Replay Review by Myself - Aug 15

This is a self-review of one of my replays. I've been trying new things, and noticed a lot of mistakes while playing the game, so I wanted to go through it in-depth. Maybe it can give you an idea of what to look at in your own replays, too.

In this game, I was focused on building higher value hands, as outlined in my Week 11 post. You should probably read that for the context before going through this replay review. With that in mind, let's focus on those concepts as I play, and hope to reinforce them in my mind.

The replay we'll be looking at is here: https://tenhou.net/0/?log=2019081508gm-0089-0000-b16b2027&tw=1

What can this hand become? Pinfu, if we get an 8s, is pretty easy, but that's not enough. A 234 sanshoku is a possibility, but we need three good draws. We don't really have a good route to itsu or iipeikou. Itsu is four tiles away, and iipeikou is three. Tanyao is a bit hard, as we'd need to cut our pair. The tiles with the best ukeire are 1m and 4m. The 4m is helping us accept the 0s, and contributing to the sanshoku, so we cut 1m to keep our options open.

Here's where the hand ended up. We managed to riichi with pinfu and a sanshoku / akadora chance. That clears our riichi + 1 condition, in my opinion. Good work!

With this hand, we have an akadora, so we've already cleared the condition. But, we can get more. Since it's still the first row, going for more value is fine. We're very close to tanyao, and can get it without much trouble by cutting this 1p, and later, the 1m.

The hand progressed very straight-forwardly, and we got a tanyao aka riichi hand. It even has a good wait! Another success, even if it was pretty much handed to us.

With this hand, we have three dora. We should progress straight-forwardly as efficiently as possible. If we can move the hand to tanyao or sanshoku and call for it, it would still be mangan, so we should keep that in mind, too. Sanshoku might require cutting the 5p, but it would still meet the 3 han requirement for open hands.

Here's an efficiency error, though. I cut the 8s since I wanted to lock in the 5p pair, but if I wanted to do that, I probably should have just cut the 6p. I could even start cutting the 12p here to enable tanyao. I think calling wasn't on my mind during this hand. When I have two or more dora, I should think about how I can call.

If this, then that is a common idiom. If something happens, do a specific thing. Most of my instructions to myself are in this form, so the new one is, "If my hand has two doras, consider how calling could become possible." Maybe I should put them on sticky notes around my monitor.

Here, I absolutely should cut the 12p. First of all, there are two 3p out and only one 7p, so the 68p shape is simply better. On top of that, cutting the 12p gives us tanyao, which lets us call to mangan. I realized this shortly after cutting it, so I think I just need to take my time more. I'll take my hand off my mouse after making a discard. The additional time taken to move my hand back to my mouse when making decisions should help me notice stuff like this.

This is similar. Here, I could cut the 6p. Then, if I got 3467m, I could cut the 12p and keep the better shape. But, I should probably just cut the 12p shape now. There's three 3p out already. It's unlikely to complete and a horrible wait.

Okay, next hand. This has a lot of promise. We're two tiles away from itsu, and we can accept the dora with our 4s shape, in addition to accepting every red five, technically. We should progress normally with 1s and 1z for now.

Now here's a choice. Of course, we want to keep the red five. If we're committing to itsu, we can just cut the 4p, which I do. But we could also cut the 1p, giving the possibility of tanyao and sanshoku in exchange for itsu. Sanshoku would probably require cutting the 0m, so I think I like itsu better. 4s can keep both options open, but it removes the dora acceptance. At this point, maybe we don't need it, though. 2m would lock in the 0m, but we could just get something to stick to it and complete the 23m shape on 1m, so I don't like that path. All said, I think 4p and 4s are relatively equal, but maybe 4s was better.

Since it's the second row, calling to tenpai is okay. But, the 2p isn't safe. I don't think I should have fought fourth place's riichi here. My out is suji, at least, but a 2000 point hand with a penchan wait isn't really worth pushing for. I have two genbutsu already.

Quite an awkward hand we have in South 1. We have a red dora, and a normal dora, so we can call these white dragons and have a three han hand. 789 sanshoku is a possibility, which could be a fallback if we don't connect to the 5s, or maybe 345 or 456 sanshoku to add value. Nothing else really jumps out at me. With all that in mind, 2s seems to be the most useless. But, I cut the 9m. I'm not really sure why. Maybe I didn't think things through enough.

