Life In 19x19

Hi all,

I've posted a review of a Chinese problem book set here; viewtopic.php?f=17&t=15712

I thought I would post one of the problems from the book here also for anyone who misses the review, to hear some stronger players thoughts on possible solutions.

Black to play, and attack white. From a section titled something like "Attacking in good shape"-

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------+
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . , . . . . . . |
$$ | . O . . . . . . . . . . . |
$$ | . X O . . O . . . . . . . |
$$ | . . X O . O X . O , . . . |
$$ | . . X . O X . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+[/go]

Thanks all!

I recognise this shape, so know the famous tesuji by sight rather than needing to figure it out myself:

Uberdude wrote:I recognise this shape, so know the famous tesuji by sight rather than needing to figure it out myself:

Nice! What is your impression of the level of difficulty of a problem like this? I didn't even know how to begin, but then this is one of the last in a series, and the indicated difficulty is above my level.

Another for you to look at. Hopefully one you haven't seen before!

Black to play

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------+
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . X . X , . . . . . . |
$$ | . X X O . . O . . . . . . |
$$ | . X O . . O . . . . . . . |
$$ | . O . O . O X . . , . . . |
$$ | . . . O X X . . X . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+[/go]

zac wrote:What is your impression of the level of difficulty of a problem like this?

Not sure, if you've never seen the correct move then it is pretty weird so a likely blind spot to miss. Also "attacking in good shape" seems an odd description which could mislead, I see it more as black is settling himself (defending) rather than attacking. I suppose a low sdk should probably be able to find it in 5 minutes?

P.S. This move could also get a reasonable result and is easier to find:

zac wrote: Another for you to look at. Hopefully one you haven't seen before!

No such luck, that's a standard joseki followup, indeed I played it in one of my British championship games! http://britgo.org/files/bchamp/2015/gam ... arelay.sgf (Andrew's response was not normal and was probably good for me).

This conversation indicates to me how important having a pattern dictionary is in go (I'm not calling it "pattern-matching" here because I think to some people that sounds like it is implying learning by rote with no skill taking place). I am currently doing a run of relatively simple problems and it is feels like there are really four very distinct groups:

• Problems I can solve instantly without thinking
• Problems where the overwhelmingly likely correct move is clear and I just have to verify that it works
• Problems where I have to start reading and rigorously searching the variation tree, but I am guided by some principles
• Problems where I really just have to brute-force the tree

I think that in general, one of the most important ways to improve is to take situations that will arise in actual games and push them earlier in that list. This is one reason that I think that (for kyu players at least) it's a better use of ten minutes to solve fifty easy problems than one brain-buster (although of course it is necessary to acquire the ability to read!). Maybe this is true for dan players too! Sometimes when I watch Guo Juan lectures about typical patterns she will say "I see low dan players get this wrong all the time".

One of the things I really like about Thomas Redecker's books on tsumego is that they provide a really detailed thought process to help move problems up the list (the workbooks for moving things into categories 1 and 2, the tsumego strategy books for moving things into category 3).

(By the way, thanks for the hide tags! I haven't tried to solve these problems myself yet.)

Uberdude wrote: No such luck, that's a standard joseki followup, indeed I played it in one of my British championship games! http://britgo.org/files/bchamp/2015/gam ... arelay.sgf (Andrew's response was not normal and was probably good for me).

Nice to know that it isn't just an artificial position. The overall quality of problems seems quite good to me. I think the level indicated is reasonable- these are meant to be 1D problems. Judging from the 5k books, this should mean that a 1D ought to be able to see the first move very quickly and only need to read it out to confirm. Seems like the books will be very good for covering the basic techniques and highlighting any such blind spots.

Having looked at the problems now:

Problem 1: Problem 2:

This is one reason that I think that (for kyu players at least) it's a better use of ten minutes to solve fifty easy problems than one brain-buster (although of course it is necessary to acquire the ability to read!). Maybe this is true for dan players too! Sometimes when I watch Guo Juan lectures about typical patterns she will say "I see low dan players get this wrong all the time".

Totally agree. I like to push my reading depth/breadth sometimes, but don't feel like it helps a lot in games. Solving many easy problems many times does seem to help. I think there is a distinction between "able to get it right" and "never get it wrong". In another lecture Guo talks about how many western players stall around the 1D level, because they lack some strong fundamental skills.

