Game log reviews accelerate improvement faster than grinding games. Players regularly ask me to review their Tenhou logs, mainly on Tenhou. Many have no idea what a game review comprises beforehand. So I'm posting an actual example of a game log I was sent recently to review. Maybe you can find some mistakes in common, and it can help you improve as well.

Mahjong game review for Tenhou (+10 bonus key advice points)

Audit, comments, guidance

Click to expand

Let the (mahjong) game (review) begin

Introduction: reviewing what to fix about your mahjong game

Thanks for sending this game log. Here's what needs work.

My first impression while looking at this game is that you don't need to dig into advanced techniques. You must fix the wounds in your fundamentals first, especially tile efficiency.

You're prone to going back shanten for no apparent reason and choosing discards that aren't optimal for defending or attacking. Your discards must be good for one of those two things (or both).

It also seems you don't know how to handle your honor tiles. You systematically discard all of them at the start no matter the kind of hand you receive. I'd like to fix that.

Hypothesis: how to fix and improve your mahjong skills

If you review basic shapes over and over, you'll stop making basic tile efficiency mistakes. There were many errors that, if you stopped and thought calmly, you would have realized by yourself.

Tenhou has a time limit, and that's the likely cause for those errors. The only way to reduce your thinking time is by throwing in more practice. "Quick thinking" is not something that can be taught. It has to be drilled. Here's an effective drill:

Grab 13 random tiles. Draw and discard quickly toward tenpai without sorting your hand. This is how a few parlor owners train their new members to play faster. This training will cut your thinking time in half when you're finally able to play with your tiles sorted. It gives you more time to think about the strategic side and spend less time figuring out the shape. You'll also learn to recall your hand without looking, which is a powerful skill.

Things you will need after this game review

I would like you to start playing while thinking about shanten number. If you decide to take a step back (or forwards), try to do so consciously. You don't need to count when you're farther away than 3 shanten, but be aware when you go from 3 shanten to 2 shanten, or when you go from 1 shanten back to 2 shanten. Internalize it.

Materials

1. This book (you don't need to understand the Japanese; look at the problems and answers for now)
2. This training (as described above), mahjong tiles for it
3. Perseverance

Procedure

1. Become able to recall every correct answer of the book, one chapter at a time.
2. Then try to think by yourself of the reason why that discard is the correct one.
3. Practice the discard exercise for 30 minutes as explained above.

Mahjong game review from a Tenhou log

(Game log link removed for privacy)

East 1, 0 Honba

Click to expand

East 1, 0 Honba

This hand has no dora, no yaku, and only one ryanmen. Pairing yakuhai will give you the speed boost this hand needs to make up for its low value.

The 1p and 9p are less useful than your other tiles. If you draw 2p, 1246p isn't an efficient shape (1p is unneeded). Same with 4689p (9p is redundant). The 13469p or 14679p shapes are slightly okay, but since you have a completed sequence (567mshuntsu) and a ryanmen on souzu (78sryanmen) you're not going for the Ittsuu here. In the end, 1p and 9p are only useful if they get paired up. Same as your haku and hatsu, except those will give you a yaku and a chance to call if you pair them up.

Considering defense for the future as well, 1p and 9p are not useful to you, but they might become someone else's winning tile later. Discard either 1p or 9p and keep your yakuhai in hands like this.

Key point #1

Hands with low value need either a speed boost or safety to compensate. ~NPMahjong

Click to expand

If you discard 2m, you're missing the ryanmen improvement by drawing 3m. Ittsuu is too far away, and it's only a 2 han yaku. Pinfu is worth only 1 han less, but it's immensely easier here. To get that ittsuu here you need to draw 2p 5p 7p 8p in the right order and within the remaining draws (before anybody reaches and wins, too). That's not realistic. Discard 9p. (Keep 6p for ryanmen and possibly sanmenchan improvement and sanshoku evolution)

Click to expand

Key point #2

Many middle tiles discarded early on, and yakuhai as a first discard are signals of a fast, decent hand.

