Knotwilg's practice

jlt · **#181**

So what would have been a better move for 107, and how to use the aji of the cut stones?

Knotwilg · **#182**

Click Here To Show Diagram Code: [go]$$ Position at move 106 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . O X . . . | $$ | . . . . . X . . X . O . . . O X . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . X . . . . . . O . . O . . . . . | $$ | . . X O . O . . . . . . . . . . X . . | $$ | . . X O . O X . . . . . . . . . . . . | $$ | . . X O X O . . . . . . . . . . . . . | $$ | . . X X O . . . . . . . . . . O O . . | $$ | . O . X O . X . . , . . . B . X X O . | $$ | . . . O . . . . . . . O . . . . . O . | $$ | . . . O . O . X . . . . O . . X X O . | $$ | . . 1 . . . W . X O O O X . . X O . . | $$ | . . . . . . . 2 X O X X X . O X O . . | $$ | . . . X . X X X O X X . . O . X X O . | $$ | . . X , . . O O O X O X X X X X O O . | $$ | . . X O . . . . . O O O O O O O X X . | $$ | . . O . O . . . . . . . . . . X . X . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

The situation is not easy. After BB jumped, WW peeped and then I (Black) played B1.
Katago thinks this is a 14% mistake. Instead, KataGo thinks about 2 other moves.
Before discussing them, let's look at the situation

Click Here To Show Diagram Code: [go]$$ Position at move 106 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . O X . . . | $$ | . . . . . X . . X . O . . . O X . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . X . . . . . . O . . O . . . . . | $$ | . . X O . O . . . . . . . . . . X . . | $$ | . . X O . O X . . . . . . . . . . . . | $$ | . . X O X O . . . . . . . . . . . . . | $$ | . . X X O . . . . . . . . . . O O . . | $$ | . O . X O . X . . , . . . X . X X O . | $$ | . . . O . . . . c . . O . . . . . O . | $$ | . . . O . O . X . . . . O . . X X O . | $$ | . . . . . . O . X O O O X . . X O . . | $$ | . . b . . . . a X O X X X . O X O . . | $$ | . . . X . X X X O X X . . O . X X O . | $$ | . . X , . . O O O X O X X X X X O O . | $$ | . . X O . . . . . O O O O O O O X X . | $$ | . . O . O . . . . . . . . . . X . X . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

The big question is if Black has the time (eyespace) to connect at ''a''. Not only can White cut off a couple of stones and build a centre, also White's central stones are currently disconnected and White ''a'' would help a lot.

A blunt connection at ''a'' would make Black very heavy. White can poke at Black's shape with moves like ''b'' or ''c'' (AI favorite).

Click Here To Show Diagram Code: [go]$$ Position at move 106 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . O X . . . | $$ | . . . . . X . . X . O . . . O X . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . X . . . . . . O . . O . . . . . | $$ | . . X O . O . . . . . . . . . . X . . | $$ | . . X O . O X . . . . . . . . . . . . | $$ | . . X O X O . . . . . . . . . . . . . | $$ | . . X X O . . . . . . . . . . O O . . | $$ | . O . X O . X . . , . . . X . X X O . | $$ | . . . O . . . . . . . O . . . . . O . | $$ | . 3 . O . O 1 X . . . . O . . X X O . | $$ | . . 4 . . 2 O . X O O O X . . X O . . | $$ | . . . . . . . . X O X X X . O X O . . | $$ | . . . X . X X X O X X . . O . X X O . | $$ | . . X 5 . . O O O X O X X X X X O O . | $$ | . . X O a . . . . O O O O O O O X X . | $$ | . . O . O . . . . . . . . . . X . X . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

Katago suggests a better shape move to guard the connection: :b1:

. Then instead of connecting, it suggests another shape point: :b3:

. After

, there's an exchange with White taking the center and Black playing ''a'' and capturing. Very high level stuff.

So what's this about? What if White cuts?

