Reading Riichi (1)

When people talk about "reading riichi," they do not mean accurately guessing the exact winning tile. Doing that is extremely difficult. It would hardly be an exaggeration to call it impossible.

Example 1

Suppose you "read" this discard row as having been one-shanten before riichi, in a shape like A below:

Assumption A
Tsumo

Even then, you still cannot narrow the wait down to a single point.

If the hand entered tenpai by drawing from to , then the winning tiles become to . If it entered tenpai by drawing from to , then the winning tiles instead become to .

These are different tenpai entry routes. As long as multiple routes still exist, you cannot read the final wait down to a single tile.

And with a discard row like Example 1, even if you write out other candidate shapes, you still cannot really rule them out:

Assumption B

Assumption C

Assumption D

At bottom, trying to "guess the exact wait" is the wrong target. What you should really be reading is:

• which tiles are especially dangerous
• which tiles are relatively safe

That is the true first step in reading riichi.

Theory

It is impossible to read a riichi wait exactly, and there is no need to do so.
The real skill is recognizing which tiles become especially dangerous, and which become quite safe, against characteristic riichi discard rows.

Below are examples of high-risk tiles.

Ura-Suji on 5

You can find detailed explanations of ura-suji elsewhere, so I will not repeat all of it here. In real play, the two most genuinely dangerous cases are basically:

1. ura-suji on 5
2. yonken-suji

First, ura-suji on 5.

There is an old rule of thumb: "When a 5 in one suit is discarded early, be careful with 1-4 and 6-9 in that same suit." I do not have statistical data for it, but in actual play it is definitely noticeable enough to respect.

Here is Example 2 from the original page:

Example 2				↓			↓	↓

The notable feature of this discard row is that came out very early. And since is already in the pond, I would rate through as clearly more dangerous.

The actual hand was:

Of course, the final wait could also easily have been in souzu or manzu. But at least for pinzu, the line from to is the main danger zone.

The key is that was an early discard. Because a highly useful 5 was discarded in the opening, it makes you suspect that the opponent may already have shapes like 23 or 78 in that suit.

If the 5 had been cut shortly before riichi instead, that would be completely different. It could have been a floating tile, a leftover from 57 after drawing 8, or a fixing discard from a shape like 455. There are too many possibilities then, so its ura-suji would not deserve special fear.

So even though Example 2 later also cuts , that does not mean to is especially more dangerous than the other no-suji tiles.

Theory

Against an early-discard 5 in one suit, be especially careful with 1-4 and 6-9 in that suit.

Yonken-Suji

There are four basic yonken-suji patterns:

Condition 1	Condition 2	Dangerous Tiles
3 discarded early	8 is in the pond	4 / 7
4 discarded early	9 is in the pond	5 / 8
6 discarded early	1 is in the pond	2 / 5
7 discarded early	2 is in the pond	3 / 6

The two related tiles are five apart, which is why this is called yonken-suji. And once again, the key word is: early discard. Yonken-suji only becomes meaningful when a useful tile from 3 to 7 is discarded early.

Take Example 3 from the original page:

Example 3

↓

↓

↓

(↓ means tsumogiri)

Since the opening is almost entirely honor tiles, the discard of on turn 7 is relatively early.

And because has already passed, the main manzu danger line becomes to . As for pinzu and souzu, there is not enough information, so they must simply be treated as "all dangerous."

The actual hand was:

Now look at Example 4:

Example 4

↓

↓

↓

↓

(↓ means tsumogiri)

This is a point that is often misunderstood, so it is worth stating again.

In Example 4, through is indeed a dangerous yonken-suji zone. But because is discarded only in the late hand, through is not especially dangerous.

If the shape had simply been something like , then would normally have been cut much earlier.

So yonken-suji certainly can be dangerous, but if you misread the conditions that make it valid, it means nothing.

The Suji of a Ryanmen-Taatsu Drop

This reading pattern is essentially derived in reverse from the tile-efficiency rule that nido-uke is bad.

Example 5							↓	↓

(↓ means tsumogiri)

The most striking feature of Example 5 is that the riichi comes after dropping the ryanmen taatsu of .

Many beginners think: "If the player threw away that ryanmen, then surely to cannot be dangerous." In reality, that line may be the main one.

Imagine shapes like these:

Assumption A
　

Assumption B
　

In shapes like A and B, the player may discard precisely because they dislike nido-uke. And the waits left behind may then be on the line of , , and .

So when you see a riichi after a ryanmen-taatsu drop, you should first suspect nido-uke. That suji is often dangerous.

Theory

When a riichi comes after a ryanmen-taatsu drop, first suspect nido-uke.

Matagi-Suji from a Kanchan Drop

This pattern is more conditional, but its danger level is very high.

If the discard order shows a fairly separated outside-to-inside pattern, and the riichi follows a kanchan drop, the matagi-suji is dangerous.

Look at Example 6 from the original page:

Example 6				↓	↓				↓

(↓ means tsumogiri)

At first glance this discard row looks unremarkable, and all no-suji tiles may seem equally dangerous. But one suji line stands out as much more dangerous than the others:

to

Here is why.

You can see that and form a kanchan-drop sequence. But because unrelated tiles like and are discarded in between, this cannot just be a simple "I had a kanchan and later gave it up."

If were merely being kept as a safe tile, the order would not look like this. And if the idea were simply to draw and make ryanmen, that is also not very convincing, because at that stage would be the more natural discard.

So a more plausible interpretation is something like:

with riichi declared after cutting one . If that is the case, the matagi-suji of to becomes especially dangerous.

Even if all the conditions are not lined up this perfectly, as soon as you see a declaration tile that comes after a fairly separated, outside-to-inside kanchan drop, you should recognize that the matagi-suji is often dangerous.

I think reading danger tiles against riichi only needs to go about this far. More important than pushing the guesswork any further is the next topic: safe patterns.

---

Original Japanese page: http://beginners.biz/mamori/mamori14.html

Previous: How to Avoid Dealing In Next: Reading Riichi (2)