Move 16 swarminglocusts vs. mackas
I’ve been taught that if you want points you play wat I did on move 17. Why is the hane from the computer or descent better in this board position? Thank you.
but why you want points here?
other places may have more points
My whole right side had been spent on points. Why or how to use influence if not recovering points already planted?
I think generally it’s supposed to be a kind of suboptimal sequence. Like you can play that way if the corner is in danger of dieing, but it lets white get really thick on the outside.
The idea of the descent or hane is that White can try to live in the corner, make black a lot stronger on the outside, and in exchange black gets to then invade and attack/split the two and one white stones on the side.
They become weaker as the black wall gets stronger.
Playing around with some ai ideas
I had a game at the go club last week that looked like this
Black takes the corner from white, and then white uses the strength to counter attack and break the bottom area of black and also gain back a few points. 16 might’ve been a slight overplay.
Anyway, later black actually dies in the corner with a ko, although I think we both misread the tsuemgo 
I think white was generally ahead, since the top right corner.
There’s also another playable sequence sometimes from
but white also has the nice endgame move A later, and again sente like the other variation.
But a better explanation might be by pro players
Possibly related
2 Likes
This one can also be kind of useful
Before answering this question, do you understand why s18 in your game was a much bigger mistake? As that’s easier to understand and correct.
Also you can link to move numbers like this: https://online-go.com/game/68615208/27
2 Likes
Connecting at Q17 was better. Why it lost more territory in the corner. But some sdk player taught me it…
If it’s that they forgot the two stones were in atari, while trying to picture variations then I can kind of understand.
Like if you were thinking, “if white connects, do they have a follow up?” concluded they do
but then forgot the premise was that they first connect the two stones and quickly clicked, then I kind of get it.
Even still, it feels like you don’t have to fix every other mistake in your game before being able to ask a question about a more “advanced” mistake say.
, but it’s easier to grab the low hanging fruit.
No point trying to master flying knife joseki when you are still forgetting to connect after hane on the first line 
1 Like
Forgetting can happen in lots of situations. If you do it the odd time under time pressure, or in a complicated situation, it’s not the same as not understanding something.
Anyway, I don’t think you have to grab the low hanging fruit to intentionally avoid answering someone’s question.
Hey can I ask you about this joseki -
yes but I’m not going to answer you because of this other mistake I saw in your game. Come back to me again when you’ve fixed that.
1 Like
Do you mean capturing the white q17 stones by playing at p17?
He was able to ask the question, and I am not saying he shouldn’t. But my usual answer to that question will assume a basic level of go skill which the s18 mistake suggests he does not possess, and therefore if my answer is to be understandable I will need to make it simpler and explain more things rather than taking for granted he will understand implicitly that e.g. “this sequence is sente” is a good thing. Perhaps he had already identified the s18 mistake and realised why it is so bad on his own and therefore didn’t need to ask about it, but this one is harder and wanted help, but perhaps not. That’s why I asked, to gain information and adjust how my replies continue accordingly.
BTW, your AI suggested variation is advanced for me as a 4d, so for a 15k who plays s18 it’s basically useless. Teachers should adjust their teaching to the level of the student, which is one reason AIs have downsides as teachers.
2 Likes
A post was split to a new topic: Links to OGS from Forums not working correctly
So in the game black played at R16 on move 17, but the AI says that loses 4.5 points and recommends black O18 instead. Black doesn’t understand why the AI criticises R16, considering that R16 secures points. How can that be bad?
First, the AI is not taking into consideration that black is a 14k player. In its reviews, it judges every move as if the players are 12d superhuman players, like the AI itself.
I’m far weaker than 12d myself, but I know a thing or two that may explain the AI’s criticism here.
The issue with R16 is that it doesn’t take enough points. At this early stage of the game, there are stilll many moves available that are worth some 15 points (like a move in the lower left corner), but R16 is only worth about 10 points. The AI assumes that white will play a 15 point move next (as a 12d player would), and then black would lose some 5 points in that exchange. I’d say that’s the simple explanation for how the AI arrives at the judgement that R16 loses 4.5 points.
Going into more detail with the next part, I might go over the head of a 14k player, as it is more aimed at strong SDK to dan level. But I’m still going to give it a try:
Over the centuries, humans have developed ideas that help stronger players to evaluate this sort of situation. One concept that applies is “overconcentration”. I think the example on that senseis page is quite good. I’ll post the first and the last diagram of the sequence:
Black successfully secured some points 15 points on the upper side. But even though all black’s moves seem quite reasonable individually (which isn’t surprising, given that this was a pro game), it’s still considered a bad result for black.
The reason is that black overcommitted, playing too safely/defensively and spending too much effort to secure those 15 points.
Black ended up with 2 powerful walls radiating influence towards the area marked by red circles. But this area is tiny compared to the potential those walls would have in isolation, radiating their influence towards a larger empty region, while in this position those walls are really close together and their influence largely overlaps. There is clearly some redundancy (a.k.a. overconcentration) in black’s shape around the area marked by red circles.
Now compare that example with black’s shape in your game after move 18:
Can you see how removing black’s stone at Q14 wouldn’t change much?
If it were absent, would black need to play Q14 at this moment?
I hope you can see that the answer is “no”.
But that means that at least one of those marked black moves caused black’s shape to become inefficient/overconcentrated. Even though each of black’s moves might seem reasonable individually, black still ended up losing points creating this shape.
Stronger players (especially AI) will try to avoid this sort of result at all costs and resist, potentially leading to a scenario of mutual destruction (like playing at O18 on move 17).
Still, as a 14k player, I think you shouldn’t worry too much about overconcentration. Your games (including this one) probably contain much bigger and easier to avoid mistakes.
3 Likes
Yeah I mean I wouldn’t come up with it either. But I wanted to find a not bad variation that showed black getting a wall and white living.
There was other variations that needed white living with a comb shape or other tenuki and I thought that might also be too advanced.
I wanted a more visual picture of the trade than emphasise the sequence too much. I think once the sequence gets long enough, it’s going to be hard to remember anyway unless you know it. So maybe the details aren’t too important. It probably won’t play out exactly the way a book suggests or an ai suggests until certain levels also.
I suppose there’s simpler ideas too.