Click Here To Show Diagram Code: [go]$$ Position at move 106 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . O X . . . | $$ | . . . . . X . . X . O . . . O X . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . X . . . . . . O . . O . . . . . | $$ | . . X O . O . . . . . . . . . . X . . | $$ | . . X O . O X . . . . . . . . . . . . | $$ | . . X O X O . . . . . . . . . . . . . | $$ | . . X X O . . 7 . . . . . . . O O . . | $$ | . O . X O . X . . , . . . X . X X O . | $$ | . . . O . . . . . . . O . . . . . O . | $$ | . 3 . O . O 1 X . . . . O . . X X O . | $$ | . a 8 . . 2 O 5 X O O O X . . X O . . | $$ | . . . . . . 6 4 X O X X X . O X O . . | $$ | . . . X . X X X O X X . . O . X X O . | $$ | . . X b . . O O O X O X X X X X O O . | $$ | . . X O . . . . . O O O O O O O X X . | $$ | . . O . O . . . . . . . . . . X . X . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

With Black having better shape at :b1:

and White worse at :b3:

the mutual cut becomes a problem for both. White's left is not as thick as in other variations and his center is in dire straits.

When :w8:

presses, Black will be able to connect underneath, provided ''b'' is sente. That's where this move comes from.

Click Here To Show Diagram Code: [go]$$ Position at move 106 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . O X . . . | $$ | . . . . . X . . X . O . . . O X . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . X . . . . . . O . . O . . . . . | $$ | . . X O . O . . . . . . . . . . X . . | $$ | . . X O . O X . . . . . . . . . . . . | $$ | . . X O X O . . . . 8 7 . . . . . . . | $$ | . . X X O . . . . . 6 5 . . . O O . . | $$ | . O . X O . X . . , 4 3 . B . X X O . | $$ | . . . O . . . . . . . O . . . . . O . | $$ | . . . O . O . X . . . . O . . X X O . | $$ | . . 1 . . . W . X O O O X . . X O . . | $$ | . . . . . . . 2 X O X X X . O X O . . | $$ | . . . X . X X X O X X . . O . X X O . | $$ | . . X , . . O O O X O X X X X X O O . | $$ | . . X O . . . . . O O O O O O O X X . | $$ | . . O . O . . . . . . . . . . X . X . | $$ | . . . . . . . . . . . . . . . . X . . | $$ ---------------------------------------[/go]

So, when Black plays :b3:

and subsequent, all that aji goes to waste and White gets a big center.

Knotwilg · **#183**

Now that we have AI to review, I'd really like to record my thought process. That's the next step. The best I can do is review my game, add what I think I thought, then review. Here's a recent win.

Knotwilg · **#184**

I played 10 games on OGS, following "dwyrin's heuristics" aka "basics":

See https://senseis.xmp.net/?Dwyrin

6 - 4 but really 5 - 5 as one was a lost game won +T

Why did I win?

1) lucky, opponent blundered in a fight
2) good splitting attack, finishing off one of the groups, leaving the other weak
3) proper usage of influence to attack, killing a group
4) winning on thickness and proper direction of play
5) convert central influence into territory

Why did I lose?

1) slack play, failure to kill an invading group
2) acquired central influence but failed to kill an invading group
3) wrong play in middle game joseki, downhill from there (opponent played strong)
4) wrong play in opening joseki downhill from there (same)
5) failure to attack for profit

What does this tell me?

1) I remain an influence oriented player
2) I like to fight and kill but that's not always the right idea or shouldn't be the win condition
3) still get better at killing: play sharper moves and read out more variations faster
4) I should learn more joseki
5) I'm performing at my level of comprehension; I didn't make any mistakes which I could easily identify without AI review

Knotwilg · **#185**

Reviewing my latest 8 games with KataGo, zooming in on my mistakes and then categorizing them, this is the result:

Fighting/middle game tactics
• Too confident resisting
• Wrong extension in capturing race (reading)
• Missing a chance to simplify, caught up in the fight
• Don’t bail out of a capturing race if you have many more liberties
• Defend higher up the chain

Shape/technique
• Playing a knight’s move for defence
• Disconnect a big group, while there was a better move that threatened to disconnect a part AND destroy the shape
• Don’t give opportunity to profit in sente
• Wrong shape in a joseki follow up (keima i/o kosumi)
• Wrong shape move

Endgame/sente
• Answering locally while the biggest move was elsewhere (deny profit in sente)
• Wrong local move losing sente
• Small, not severe move
• Just defending territory is small