My feeling is that learning through experience is the best way- so lots of games and lots of problems. The harder problems introduce new ideas, and the easier problems perfect them. I think later I can worry about more theoretical concepts like opening theory, direction of play etc, beyond the most basic guiding principles.

I remember reading somewhere sometime like "calculation is most important. Some kyu player thinks their opening is weak, but I don't feel that. As soon as the stones touch, they fall apart. Opening theory doesn't work for them." I think some general principles are good to have, but feel like a lot of players have it a bit backwards. I can't tell you how many times I've seen someone who is just in to the SDK range ask how to get stronger in fights, and usually they want a book that will teach them the theory. But if you ask how often they solve problems, it's not much. Intellectualising the problem is interesting, but like getting good at almost anything, it just takes lots of hard work and repetition. Theory can give an illusion of understanding, but it is being able to put it in to practice that is really important.

zac wrote: Nice to know that it isn't just an artificial position.

Indeed not, it's a common double approach joseki (if slightly old-fashioned now with the AI preference for attach instead of push). It occurred (with slightly different black defence on lower side) in the final of the recent European Grandslam tournament between Ilya Shikshin and Pavol Lisy. A followup tesuji problem is why black shouldn't play the atari on the outside, as it means white then has a nice counter tesuji inside, see Catalin Taranu's review with Ilya from https://youtu.be/dfohoEZEW2k?t=12m58s. Doing these joseki follow-up problems is what it means to properly study and learn from joseki (in the joseki white usually adds a move to prevent black's tesuji, so to understand why that defence is (sort of) necessary you need to know the threat).

Uberdude wrote:

zac wrote:What is your impression of the level of difficulty of a problem like this?

Not sure, if you've never seen the correct move then it is pretty weird so a likely blind spot to miss. Also "attacking in good shape" seems an odd description which could mislead, I see it more as black is settling himself (defending) rather than attacking. I suppose a low sdk should probably be able to find it in 5 minutes?

By low SDK you mean like 2-kyu?

And when I saw this problem, I didn't think, "How am I going to attack White with good shape?" I thought, "Oh, mama!"

P.S. This move could also get a reasonable result and is easier to find:

$$B
$$ +---------------------------+
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . , . . . . . . |
$$ | . O . . . . . . . . . . . |
$$ | . X O . . O . . . . . . . |
$$ | . . X O . O X . O , . . . |
$$ | . . X . O X . . . . . . . |
$$ | . . . . 1 . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------+
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . , . . . . . . |
$$ | . O . . . . . . . . . . . |
$$ | . X O . . O . . . . . . . |
$$ | . . X O . O X . O , . . . |
$$ | . . X . O X . . . . . . . |
$$ | . . . . 1 . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+[/go]

Uberdude wrote:Doing these joseki follow-up problems is what it means to properly study and learn from joseki (in the joseki white usually adds a move to prevent black's tesuji, so to understand why that defence is (sort of) necessary you need to know the threat).

Hear, hear!

The first joseki book I bought was a small pamphlet by the Nihon Kiin. Both Sakata's Killer of Go series and Takagawa's Go Reader set had a joseki volume, but I did not put much study into either one. The next joseki book that I bought could be called All About Joseki Mistakes, and I found it very valuable. The next one was How To Play After Joseki. At the time I thought it was the most valuable joseki book I had. But thinking about it now, I don't think I could have gotten as much from it if I had not already studied All About Joseki Mistakes. All of these books, except perhaps the Nihon Kiin pamphlet, were at SDK level.

dfan wrote:This conversation indicates to me how important having a pattern dictionary is in go (I'm not calling it "pattern-matching" here because I think to some people that sounds like it is implying learning by rote with no skill taking place).

I agree. You want to build up a "vocabulary" of at least 50,000 patterns.

I am currently doing a run of relatively simple problems and it is feels like there are really four very distinct groups:

• Problems I can solve instantly without thinking
• Problems where the overwhelmingly likely correct move is clear and I just have to verify that it works
• Problems where I have to start reading and rigorously searching the variation tree, but I am guided by some principles
• Problems where I really just have to brute-force the tree

I think that in general, one of the most important ways to improve is to take situations that will arise in actual games and push them earlier in that list.

OK. OC, that tends to happen naturally.

This is one reason that I think that (for kyu players at least) it's a better use of ten minutes to solve fifty easy problems than one brain-buster (although of course it is necessary to acquire the ability to read!).

