These notes are all written with the walls closed and hands hidden, just like a live game. I'm writing on the first view, without the benefit of hindsight. This is so that I can judge whether a decision was correct given the information available at the time. When I say a tile is dangerous, it's not because you dealt in, it's because it looked dangerous according to the pond.

http://tenhou.net/0/?log=2015122714gm-000b-11052-9b128c2c&tw=0

E1:
- Basic jihai order is otakaze -> bakaze -> sangenhai -> jikaze. In this case, xia -> pei -> hatsu. Personally I also like discarding isolated terminals before fanpai.
- 9p discard is poor. With the 9s pair, forcing tanyao will likely be slow. Should drop 2s instead to keep the 7p ukeire.
- Well okay you paired 2s, but with four pairs you should be ready for chiitoitsu, so discard 7m or 5p instead of dropping the 9s pair.
- Why discard 2s against that riichi? 2s was completely live. 1p was one chance. Even though he discarded 2p right before riichi, 1p was your safest tile there.

E2:
- Instead of discarding 5m, 2s would've put you in a very wide iishanten.

E3:
- The 35s kanchan had better chances for improvement than 24m, so you should've kept it.
- 788m566s, discarding 7m give you the wider iishanten.
- You've got kind of a weak iishanten for a small hand, and you need to push dangerous tiles to reach tenpai, so you should've folded against that riichi.

E4:
- Jihai order.
- 8s is wider than 3m.
- Why pass up tenpai? You've got two dora and you're dead last. Riichi would've been fine.

S1:
- Discarding 9s instead of 9p would've been a tiny bit better since it's next to the dora.
- Passing up tenpai here isn't too bad, but discarding 2p would be wider than 8m.
- Tanyao nomi? No point in damaten here. Even if your wait isn't great, it's worth declaring riichi immediately just to secure your dealer turn. Even if you don't win the hand, the noten penalties are enough of a benefit, and if you're lucky then they might be worth more than tanyao nomi anyway.

S2:
- On drawing 4s, I'd rather discard 1s instead of 2m.
- When you go full betaori, it's good to discard the most dangerous safe tiles first. For example, toimen discarded 7m, so you should follow up with that before discarding 8s, which is safe against all three players. This helps you to defend better in case other players push against riichi.

S2-1:
- Instead of 5m, discarding 89p would leave you with a better shape, plus tanyao potential.
- Toimen made a third call after riichi, you should assume tenpai and discard xia or 2m instead of 1p.

S3:
- Kokushi from eight tiles? Good luck. You're basically throwing away this hand, and that's going to make the last round much harder for you. Your goal here should be to lower the gap to third as much as you can.
- That said, if you're going to try for kokushi, it's good to discard two suits first to give the impression that it's just honitsu.

S4:
- Your options now are something like dabunan honitsu with a direct hit on third, or riichi tsumo pinfu ittsuu dora or something. Riichi with one dora and a direct hit on first doesn't help you at all.

http://tenhou.net/0/?log=2015122512gm-0029-0000-7d665a48&tw=0

E1:
- Should discard 9m before 1m.
- 8p before 5s.
- Hatsu nomi iishanten, better fold against that riichi. 1m may have been one chance but at that point in the round it wouldn't have been strange for shimocha's 3m discard to have come from 233m. Discarding 2m after that was even more dangerous.

E1-1:
- 1p and 1s might both seem to be isolated terminals, but 1p is a better discard. When it comes to early game tile efficiency, think of 134s as 13s and a single 4s.
- Discarding 2m would be better than 9m, considering the potential for future taatsu.
- 9p gives a much wider iishanten than 3s.
- You've got tanyao two dora for 40/3, which is 7700 as dealer or 11600 by tsumo. You don't need to riichi here.

E1-2:
- Why break up the 9m pair and pass up iishanten? The 45556p shape can easily be improved, and the odds of completing the 468m ryankan are similar to the odds of forming a souzu ryanmen.