Strategic choices
• Too territorial choice while splitting was better
• Missing a chance to cut off a major group, focusing on central influence instead
• Don’t reinforce the high 2 space enclosure for mere sabaki
• Attack too close to strength, allow to resist

Invasions/reductions
• Play what looks like a reduction (a shoulder hit) but it’s actually too deep an invasion
• Better Invade where the weakest group is
• Don’t invade just for scooping out territory; instead attack the weakest group

Ko
• Should answer a ko threat by playing another ko locally
• Wrongly evaluate winning the ko vs answering ko threat

Life & Death
• Wrongly evaluate a group to be killable, losing sente

Knotwilg · **#186**

I'm using a fantastic app, TsumeGo pro, one of the few for which I pay because it's so awesome. After this advertisement, let me discuss one of the problems I solved today.

First some context: there's a daily output of 6 problems, 2 Easy, 2 Medium, 2 Hard. That's spot on for my level: Easy I can solve at a glance, Medium I can always solve with some focus but occasionally make mistakes, Hard I can solve, say 50% of the time.

Here's one of today's hard problems:

Click Here To Show Diagram Code: [go]$$B Black to kill the whole group $$ -------------------+ $$ . . . . . . c a . | $$ . . . X X X O . O | $$ . . . X O O . b . | $$ . . X X O X O O . | $$ . . X . O X X O d | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

More context: I always need to apply myself to solve such a problem thoroughly, not just by clicking.
Solving-by-clicking usually starts at one of the vital points. Here I see A and B.
The systematic approach is to reduce the eyespace first with the hanes of C or D.

Click Here To Show Diagram Code: [go]$$B reduce eyespace 1 $$ -------------------+ $$ . . . . . . 1 2 . | $$ . . . X X X O . O | $$ . . . X O O 4 3 . | $$ . . X X O X O O . | $$ . . X . O X X O d | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

We can quickly see how :b1:

fails. I would normally not explore this variation: it's useful to discover the vital point of :w2:

which I already have discovered.

Click Here To Show Diagram Code: [go]$$B reduce eyespace 2 $$ -------------------+ $$ . . . . . . 4 3 . | $$ . . . X X X O 7 O | $$ . . . X O O 6 5 . | $$ . . X X O X O O 2 | $$ . . X . O X X O 1 | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

This line already works. :b3:

is played at the vital point because reducing the eye space on that side didn't work. Next we play the other vital point at :b5:

. If

at

then we falsify the eye at :w6:

. With

we create a known "nakade" using the "special properties of the corner". If :w4:

elsewhere, then Black connects his stone and there can't be two eyes. So how about :w2:

?

Click Here To Show Diagram Code: [go]$$B reduce eyespace 2 $$ -------------------+ $$ . . . . . . . 2 . | $$ . . . X X X O . O | $$ . . . X O O 4 3 6 | $$ . . X X O X O O 5 | $$ . . X . O X X O 1 | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

If

on the vital point, then this line fails for Black

Interestingly though ...

Click Here To Show Diagram Code: [go]$$B reduce eyespace 1 $$ -------------------+ $$ . . . . . . 2 3 . | $$ . . . X X X O 7 O | $$ . . . X O O 6 5 . | $$ . . X X O X O O 4 | $$ . . X . O X X O 1 | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

This

reactivates the killing line. We can see that :b3:

really is the vital point. So ...

Click Here To Show Diagram Code: [go]$$B main line $$ -------------------+ $$ . . . . . . 2 1 . | $$ . . . X X X O 7 O | $$ . . . X O O 6 5 . | $$ . . X X O X O O 4 | $$ . . X . O X X O 3 | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

... this becomes the main line. :b1:

forces

and now

executes the killing shape.

Click Here To Show Diagram Code: [go]$$B variation $$ -------------------+ $$ . . . . . . 2 1 . | $$ . . . X X X O 4 O | $$ . . . X O O 5 b . | $$ . . X X O X O O a | $$ . . X . O X X O 3 | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

doesn't help: after :b5:

A and B are miai.

