Wednesday Meditation

Wasn't able to do a review yesterday because I was a little tired. Going to limit these to one in-depth game review / day, as anything else will take too long.

pts after session: 860/2800pt

Music: Shojoskip - Cosodorokitsune

A staple album of mine that I've been listening to for the longest time. Very familiar and has carried me through many a Tenhou session.「glory of the snow」has to be one of the best shoegaze tracks I've heard.

Mental: Mental state felt good overall. No real stressful moments outside of the initial 7700 deal in, but I recovered pretty soon after. Got mildly distracted thinking about the possibility of a Tenhou version of Dota chatwheels. Might've ALT+TAB once or twice during the first game. 2nd game had a lot of good reads, though I was still uncertain about a few decisions and made some small mistakes.

Game 1: 2nd (+45)

Nothing too exciting except for the first hand, where I could've thrown 5m on the read that both toimen and simocha are probably not using 78s. 5s -> 78s slide is also a read I should make in the future, especially since it came after the """safe tile overflow"""?

Game 2: 1st (+90)

East 1-0

Pretty fundamental chi, but it's still important to understand why 58m is a good wait here. I see 2 8m, and I have a mini-wall of 6m. Wall of 6m is important because it makes the 8m harder to use, and also strengthens the read on simocha's early 9m; 6789m -> 9m is less likely. The fact that we see 2 8m is also important, because it is a lot less likely that someone is holding a toitsu in their hand. 5m we can generally assume at least 1 wall, so ~3 58m wall should be the read here.

East 2-0

Pretty careless and uncharacteristic mistake.  I seem to remember being drawn to the idea of drawing dora and switching to pinfu for a yaku, but there's no reason I can't throw 8p for now.

East 2-1

Lots of stuff going on here. 3s is probably better for agari-rate, since simocha is confirmed not using it. No signal from toimen or kamicha, but we can probably assume 1-2 wall, and 3s will 100% drop from simocha.

Things I missed:

I chickened out because I didn't want to feed simocha tenpai / risk dealing into simocha's potential mangan if they happen to be in tenpai. In hindisght I should probably push the south and dama on 3s here.

East 2-2

No reason to risk riichi here into two other riichis, but there isn't much to fold with either. 1s -> 1s is probably the correct play, as we can get into a good tenpai with 58p draw. Really don't know what the correct play is here...

East 4-0

Kan 3p is weak, but this tenpai is just too good to pass up. Hard to discern how many 3p are in the wall, but I assumed 1 and rolled with it. Got a really lucky tsumo here.

Things I missed:

Toimen's discards are really suss; 5p wall + 2 4p makes a bit harder to use 3p, even if it's next to the dora.

South 2

Read is that 9p is 2 wall here, since simocha cut early 8p and kamicha cut early 9p. Assuming toimen's hand is random, and wall is ~5x the size of toimen's hand, so 2 wall chance is (5/6)^2 ~= 67%. Since oya is open and pushing in South 2, decided to play variance control and stayed dama here.

Things I missed:

Actually, the calc can go further than that, because toimen's discards, and by extension toimen's hand, aren't actually random. Given toimen's discard pattern in this round (and really I should be tracking prior discard patterns too), I can assume that their discard order prioritization looks something like: Lone 19 tiles -> Wind tiles -> Yakuhai tiles. Because I see 3 8p, the only way that toimen has 9p in their hand is if the 9p is connected to an 8p or 7p, which makes the 2 wall chance even higher than the original 2/3 estimate.

Now, this isn't rigorous at all, but I can use these probabilities to get a better estimate of what the chance is. It's something like P(9p | 8p) * P(9p | 7p | !8p).

The chance that toimen has the last 8p is about 1/6, and the chance that toimen has at least 1 7p is 1 - (5/6)^4 ~= 1/2.

So, the chance that toimen has at least 1 9p is roughly 1/6 * 1/3 + 5/6 * 1/2 * 1/3 ~= 20%. Which means the chance that we have 2 9p in the wall is ~80%.

South 3

Riichi is correct, but there's some better reasons as to why.

Things I missed:

Early 9m from simocha, early 9m and 7m from kamicha, early 6m from toimen. All of this is evidence for 3 7m in the wall; at worst 2 7m, with kamicha or simocha holding 1. Only way there's 0 wall is if simocha is going for toitoi with a 7m ankou in hand, but that possibility is extremely unlikely.