Re-promoted from 5th to 6th dan in Tenhou (Mahjong): Comparison and analysis of areas where improvements were made.

I was demoted from 6th dan to 5th dan in early December on "Tenhou," which was disheartening, but I've managed to get promoted again.
It took me about 400 matches to go from 5th dan to 6th dan the first time, but this time I was able to get promoted in just 53 matches.
While I admit there was a fair bit of luck involved, I've managed to improve on some of the areas I was working on, so I'd like to analyze the results.
We will compare our performance when we first reached the 5th dan rank with our performance after being demoted to see if we were able to improve.

(When I first reached rank 5)

(This is my record since being demoted to 5th dan.)

*Some of the displayed values, such as "average turn order when reaching a winning hand," are slightly inaccurate.

Points that were the problem

The key points I highlighted in the previous article were:

- Improved overall defensive awareness
(Consider keeping your hand hidden, being wary of two open melds, and refraining from pushing when you're one shanten away from a winning hand.)
- Aggressive calling, creative waiting strategies for awkward shapes, and building hands that will result in good shapes.
Those are the two points.
Speaking of defensive challenges, dealing with Dama was quite difficult.

I was quite concerned when someone who was in a position where they could have folded, discarded a slightly risky tile against someone who had declared riichi.

If you're playing defensively and luckily keep drawing safe tiles and discarding them, it's impossible to know what's going on, and I suppose there's no point in worrying about it.

From the middle of the game onward, I treated players who had two melds with the same caution as those who had declared riichi.

Also, when facing a declared reach, I made sure not to push if I was one step away from a winning hand, unless it was a very good reason.

Regarding the attack, we aimed for an aggressive approach.

If there were still a lot of awkward shapes remaining, I would quickly cut my losses and call a tile to avoid falling behind.
Since there doesn't seem to be any quick fix for waiting strategies or making adjustments to create a good shape, I've decided to stop worrying about it for now.

Winning data

Win rate 21.3% → 24.7%

Earnings upon winning: 6999 → 6658

Income during the period of deception: 6577 → 9172

Winning average order 11.620→11.073

Furo winning rate average order 11.605→11.304
Average success rate 11.630→10.838

Reach rate when winning 47.3%→41.6%
2 furo rate at winning 20.6% → 27.7%

The numbers related to winning hands were a little too good, so it seems like I was in pretty good form.

The decrease in income per win and the earlier average turn to win suggest that I'm able to play an aggressive, quick-calling style.
These numbers clearly show that improvements have been made.
I think the sudden increase in my income from deception is just a coincidence.
The proportion of manganese in clumps has increased, so that might also be a contributing factor.
If it's an early turn, it seems that it's often better to try to hold off the opponent with a reach.

Discard data

Dealing into a winning hand: 11.41 TP3T → 9.91 TP3T

Number of shanten (shanten) at the time of discarding: 0.791 → 0.745

Discarded to Fuuro 46.3% → 40.0%
These are also very good numbers, but the sample size is small, so we shouldn't get too excited.

However, the reduction in the number of tiles needed to complete a hand when dealing into a winning hand demonstrates an improvement in addressing the problem.

I often told myself that there were very few situations where I could attack by discarding a dangerous tile when I was one step away from a winning hand, so I simply folded.
I've also reduced instances of dealing into other players' melds, and I think this is a result of treating two melds from other players (mid-game onwards) as a "quasi-reach" and carefully considering when to push or fold.
In other words, if my own actions seem very slow, I'll prioritize minimizing the number of goals conceded rather than pushing too hard.

Reach data

Reach rate 18.6% → 16.8%
Revenue at the time of reaching: 9244 → 8900

I think this is simply because the rate of calling out has increased, which has led to a decrease in the rate of reaching.
Consequently, income will naturally decrease.
 

Draft data

(2 decoys on the left / 1 decoy on the right)

Secondary dew rate 10.9%/28.6%→15.9%/35.9%

Secondary dew average order 9.053/7.450→8.773/6.317
Percentage of players who open a hand first: 12.21 TP3T / 60.71 TP3T → 17.71 TP3T / 70.91 TP3T

Vice Russian time winning rate 40.1%/30.1%→43.2%/34.7%
Dealing into a winning hand while open: 14.01 TP3T/13.51 TP3T → 11.41 TP3T/11.61 TP3T
A very significant improvement can be seen.

We were also able to raise our rate of making at least one meld to our target level.
I think it would be better if it could sing a little more. (Around 37-38%)
Taking the initiative and making a meld isn't always the right move, but it seems that when I do make a meld, I'm able to take the initiative about 70% of the time.

Since the winning rate when you declare riichi is close to 60%, it seems that calling tiles should be considered roughly equivalent to declaring riichi.
However, naturally, exposing tiles reduces your defensive capabilities, so your chances of dealing into a winning hand worsen by about 2% compared to normal.
Since the rate of dealing into a winning hand after declaring riichi is around 9%, declaring riichi might be the next most effective defense after simply folding.

I believe that having a good hand ready to go first is both the best offense and the best defense—that's a kind of truth.

Ranking data, etc.

Number of matches: 402 → 53
Average rank 2.443→2.170
Table average R 1941→1957

Since neither side has played many matches, there isn't much to say about them.
However, I think it's the result of improvements coming together with good waves, so I'll just be happy about it.
I'm also a little pleased that my results this time were better, even though I was playing against opponents with a slightly higher average rating than before.
 

Useful ideas and perspectives

The basic goal is to reach a ready hand (tenpai) around the 10th turn.
You need to determine whether you're going for a Tanyao-type hand or a hand that relies on honor tiles.
I assess how many more sets of tiles (excluding the head) I need to make, how many of those sets are awkward shapes, and whether I can finish within about 10 turns.
If you're in a situation where you need to go on the offensive, you should start calling tiles early on to make hands like flushes or Chanta.

If you have even one dora tile, you can compete not only with a closed hand, but also with a player who has called a tile and is waiting on a two-sided wait, at least against a child player who has declared riichi.

Thus, whether you're playing with your hand closed or calling tiles, the important thing is whether you can reach a ready hand (tenpai) by around the 10th turn.
By calling a tile, you can put pressure on other players and lower the stakes of the game, and if your hand is bad, you can also fold.

Another important principle was that there are very few situations where you would discard a dangerous tile and push against another player who is already in a ready hand when you are one step away from a winning hand yourself.
Especially with Tenhou's rules, which are very strict about losing last place, preventing point losses is of utmost importance.
Meenin's book "Mahjong tactics in statisticsThe following is included in "Winning Value IndexThis approach helps alleviate the stress of getting off the train.
To put the part of that book I referenced in extremely simple terms,

- When you have a low score, the value of your winning points increases as the game progresses.
- When you have a high score, the value of your winning points decreases as the game progresses.
That's the gist of it.

In Tenhou, the rate at which the value of a hand decreases is extreme; even if you have 24,000 to 26,000 points, in the second round of the South wind, the value of an 8,000-point win becomes 0.71 times that amount.
If I have a score close to the starting point, I should be around 2nd or 3rd place, and the last place should be someone else.
In that case, the potential rewards of taking on the risk of falling to last place become considerably less.
If that's the case, the decision of whether to push or pull when you have more points than the starting score will be heavily biased towards pulling, which I think will result in fewer fourth-place finishes.

Conversely, if you only have 14,000 to 16,000 points in the second round of the South wind, the value of an 8,000-point win jumps to 1.57 times its original value.

In other words, to avoid last place, the decision becomes to press hard if you have a mangan.

Mahjong Ranking

Nihon Blog Village