Click Here To Show Diagram Code: [go]$$B failure $$ -------------------+ $$ . . . . . . 2 1 . | $$ . . . X X X O 4 O | $$ . . . X O O 5 3 6 | $$ . . X X O X O O . | $$ . . X . O X X O a | $$ . . . X O O X X . | $$ . . . . X X X . . | $$ . . . . . . . . . | $$ . . . . . . . . . |[/go]

Incidentally, this :b3:

doesn't work. After :w6:

White can live at :b5:

or A.

This problem has a depth of 9 and a width of 8, for me at least. That's quite a hard problem but the systematic approach of it reduces the complexity of the variation tree.

Knotwilg · **#187**

For reference, this is the second hard problem today. It's one that I can almost solve at the spot. There's a clear vital point and I have often enough seen problems where the threat to create an eye by capturing three stones leads to shortage of liberties in the corner. In other words, this is almost an intuitive one for which I won't apply a systematic approach.

Click Here To Show Diagram Code: [go]$$B Black to live $$-----------------------+ $$ . . . . . . . . . . . | $$ . . . . . . . . . O . | $$ . . . . O . O O O . . | $$ . . . . . O X X . X . | $$ . . . . . O X . . . . | $$ . . . . . O X X X X . | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

Now if you do, it goes like this:

Click Here To Show Diagram Code: [go]$$B Increase eyespace 1 $$-----------------------+ $$ . . . . . . . . . . . | $$ . . . . . . . . . O . | $$ . . . . O . O O O 2 . | $$ . . . . . O X X 1 X . | $$ . . . . . O X . a . b | $$ . . . . . O X X X X . | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

Failure: A and B miai

Click Here To Show Diagram Code: [go]$$B Increase eyespace 2 $$-----------------------+ $$ . . . . . . . . . . . | $$ . . . . . . . . . O . | $$ . . . . O . O O O 1 2 | $$ . . . . . O X X 3 X . | $$ . . . . . O X . . . . | $$ . . . . . O X X X X 4 | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

Fail by reduce

Click Here To Show Diagram Code: [go]$$B Increase eyespace 3 $$-----------------------+ $$ . . . . . . . . . . . | $$ . . . . . . . . . O . | $$ . . . . O . O O O . . | $$ . . . . . O X X 3 X . | $$ . . . . . O X . 2 . . | $$ . . . . . O X X X X 1 | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

Only in this 3rd attempt the vital point must be played to invalidate it. :w3:

dies in gote ...

Click Here To Show Diagram Code: [go]$$B Increase eyespace 2b $$-----------------------+ $$ . . . . . . . . . . 8 | $$ . . . . . . . . a O 7 | $$ . . . . O . O O O 1 2 | $$ . . . . . O X X . X 6 | $$ . . . . . O X . 3 9 4 | $$ . . . . . O X X X X 5 | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

In the previous attempt, Black can play the vital point and live. See below main line.

Click Here To Show Diagram Code: [go]$$B Increase eyespace 2c $$-----------------------+ $$ . . . . . . . . . . . | $$ . . . . . . . . . O . | $$ . . . . O . O O O 1 4 | $$ . . . . . O X X 3 X . | $$ . . . . . O X . 2 . . | $$ . . . . . O X X X X . | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

So White must kick in :w2 here.

Click Here To Show Diagram Code: [go]$$B Main line $$-----------------------+ $$ . . . . . . . . . . 8 | $$ . . . . . . . . a O 7 | $$ . . . . O . O O O 5 6 | $$ . . . . . O X X . X 4 | $$ . . . . . O X . 1 9 2 | $$ . . . . . O X X X X 3 | $$ . . . . . . O O O O . | $$ . . . . . . . . . . . | $$ . . . . . . . . . . . |[/go]

Which eventually leads back to 2c in a different order. :b1:

takes the vital point right away; :w2:

takes the next vital point
:b7:

is the tesuji, a sacrifice that leads to a shortage of liberties for White. When she connects at :b7:

against Black's atari at :b9:

, Black A captures all 6 stones. If White doesn't connect, then Black captures three stones to make an eye.

Formally, this problem would have width 6 and depth 11. That probably meets the criteria for "hard". But I can see it instantly, thanks to doing many problems. Now if only real game problems of width 6 and depth 11 were intuitive to me ...

Knotwilg · **#188**

Not sure what happened but I'm on a 12 game winning streak on OGS and comfortably on 1d.

I recently started learning with Baduk doctor but that sounds too soon to have any effect.