I fall somewhere in between. First, I am not at all convinced that problems provide the best way to build up one's go vocabulary, although that can and does happen. Second, and more importantly, I think that easy problems do not provide enough of a workout. Sure, do a few over breakfast, but for a 10 minute workout at the SDK level, I think that working on 5-10 problems that are difficult enough so that you can solve only half of them in that time is better. In a real game there will be positions that you should spend a minute or two on, and I think that that's the right time to spend on problems, as well.

Edit: OC, you need to overlearn the problems that you missed.

Maybe this is true for dan players too! Sometimes when I watch Guo Juan lectures about typical patterns she will say "I see low dan players get this wrong all the time".

Low dans need to build their vocabularies, too.

One of the things I really like about Thomas Redecker's books on tsumego is that they provide a really detailed thought process to help move problems up the list (the workbooks for moving things into categories 1 and 2, the tsumego strategy books for moving things into category 3).

Redecker's books are a valuable resource.

Bill Spight wrote:
I agree. You want to build up a "vocabulary" of at least 50,000 patterns.

I fall somewhere in between. First, I am not at all convinced that problems provide the best way to build up one's go vocabulary, although that can and does happen. Second, and more importantly, I think that easy problems do not provide enough of a workout. Sure, do a few over breakfast, but for a 10 minute workout at the SDK level, I think that working on 5-10 problems that are difficult enough so that you can solve only half of them in that time is better. In a real game there will be positions that you should spend a minute or two on, and I think that that's the right time to spend on problems, as well.

Redecker's books are a valuable resource.

Very interesting to hear your thoughts Bill.
I had another look at the section of the book that the first problem is from. Looking at the examples given and some of the other problems, it looks to be problems looking to take advantage of white's bad shape. The "attack" might mean "attack white's shape". I'll ask someone next week who might give me a better idea.

A much easier example;

Click Here To Show Diagram Code

[go]$$B
$$ +---------------------------+
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . , . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . X . . . . . . . |
$$ | . . . X . X O . . , . . . |
$$ | . . . X O O . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+[/go]

If problems are not the best way to build a persons go vocabulary, what is, in your opinion?

It's been my experience when playing against western players that it is their reading strength, and basic tactics during contact fights that lets them down- I see a lot of broken shapes, allowing hane at the head of two stones, not cutting when it is possible etc. And then against Asian players, I see much of the opposite- good reading and tactical skills, but mistakes like playing too close to thickness, invading deeply when it's not needed or jealousy of territory, poor choices in opening (preferring to start an immediate fight rather than take a big point).

Your comments r.e. easy v. challenging problems is something for me to think about. Perhaps I'm biased towards easy problems because I just enjoy them more I like to feel like I have accomplished a lot, rather than failing half the time. Actually, thinking about it now, if failing half the time scares me maybe I should choose a different game to play! Actually I feel that I maybe missed some elementary education early on when I started playing. I lost a game a few weeks ago by responding to a hane on my L+2 group incorrectly. Actually reading it out shouldn't be a challenge to me, and if it was presented as a problem I would likely get it right. But alas I did not read accurately during the game. I think had I solved many problems involving that shape, I would never get it wrong during a game. Immediate recognition of the shape would have given me the clue I needed. I miss a lot of opportunities to cut, capture, connect etc., that would be more apparent if I could just "see" that there is a better move there.

For me, solving problems seems to have two purposes; to improve brute reading strength- depth/breadth, secondly to improve (call it what you will) intuition- the ability to immediate see a shape and know that there is a good move to play. Harder problems seems to improve the former, as well as introduce new moves for the latter. Easy problems seem to cement the latter, ensuring that during a game those moves become immediately apparent. I'm guessing that your suggestion of over learning helps with the second; to practice the harder problems (hard because you cannot see the possible move) until they become easy (you can see the first move).

Bill wrote:You want to build up a "vocabulary" of at least 50,000 patterns.

EdLee wrote:

Maybe one option:

$$B
$$ | . . . . . . , . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . X . . . . . . . |
$$ | . . . X . X O . 1 , . . . |
$$ | . . . X O O . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . , . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . X . . . . . . . |
$$ | . . . X . X O . 1 , . . . |
$$ | . . . X O O . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ +---------------------------+[/go]

I came across this for the first time a couple of days ago, in Jump Level Up 1, otherwise I would have gotten it wrong. I think it's the kind of problem that lots of SDK would be pretty lost on.