Preface
In the online mahjong game "Mahjong Soul," after dropping from Mahjong Saint 3 to Mahjong Master 3, I've finally made it back up to Mahjong Saint 2.
(Perhaps it was beyond my means, so saying "I went back" might not be appropriate...)
I haven't quite reached my previous peak yet, but I feel like I have more skills and techniques at my disposal compared to when I was a Mahjong Saint 3.
I've been thinking a lot about what it means to deal with mahjong that focuses on avoiding last place, and I've been trying out various things through trial and error.
Meanwhile, I recently read a tactical book...Mahjong AI NAGA's Ironclad TacticsI found this to be quite helpful to me, so I'd like to share it.
What kind of book is it?
Speaking of Mahjong AI, there are nowadays like "SuphX," "Bakuda," "LuckyJ," and "Mortal," but this book explains the tactics of "NAGA," which is publicly available for a fee.
Mahjong AI can give you a single, definitive answer for each situation, but it doesn't explain the reasoning behind its decisions.
The observer must analyze what criteria are being used to make that judgment.
This book clearly explains what NAGA likely prioritizes.
"NAGA" allows you to change the situation when reviewing game records.
You can experiment with the values of each element to see how the judgment changes.
For example, in a situation where one side has to push and the other has to pull, let's say NAGA decides to push.
So, what happens if the tile you push is a dora tile? What happens if it's not a sequence tile? What happens if your opponent is the dealer? What happens if the score situation is different? You can change these values and verify how your decision-making changes.
By analyzing these changes, NAGA's ideas are generalized to some extent, and concrete examples are provided to make them easier for readers to understand.
The content is divided into four chapters: Chapter 1 - Hand formation, Chapter 2 - Calling (deciding when to call), Chapter 3 - Deciding when to declare Riichi, and Chapter 4 - Pushing and folding.
Each chapter deals with a common scenario, and I think you'll find it useful in many situations.
The parts that resonated with me
There are many points in this book that I found helpful, but since it's a commercial product, I can't go into detail about its contents here.
However, I will write a little about the parts that I think won't cause any problems if I write about them.
In the first theme of Chapter 1, "Hand-Assembly," of this book,"Don't overlook the possibility of a Honitsu (all one suit) when you open your hand!"That's what it says.
I can't go into details here, but basically, even if a hand is disjointed and doesn't seem likely to win, NAGA will still be quite conscious of the possibility of a Honitsu (all one suit) hand.
(Even if it's just a Honitsu without that)
僕はラス回避麻雀を打っていくにあたって、放銃率が上がりそうな副露はどうしても控えめになってしまいがちでした。
Even if you play with a disorganized hand and focus on calling tiles, you'll end up having to fold if another player declares a reach.
Therefore, the rate of open calls remained relatively low, around 30%.
However, as I put into practice what was written in this book, I noticed something.
The advantages are that even if you have melds, you can still fold fairly easily, and if you have two or more melds, it's like a semi-reach, and other players will be wary of you.
I noticed that when I made a particularly blatant attempt to conceal a hand, other players who weren't aiming for a win would hold back and play defensively.
While I wouldn't recommend calling a tile as a bluff, I thought that forcing a hand into a Honitsu (all one suit) when you have a bad hand isn't a bad idea as it lowers the chances of other players winning.
Of course, if I also draw honor tiles or dora tiles, it's not uncommon for them to make a hand worth 7700 points.
If you keep playing with a closed hand because you have a bad starting hand, you'll often end up not being able to reach the point of declaring "reach," being unable to do anything, and having your points chipped away by other players.
I've come to realize, rather belatedly, how important it is to use moves like "reach" and "opening tiles" to play in a way that forces other players to be wary.
With this mindset, my rate of calling tiles naturally increased to around 32% to 34%.
There were many other things that were very helpful and made a lot of sense to me.
I usually have the mahjong AI "Mortal" analyze the hand after I've played, and then I quickly go through it.
When I look at the analysis results, I sometimes think, "Was this my mistake?" or "No, this is correct," and I try to think about the reasons why I arrived at this conclusion.
"NAGA" and "Mortal" are different AIs, but I don't think the factors they prioritize in their decision-making are completely different.
Perhaps because of this, even though I don't usually interact with NAGA, there were many parts of this book that I found myself nodding along to and thinking, "That's so true!"
Of all the mahjong strategy books I've read recently, this one was the most convincing and easy to accept.
Afterword
The aboveThe ironclad strategy of the Mahjong AI "NAGA"I briefly introduced the contents of "[Title of the book]" by relating them to my own recent activities.
Lately, I've been enjoying trying to force a hand that's a mix of random tiles and likely to be cheap into a single flush.
I think I've started to enjoy not only getting hands that seem likely to win, but also how to make my opponents wary of hands that don't seem likely to win.
I think there will come a time when you start losing a lot and lose confidence in your playing style.
However, the experience of falling and then climbing back up should serve as a guide to help you maintain your composure.
I hope to reclaim the title of Mahjong Saint 3 in 2025.
A collection of articles on improving your mahjong skills, techniques, mental game, Tenhou (online mahjong platform), etc.