As you can see, I passed up calling the first white. I wanted to guarantee that I could use the 5s before doing so, but I think I should have just called it. I could force it if necessary. Anyway, here, cutting the 3p is bad. It removes the chance of the 345 sanshoku. I should cut the 89s instead. I've already declined that sanshoku, and it's the same ukeire.

After discarding 5p, I should just tsumogiri this 0p instead of getting greedy. It's too late to be going for value. First row value, second row speed. Plus, it just passed last turn, so this is basically the best chance to get rid of it.

This hand is an easy tanyao just by cutting 9s. It also accepts the dora, 4s. That's pretty clear cut.

This was a bit silly, though. I should cut either 3s or 8s to add the 7p ukeire. So, which one? Cutting 3s removes dora acceptance, while cutting 8s removes the 678 sanshoku acceptance. But, we'd need to get an 8m to swap with our 5m for the sanshoku, so I think cutting 8s is better. If we needed the points, we could even then riichi on the 35s shape to guarantee it, but in this case we don't need to worry about that I think. We have our +1 in tanyao already, and 56s still accepts the dora anyway.

Now that it's the second row, I opt to call the dora. With two 7s already out, this is an awful shape for a 2000 point hand, so maybe I should have held off. It'd be a lot better if I'd kept that 8p. My call decision making still needs work, but let's ignore it for now. We're focused on looking at hand building. Something to keep in mind, though.

We have our +1 with the red dora, and it's at the end of the first row. It's getting too late to mess around with value, but for practice's sake, let's think about what we could get. Any sanshoku involving 5 is a possibility. Tanyao is pretty far away, as is itsu. Some lucky draws can get us pinfu. We can't accept the dora 1p. Overall, this just seems like a hand we should play efficiently and hope to riichi.

The next hand is in a similar predicament, with it being quite late. We can accept the dora 6m, and have the 0m. Tanyao is quite achievable, and cutting the 19m looks good. I suppose 1m is better to cut, since there's less ura chance for it, but it doesn't really matter.

Last turn, I cut 7m, which was a mistake. I should have cut 3m instead, since 7m connects to the dora. Looking past that, we can get tanyao if we cut that 9s, but we're also pretty close to itsu. I like just cutting the 3m now to keep both options open. 2p would lock tanyao in harder by removing the 1p acceptance, but I don't think that's worth it.

We ended up losing tanyao, but itsu is still open. Here, the 3p isn't really doing anything for us. We should cut that instead of the 2s. Then again, the 2s isn't really doing anything for us, since our hand has no value if we get, say, 7p, so we don't want to riichi anyway. A higher chance of reaching tenpai for tenpai payments, I suppose.

We have our red five, so our goal is met already. What else is possible? 123 sanshoku, or cutting all the ones for tanyao. By the time we cut all our floating tiles, we'll already be in the second row, though. I don't think we should cut the 12m in that case, since it is helpful and this hand needs all the blocks it can get. With shimocha having two calls already and our hand looking like this, it's not likely we'll win. But, we can't just start folding, as we're fourth in South 3, so proceeding normally it is.

I should have just riichi'd here. There are three sticks in the pot and five honba, on top of us already having a dora. If we win, we get second. I wanted a good wait, but there's no time to waste.

Here, with 6 sticks in the pot and 6 honba, we don't really need to care about getting a +1. The pot is our +1. That's something I also need to keep in mind when it does come up. So, we should proceed normally. 123 and 789 sanshoku are possibilities, though. I ended up getting to riichi on the 79s wait and got the ippatsu for second place.

Overall, I think I did pretty well aside from some silly efficiency errors. The two action items I'm taking away from this game are:

1) When my hand has two or more dora, look for yaku to call with. It's similar thinking to looking for value, but in this case, we're looking for speed.

2) Take my hand off the mouse in between turns. This will give me a bit more time to notice the state of my hand, and hopefully help cut down on silly mistakes brought about by haste.

Of course, there are other things. But I can only focus on integrating so much into my play at once. For now, I'll work on those, and fold in more when I feel like I have a better handle on my current things.