Knotwilg · **#189**

Ranking up to 1d OGS
KGS account has gone into decay, currently 1k.

Goal: rank up to 2d OGS, using KGS as a practice account
How?

Focus on reading. Play as sharply as possible on KGS.
Review the game without AI to record what I read.
Review with AI to evaluate reading

Performance games on OGS.

Knotwilg · **#190**

I plqyed 10 games on the first 10 days of the year and analyzed them with KataGo to discover patterns in my biggest mistakes. In each game there were at least 3 moves which lost more than 10% chance of winning, once as much as 60% and then there was the blunder that swapped almost 200%. I could also have ranked them by points lost but went for this metric. This means that some big mistakes remain under the radar as they don't affect the metric towards the end of the game while smaller errors in absolute terms come to surface because they affect an otherwise tight game.

Here are those 30 mistakes categorized.

1) poor choices in cutting and connecting, either doing so when low value or not doing so when high value - 6

The biggest lesson that I can draw is probably that my intuition for valuable cuts or connections is off. On 3 occasions I lost points because I went for a connection of low value or a cut that got me into trouble. On 3 other occasions not cutting or connecting turned out to be a negative turning point. This is not easy to learn from because I can't redress the balance in any direction. I need to look out for cuts and connections and evaluate if they're important, pointless or even negative.

2) lack of fighting spirit, not finding the (complicated) sharper move - 6

A big category consists of moves that I didn't consider at all and even after seeing them with AI I have a hard time understanding them. They are often about large scale evaluations where I could have sacrificed a group which would have taken the opponent a lot of time to capture, giving me the potential on the other side. Large scale, long term sacrifices.

3) Not identifying a weak group and failing to defend or attack it - 4

This is the 2nd big lesson: I underestimate the value of a play that defends or attacks a weak group and switch to big points too early.

4) slack move / wrong haengma - 4

On 4 occasions I make a slack move while the good move is easily available.

5) bad endgame - 4

Endgame was an issue in 2 games, one of which involved again a sacrifice that was not easy to see.

6) timing of a peep - 2

In one game there was a peep which I didn't play when I should have and when I played it I shouldn't. Not easy to understand why the timing was off.

7) ko - 2

Only one game involved a ko which I messed up twice. It was a 2 step ko which I first allowed to become a direct one and then I responded to a threat while I could have resisted.

The remaining 3 are single occurrences.

8) fail to surround in the opening - 1
9) missed opportunity to invade - 1
10) reading blunder - 1

The majority of these big mistakes occur between the 40th and 140th move. Only two mistakes can be seen in the opening. In the late endgame I blundered only once. That is quite comforting. Of course the middle game and early endgame are the expected arena for big mistakes at my level.

So, cut & connect properly, weak groups and endgame techmiques will be the focus areas for study and practice games

dfan · **#191**

I've learned a lot from your thoughts on improvement so I'm reluctant to offer any of my own, but here are a couple anyway - take them as seriously as is appropriate:

Knotwilg wrote:

1) poor choices in cutting and connecting, either doing so when low value or not doing so when high value - 6

The biggest lesson that I can draw is probably that my intuition for valuable cuts or connections is off. On 3 occasions I lost points because I went for a connection of low value or a cut that got me into trouble. On 3 other occasions not cutting or connecting turned out to be a negative turning point. This is not easy to learn from because I can't redress the balance in any direction.

Do you understand in retrospect why these moves were mistakes? If so, this is the sort of thing that flashcards (real or virtual) can be good for. I have about 200 cards with mistakes from my games: nothing involving a lot of a complicated reading, just "I should know better than this" sort of moves.

Quote:

2) lack of fighting spirit, not finding the (complicated) sharper move - 6

A big category consists of moves that I didn't consider at all and even after seeing them with AI I have a hard time understanding them.

If you have trouble understanding why the "correct" move is best then I think not playing it can't really be categorized as a mistake for your level, and it's better not to worry about it until it becomes more understandable. (Still of course it is useful to know what fraction of your "bad" moves fall into this category. And even if you don't understand the best moves, it's valuable to at least turn them into moves you are capable of considering.)

Knotwilg · **#192**

dfan wrote:

I've learned a lot from your thoughts on improvement so I'm reluctant to offer any of my own

Much appreciated!