Defending here was a good call, good job. East opened up his hand, and his discards look fast (plenty of middle tiles and Yakuhai as his first discard, both signals of speed and value), while your hand is both slow and cheap. It was an excellent choice to fold. However, next draw...

Click to expand

This 3m is not a safe tile. Sure, it is also not "incredibly dangerous," but it's an unnecessary small risk. You have plenty of safe tiles to defend all the way to the end. The sum of these small mistakes will add up in the long run. Discarding 3m doesn't move your hand forwards, and it's not a safe tile (it can still deal into toitoi, chanta, junchan, ittsuu, sanshoku, and yakuhai ankou, etc.; all of them can use the dora as pair as well). Don't give people chances like that when your hand is as cheap and unwinnable like this one.

Key point #3

With cheap, slow hands it's crucial that you do not give people chances, however small.

Click to expand

Your hand is cheap, your tenpai chance slim: only 6s and 9s, 5 tiles in total. Even if you do get tenpai, your wait will be 4 tiles at best. On top of that, you're against a riichi and East who was already dangerous. 8m itself isn't "too dangerous" (given that you see three 7m), but it's an unnecessary risk for no reward. Choose 7m and fold correctly.

This hand is too terrible to close your eyes and chase it blindly like that. Save those points for when you get a decent hand.

Key point #4

When you get poor hands, don't gamble unnecessarily; save your points to invest them into your better hands.

East 1, 1 Honba

Click to expand

East 1, 1 Honba

Here you went back from 1shanten to 2shanten for no reason, and tearing apart the sanshoku you were trying so hard to build. You still don't see a single 7m on the board yet, so is there a reason why you'd break up your hand and discard a tile that is neither safe nor leads you to tenpai?

It's normal to discard 3p here. You might feel uneasy about having a penchan shape, but this kind of detour is only playable in the first row of discards. After that, you fall behind everyone. You're lucky it worked out in the end and you didn't deal into anybody.

Key point #5

Long detours are only playable in the first row of discards.

East 2, 0 Honba

Click to expand

East 2, 0 Honba

By this point you had probably already decided you'd go for hon itsu. Good. There's a technique you can use early on when you have decided to go for hon itsu to make your hand easier to win: discard the suit you have the most tiles in first. In this case, souzu tiles. By doing so, you don't reveal the actual color of your hon itsu to your opponents right away, making it harder to respond to your calls.

If you discard 1m 7s, then you immediately inform everyone that you're going for pinzu right away. If you dump 7s 2s 1s, then nobody can pinpoint whether you're going for manzu or pinzu yet.

Other than that, this hand was average, and anybody would've played it with the same result. That's a good thing. It's essential that you don't let the simple hands slip away.

Key point #6

When locking in towards hon itsu, start by discarding the obsolete suit you have the most tiles in.

East 3, 1 Honba

Click to expand

East 3, 1 Honba

The 3m becomes ryanmen with 2m or 4m. 9m never becomes ryanmen. The value of 3m is higher than that of 9m. Discard 9m. There's also a small chance for 234 sanshoku naturally in the distance.

For instance,

or

depending on the timing and order in which you draw your useful tiles.

Click to expand

There's no good reason to discard the dora ryanmen instead of the 58mryanmen ryanmen. Riichi, tsumo, iipeikou, dora 1 is already mangan chance. It's a waste. If you were playing with aka dora, 56sryanmen is also an option.

Click to expand

3p passed. You get a free pass to discard it safely and keep the larger 1 shanten. By shedding 5s, you lose the 47s tenpai chance, iipeikou chance, and the chance to discard dora safely. The only thing you gain in exchange is participating in the lottery of drawing one of the 2 remaining dora while dragging the risk of having to release it later on. I don't recommend you buy into that lottery. Keep the larger 1 shanten shape. Riichi, tsumo (and iipeikou) are already 2-3 han (consider the chance for ippatsu and ura as well). Menzen riichi is strong.

Key point #7

Mahjong is a game of choosing what lottery you want to participate in. Choose carefully, since every lottery has a different cost, probability, and payout.

Click to expand

If you're going to take the chance of discarding dora vs. oya riichi, you should at least riichi yourself for the value boost and for the right of winning on 36sryanmen as well. It's a better lottery.