E2:
- 5p gives the widest shape, but since it's dora, 3p is the next best choice. Either of these were much wider than 6p.
- Opening up here and discarding 3p locks you into waiting on 8p since 2p is already dead. You've already got a decent lead, you don't need to be so aggressive as to chii a ryanmen.
- Why break up your hand to hang onto the pei?
- Shimocha kept pei as a safe tile. Discarding it indicates that he's either tenpai or iishanten. 3p is one chance but since he discarded 27p there's a 36p aidayonken, so you shouldn't take the risk.
- Discarding 6m against that riichi is super dangerous. Given that 5p was kabe and there were three of both 78p visible, 6p was much safer. And then 8s, why on earth?
- That last 9s was a pointless discard. It's your last draw, there's no way you're going to reach tenpai now, so you should discard the 8s that just passed.

E3:
- Looks like you tilted here. 8s? 3s? Why hold onto the ton? If you're forcing tanyao, why discard 5p instead of 2p? If you're going for sanshoku, why chii 4m? Why then discard 6m instead of 2m?

E4:
- 4p?

S1:
- Discarding 4s practically locks you into waiting on 58m twice, which is relatively weak. Otherwise, you're stuck with an even worse shanpon. Better to keep the 5s ukeire and discard pei instead.
- Toimen discarded 2578m and 35s from his hand. There's a good chance he's waiting on lower pinzu, so discarding 23p is risky, especially after he discarded 1p. Oh, yeah, ron.

S2:
- I'd pon that 4p to go straight to tenpai on 25m.

S3:
- Riichi is a little aggressive here but it's fairly reasonable.

S4:
- Why break up the 12m penchan? Even if you wanted to keep the nan, 8p is a much better choice.
- Okay, tenpai. I'd advise staying damaten in a situation like this. 14s don't look particularly likely to come out, much less against riichi, so you're just exposing yourself to needless danger. Tsumo is enough to win.

http://tenhou.net/0/?log=2015111714gm-0029-0000-350b2447&tw=3

E1:
- 2p tsumogiri was too early. You've got a pair of fanpai dora, so you should maximize your speed. Drop ton instead.
- The 6s isn't a good tile to hang onto considering you just discarded 58s. Tsumogiri and keep the safer 9s.
- Should've kept the second 5p so you could improve your wait. The 35m kanchan is a crucial weakness.

E2:
- No reason to keep that nan for an extra turn.
- Should discard 7m instead of 8s.
- 1m? You're ryanshanten for sanshoku. It's better to just discard 7m for the widest iishanten.
- Damaten? There's one 1p out already. It's more like for 3m to come out or for you to draw it before you draw 1p.
- It's almost always a bad idea to tsumogiri riichi. If you're going to riichi, you should do it as soon as possible, or if you're set on sanshoku then stick with it. As it is, you gave the other players another turn to advance their hands, which will affect how they respond to your riichi.

E3:
- Should drop ton first. 2m is useful, especially considering it's next to the dora.
- 8s is wider than 7s. 466p788sEE, your ukeire are 56p689sE, discarding 7s loses 69s ukeire and locks you into a weak shanpon. 8s would leave you with 56p69sE, losing only the 8s ukeire.

E4:
- Jihai order. Or were you keeping xia as a safe tile? That's fine but pointless if you're going to discard it only a few turns later anyway.
- 2p is more valuable than 2m. The chance to draw 234p is the same as 234m, but if you draw 1p you end up with a better waiting shape.
- Look at shimocha's calls and think about the shape of his hand. He started with 33445s -> chii 3s -> 334[345]s -> discard 3s -> chii 234s -> discard 2p. If 2p was more useful to him than 3s then it's pretty much guaranteed that he's tenpai on lower pinzu, so discarding 4p was very dangerous. It looks like a cheap hand but you're only iishanten for chun nomi, you don't have enough of a lead to justify pushing that.
- Okay I checked shimocha's hand and I see that's not how things went this time, but only because shimocha's own tile efficiency was bad. You shouldn't assume that other players will make mistakes like that. The fact remains that 4p was very risky.
- Kamicha chii'd a ryanmen after passing chii twice, then discarded a middle tile. You should assume tenpai there, so 8p was dangerous.

