players win or lose.
However, that's not how it works in Japanese mahjong.
Japanese mahjong compares position, not points. Tonpuusen last hand, you deal in to a 12000 points hand and drop from first to last place. If you win a 12000 points hand the next round, would that offset your loss from the previous round? Of course not.
Also, because of special characteristics of position battle in Japanese mahjong, the point sticks of the other players will affect the outcome of your game. For example, if you deal in to a 8000 points hand, not only will you have a 16000 points gap with the player you deal in to, you will also have a 8000 points gap with the other two players. If the player tsumo instead of winning from your discard, there wouldn't be this problem. Therefore when dealing in while playing Japanese mahjong, although the points are only given to a single player, the reality is that you lose to the other 2 players as well.

To iterate my point, defending is an important skill in Japanese mahjong. I understand that sometimes giving up a win is depressing, but this is an important process in Japanese mahjong. If a beginner wants to quickly improve, practicing to abandon a win is a good method, as attacking too much is a common problem with Japanese mahjong beginners.