East 4, 0 Honba (last of East round)

Click to expand

East 4, 0 Honba (Last of East Round)

Shape. Discard 5p. There is no need and no "natural" way of turning this hand into 345 sanshoku anymore. Drawing 5s would mean that you missed your chance for tenpai. You're forcing it too hard, and you don't need the value so badly right now. A good shape East Riichi is a stronger choice. Discarding 5p gives you the most considerable tenpai chance, all of which leave you with a good wait.

Key point #8

An early good shape riichi hand is your most reliable ally when you're East ~NPMahjong

Click to expand

The 5m passed, and two people didn't change their hands. Don't give your opponents the chance to call pon/chii/kan/ron on 2m. Discard 5m.

South 1, 0 Honba

Click to expand

South 1, 0 Honba

If you discard this live Chun 10 times, it's bound to deal in once. Because it's an honor, it doesn't mean it's safe. Since you're folding to this East riichi (because this hand is terrible in shape and value), it's an unnecessary risk. Discard 3p first.

Click to expand

East riichi and 1 person pushing. You're not supposed to push 2m while noten (and with that 7s penchan that's half dead). It's too naive. You're only giving people chances. Fold.

Click to expand

The 9p and 3p are safe, but 6p hasn't shown up yet. It's not "safe." Therefore, that's another unnecessary risk.

Key point #9

Double check your safe tiles. Suji doesn't equal safe.

South 3, 1 Honba

Click to expand

South 3, 1 Honba

Key point #10

Even if all options are "right", there is usually one better option ~NPMahjong

Disclaimer: All things considered, riichi isn't bad. It's a bet that will give you a good enough average placement. Also, dama wouldn't be a bad choice either. After all, if you were playing Tenhou Rules I might choose dama in some cases. So let's get that out of the way first.

I do believe that neither of them is the best choice under these (and under most of the existing) rules. The best discard is 8m for the points and shape.

The 3456mnobetan shape is so famous and strong that it has a name for itself, "yonrenkei". It can quickly evolve into ryanmen, or sanmen waits. You also keep the chance for 456 sanshoku. You can also improve the wait to a 36 nobetan after drawing 4p 7p. There are too many tiles that will enhance the hand, and it's still before the 9th draw, where you can use the extra points to improve your placing in the tournament. After discarding 8m, you will have the following hand improvements:

1. 2m (furiten 3menchan)
2. 3m (47p wait, sanshoku)
3. 4m (25m ryanmen wait)
4. 5m (47m ryamen wait)
5. 6m (47p wait, no sanshoku)
6. 7m (furiten 3menchan)
7. 4p (36m wait, sanshoku)
8. 5p (467p three sided wait)
9. 7p (36m wait)

And yes, you will riichi with a furiten 3menchan wait. A kanchan wait for a 3 or 7 and a furiten 3-sided-wait have the same winning chance. The latter is more expensive than the former (thanks to the extra han of menzen tsumo). Therefore, if you're willing to riichi this kanchan, you should be prepared and ready to riichi that triple wait.

Bonus Key Point

Reaching your furiten hands increases their hand win percentage, as it increases the likelihood other people will step out. A three-sided wait furiten riichi has almost the same winning chance as a middle tile kanchan riichi.

Considering all of the above, I don't fully understand why you chase improvements and yaku that are so out of hand in the mid-game, but then you won't pursue them when there are so many tiles and possibilities. Keep this kind of shape in mind. This could have been the turning point of this game.

As for what happened in the last hand in All Last, I don't think there's much you could've done to avoid that. There wasn't much of a chance to do anything.

Those were the critical errors. Given time to think, you can make the right choices. To improve that aspect, don't give up on studying and practicing shapes. Then, after you master the basic rules, you can begin bending them to your will.

"Learn the rules like a pro, so you can break them like an artist" ~Pablo Picasso

Want your own game reviewed? Pledge on Patreon, send me your log, and I'll make time for you.

If you can't, that's fine. Share this post with your friends online if you liked it. It helps a lot.

Good luck at the tables and see you next time.