S1:
- Why discard 2s to keep the chun? You lose 3s ukeire and you're left with a very weak shanpon.
- Now that you drew the 3s it'd usually be better to discard 4p instead, but since there are three 2s discarded already following up with 3s is fine.
- But on drawing 8p, 4p or 9p give wider ryanshanten than 4s. There's one 9p discarded but 4p is more likely to be used in other people's hands, so I'd go with that.
- Also, considering kamicha pon'd nan and discarded haku, you should expect tenpai. Since there are two other haku visible already, he could only have been holding it as a safe tile. Looking at the pond, lower/middle souzu is most suspicious, so that's another reason to discard 4p instead of 4s.

S2:
- 2p? Why keep hatsu?
- Why 2p again? 5m is the most direct path to tenpai, but if you want a decent shape you should discard 1s. You've got chun dora, with these scores you don't need to grasp for dora here.
- 8m may be suji but you're so far from tenpai that it's not worth chasing. Just discard 1s first, followed by 5m and 1p.
- 9m is safer than 9p. It's safe against kamicha and shimocha can only have a hell wait on it. Oh, plus you still have 8m that's safe against both.

S2-1:
- Discarding 7p right off the bat hurts too much. You've got a nice shape, and since there are two riichi sticks and a honba on the table you don't need to grasp for dora here. You'd at least discard haku first.
- You managed to pair the dora, but it's better to discard 4m before 9m. 9m will be safer later.

S3:
- A haipai like this is certainly beautiful but don't be tempted to overstretch yourself. Haku chun is enough as it is. You don't need honitsu. Your priority should be rushing the game to an end - you're the last dealer, so if you can make it to oorasu in first place, you've got a good chance of keeping first.
- Calling to break ippatsu is one thing (which I wouldn't advise in this situation anyway) but you've got no reason to push here. You're in first, last has riichi'd, second is dealer, you've got a bunch of tiles that are very very likely to be safe. This is the perfect time to fold. If you fold, last place has to tsumo haneman to hurt you, but if you push then you only have to deal into 2700 to get screwed over. Being over 30k in oorasu is far, far better than being under 30k. Always try to avoid a situation where you absolutely have to win in oorasu.

S4:
- You have a really nice hand but as nice as it is, you've gotta fold against that riichi. If you deal into mangan, it's all over, but if it goes to west round then you've got a great advantage since you're closest to 30k. Even if toimen tsumos this riichi or if someone else wins in west round, you'd still most likely finish second. Last place as 5d isn't all that terrible but it's still important to remember the basics of last place avoidance.

http://tenhou.net/0/?log=2015122212gm-0029-0000-5197a49a&tw=2

E1:
- Rather than 9s, you should've discarded 9m instead, since the other two are already discarded.
- Why pass up pon on the ton? You're passing up tenpai, plus you're locking yourself into a ton hell wait. That's not worth staying menzen for.
- You were keeping xia as a safe tile, but dropping it with that timing is bad. It's not like 8p is especially dangerous right now, but it will be later on.
- Folding is the right move since you've already lost your chance to win the hand.
- 8p now? 4s was safe against everyone. Shimocha seemed to be folding, so 5s and 7m were also safe.

E1-2:
- 6s is bad tile efficiency. Sanshoku doesn't look likely, so just drop the 2m.
- I dunno about riichi nomi here. You've got kind of a bad wait, you're in last, and it's still fairly early in the round, so maybe you could afford to delay a little so you can try and improve your wait.

E2:
- 1s is wider than 2s. You can pon 2s for tanyao tenpai, plus you keep the 36p ukeire.
- Discarding 1s now is fine but 5s shanpon is weaker than 2s.
- It's too early to discard 4p there. With a middle tile shanpon it probably won't hurt you too much, but everyone else still looks slow so you don't need to get defensive just yet.
- Why switch chun for pei? If you're going to shibori then you have to shibori the whole way. If not, then you should discard it immediately, because it's only going to get more dangerous as time goes on.
- Pei would probably be safer to keep than 6s.
- It'd be safer to discard 2s than 5s.
- That early 4p discard really backfired.

