/ Tenhou Review

Game Log 1

During a game I have a piece of paper and pen in front of me. Any time I encounter a situation I'm uncertain about I'll jot down the Round and turn number, along with some notes if needed. These uncertainties can be anything from intersting hand formations, to Attack/Defend questions, Riichi/Dama, or even what certain discards of my opponents mean. Using these uncertainties I review the points of the game I wrote down. I won't write down all questions here, as many are a sanity check. If you yourself see the game and have any questions or concerns about my play, please let me know!

Date: 17/07/29
Game Type: 特南喰赤
Link: http://tenhou.net/0/?log=2017072912gm-0029-0000-05b6ae7e&tw=0
3位 A:Dasuke(-30.0) B:Dorako(+77.0) C:ふぇいとん(0.0) D:Px10(-47.0)

E1-0 4th Turn:

Pon Double east and push the early riichi?
Non dealer gets an early riichi. As dealer, I currently have a 3 shanten hand with about 4 confirmed blocks. Double east comes out. Take it and go Full Push since its early in the round with no safe tiles as well as dealer? Or pass it up and take the two extra safe tiles given?

Kinda split on this one. The dora is a hard to use 9m that can only be used in 789, 99, and 999 simple shapes; which compared to a 3-7 dora that can be used in five shapes, its slightly more unlikely they have some in early game.

I called in game but this hand seems to be a bit far away still. If this had a single block more, for instance instead of 48m having 34m or 78m, I definately think it would be worth pushing this early with the information given.

E2-3 13th Turn:

In face of last place riichi opponent drops a seemingly dangerous tile, are they pushing?
6s is pretty live. While most of the sozu seem to be pretty safe at this point the 36s suji is still unaccounted for. Unless they had two more 5s in their hand it seems to be pretty risky. Even the two tiles before that were signs of pushing. The 9m is questionably pushing since it might have been their best option with limited resource. Same with 2m but a bit more unlikely since its a complete no suji, in our view at least. Throw in the 6s and its pretty safe to say they're still fighting to some degree with their hand.

The result in this case:

E4-0 7th Turn:

Tenpai with 3 dora, but on a pretty bad wait. with the 2m through 8m shape, drop 4s and try to improve wait, or take the riichi?
Dropping 4s here the following tiles would give this hand a better wait: 134679m6p35s. 9 types of tiles with 24 individual tiles remaining. For reference in many mahjong books it suggests that if you have the option between a 4 tile wait and the chance to improve the hand with two 4 in a row shapes, take the improvement route if its not too late in the game. Assuming all tiles are available and the shapes are between 2 and 8, each shape gives 14 tiles each, or 28 advancement tiles in total.

Given that there were 24 available tiles, I really like the advancement path. The only potential deterent is that mangan level...but even then, the winning rate seems higher with advancement.

S1-0 12th Turn:

Running away from not only a riichi but both opponents who had calls and dangerous drops after the declaration, showing their intent for a showdown.
I dropped 6s the turn before as it become safe. The player to the right also cut their drawn tile, so it looks to be safe there. The problem lies with North seat, who just called 6p and dropped 9s. Why were they keeping 9s this late? Possibly because it was a dangerous tile up untill last turn, or they could have a shape they were using it in within their hand. Instances of those shapes include: 889, 899, 779, 799, as well as 579. The 579 in particular makes dropping the 6s again a bit scary.

During the game I got so focused on dropping my pair of 6s I didn't process that I had drawn an also safe 1p tile the turn before. So when it came down to this turn, even though I was uncertain about the 6s drop it never occured to me to drop either the 4m or 1p tiles, which seem to both be reasonably safe for all players involved.

Getting too focused on your immediate set of actions can be pretty bad if you're tunnelvisioning on them to the point that the ever changing choices and situation is ignored.

S2-0 6th Turn:

At the crossroads between a shiny hand a winning one, take either the 2.6k tenpai with potential advancement to toitoi, or commit to a 12k hand chance with 1p call?
I posted this one on Twitter. xKime, also known as Kyoukai Pro Nicolas, had the following to say: https://twitter.com/_xkime/status/891176217054855169
Which basically is to cut the North, and if you're thinking of point upgrades than bet on 67p pull for points

Especially with this close of game I think the tenpai would of been a better choice here than the pipe dream honroutou.

S2-0 8th Turn:

Last place in S2, 3rd place has declared a 6th turn riichi. Attack? Defend? Both?
I really don't wanna deal in here. A direct hit by close 3rd place at this point with no dealer turn left going into South 3 doesn't bode well with me. Since thats the case 58m5p seem pretty difficult to drop. If I really want to win this hand, there is a route that has not only the potential to win, but also call for it. 2m is currently walled off since I can see all 3m so the only shapes able to wait on that now are 2m and 22m. By droping 2m and later the two 1p, this hand could advnace to a tanyao dora 3 and also call for tiles (not that they're likely to come out though).

The reason 2m first over 1p first is that if this hand DID get tenpai by drawing the 4m, left with a 36s wait, I think fighting with this hard becomes worth a 5p drop at that point.

S3-0 2nd Turn:

An interesting WWYD.
Loading this up into Single Player Mahjong Simulator (Hitori Mahjong Simulator - HTM: http://ara.moo.jp/mjhmr/) We receive the following:
Which show Estimated Value wise, the 4m is the best drop in the scope of the simulator. This simulator has a couple of limitations. First off there are no red dora. Second it doesn't account for ron or calling as this is for single player, or tsumo, only. Finally all estimates are calculated with the assumption of draws all the way till Exhasutive Draw.

By changing the parameters here around a bit, for instance if 9m Wasn't the dora and a 7m was discarded at this point, the answer changes over to dropping 9m instead. Simulators can't give you perfect answers, but they can give you some very useful numbers including tsumo tenpai chance and winning chance, along with showing how the changing characteristics of a hand effect the best available choice.


I should probably stop with this. Note to self, of the various questions I point out during a game, only add about 3 or so to the post. Else its gonna take way too long to get through reviews before I can start the next game... And with that off ot the next game!