E3:
- No reason to discard 4s to keep 7p. You're making your iishanten narrower, you still need to wait for another effective draw, and there's no guarantee you'll end up with a better shape.

E4:
- 7p off the bat? 1m is way less useful, even if it's next to the dora. You're in first, the dealer's in last, your priority should be to rotate the winds.

S1:
- Jihai order.
- It's faster to just discard the 1m rather than breaking up the 12s penchan. You've got too many pairs as it is, keeping the 1m will lock you into a weak shanpon.
- Why break up the 7m kanchan to keep 4p? You're going back shanten.
- I don't think this is a good time to go chiitoitsu. 56p will be hard to pair, and dangerous to switch out.

S1-2:
- 1s may seem isolated but it's actually rather strong. You have 5s, so you only need to draw one tile (3s) to get a taatsu waiting on two tiles. Look out for 159 shapes in your haipai. They're much more useful than otakaze ton.
- 2s timing is bad. Either discard it right away or keep it until you've discarded your other junk tiles.
- Why keep the 2s now? 8p is far more useful, and 2s looked safe anyway.

S1-3:
- Defending against the riichi is right, but it'd be best to start with 4m rather than 9m. With a hand like this there's still a chance you can get tenpai safely.

S1-4:
- I don't think you need to break up your 9p pair. 2m is safe enough, and 5p is also probably safe. Giving up this hand means letting kamicha win his mangan.

S2:
- 4s tsumogiri? Given the choice between tanyao and dora, take tanyao. Otherwise you're stuck with sanshoku, which won't necessarily pan out.
- Better to give up on sanshoku when you draw the red 5p. Discard 8p instead. Ideally you'll draw 5m or 7m first, but even if you don't, it's better to have a 7m tenpai than a 7m iishanten.

S3:
- 8m was too early. Drawing either of 67m gives you a double sided taatsu, so discard the less useful chun instead.
- You've got a messy furiten wait, but it can still recover, so just suck it up and discard 7s. Even if you were going to avoid furiten, discarding 7m would let you keep the 5m ukeire.
- If you don't need the 8s kanchan anymore, then it's better to discard the 7s first. Discard the dangerous tiles first before they actually get dangerous.

S4:
- You need chun dora with either a direct hit or tsumo with haku to take second, so keeping the haku isn't unreasonable, but haku isn't worth more than 9p.
- Why discard 2s? You've got too many pairs, discard 7p instead since it's the least useful of them.
- 9m tsumogiri? A pair plus an adjacent tile is more useful than a pair. Are you going for toitoi from four pairs? That's not likely to work out.
- Discarding 5p means you lose the 36p ukeire. If you're going to chase toitoi then you should start with either the manzu or 4p so you can at least keep the 6p ukeire.
- Huh what are you doing, honitsu? That's too far away, and your previous manzu discards hurt too much.
- You're pushing dangerous pinzu against last place dealer now, just for chun nomi. That's not worth the risk. Chun nomi will only let you keep third, but you also keep third if someone else wins or if the dealer finishes in noten.
- This honitsu isn't happening, you shouldn't pass up the win.
- But then why take the win there? You gotta be decisive in mahjong. Once you make a decision, stick with it, because being ambivalent means putting your previous choices to waste, which only benefits the other players.

Overall, your tile efficiency could do with some refinement. Try and work on the way you balance hand speed with shape and value. You could also stand to improve your defense. Defense in mahjong isn't just about "not dealing in", it's about minimizing risk at all times. An effective way to do this is to predict the final shape of your hand and to discard everything else starting from the most dangerous tiles. This lets you keep safer tiles for later. It's tricky, but once you get the hang of it it'll be very useful. You need to play decisively to pull it off - make the right decision to begin with, then stick to it rather than flipflopping back and forth.