Quote:

Do you understand in retrospect why these moves were mistakes?

I'll add all 6 examples in later so that I can use the wisdom of this crowd. At least two were of the nature "connecting without value / preventing a connection without value".

Quote:

If so, this is the sort of thing that flashcards (real or virtual) can be good for. I have about 200 cards with mistakes from my games: nothing involving a lot of a complicated reading, just "I should know better than this" sort of moves.

That wouldn't work for me, especially not 200 things to keep in mind. I liked Tamsi's "compass" which resembles an idea I picked up from a top referee in our country: in each performance, focus on one single thing to improve. Then after a while these become part of your regular performance.

Quote:

If you have trouble understanding why the "correct" move is best then I think not playing it can't really be categorized as a mistake for your level, and it's better not to worry about it until it becomes more understandable. (Still of course it is useful to know what fraction of your "bad" moves fall into this category. And even if you don't understand the best moves, it's valuable to at least turn them into moves you are capable of considering.)

Agree: they were indeed not mistakes at my level and rather a source of inspiration for the future. On the other hand, if I still want to break out of the low dan level where I have been stuck for the past 20 years, probably something more fundamental has to change than fixing mistakes at my current level. So, while I'm not going to focus my practice on these, I don't want to ignore their existence.

Thanks a lot for the feedback!

jlt · **#193**

You categories look like everyone's mistakes, except that your moves are less obviously bad than kyu mistakes.

Quote:

Not identifying a weak group and failing to defend or attack it

Knowing if a group needs a defensive move or if it can endure an attack is a difficult question in general and requires reading skills - life and death, tesuji and endgame.

John Fairbairn · **#194**

Quote:

If I still want to break out of the low dan level where I have been stuck for the past 20 years, probably something more fundamental has to change than fixing mistakes at my current level.

Quote:

Knowing if a group needs a defensive move or if it can endure an attack is a difficult question in general and requires reading skills is a difficult question in general and requires reading skills

Although I personally do not have the qualifications at playing level to back up what I am about to say, I think I've seen enough in the go literature to be able to comment.

I agree with the first quote. I think the second quote is a perfect example of where a fundamental change in thinking is needed.

Let us do a little explication de texte. "Knowing if a group needs a defensive move": I think the writer means us to infer that such a group would be a "weak" group. Similarly, "if it can endure an attack" seems to expect us to infer that such a group would be maybe not strong but "not weak." I would say that view is typical, in varying degrees of course, of almost all amateurs up to about 2-dan. Beyond 2-dan, amateurs seem to have a fuzzy idea that this notion is not quite right, but they can't really put their fingers on why.

To put this in stark terms, and allowing myself a smidgeon of hyperbole to make a point, if you imagine ten groups on the board early in a typical game and ask amateurs which ones are weak, most would say one or two. A higher dan would say 3 or 4 and (as an expression of his fuzzy suspicions) "maybe a couple more". A good professional would say 9. A bot might say "10 at the moment, 12 in a few moves time."

Precise numbers to one side, why does this matter? Well, obviously if you think a group is alive and the opponent can kill it, that's a problem. But it's not too common a problem in practice - rare, even, at dan level. Back to explication, "it can endure an attack" is another way of saying "can be bullied". Ijime (bullying) is a word you would soon become familiar with if you read Japanese pro commentaries. I very rarely hear it from western amateurs. I think that's partly because the term gets translated in many different ways and so ijime has not really established itself as an identifiable concept in English. But imagine if you allow (as amateurs do) four of your typical five groups to be bullied, and they each lose about 5 or 6 points on average. By not allowing yourself to be bullied you could improve very significantly - at a stroke even.

This is especially obvious at DDK level, but it occurs even at dan level. And in fact it occurs occasionally at pro level. It's often not easy to spot a group is weak enough to succumb to bullying. But the important thing is to learn to look. Pros look. Amateurs don't. After all, saying "it can endure" is just a way of not looking - sweeping problems under the carpet.

Now, deciding why a group may be weak is hard. But you can get round that problem quite simply: just assume all your groups are weak. Imitate the pros.

You then have a simpler task. Find ways of defending your weak groups (i.e. all your groups) that are not too inefficient if it turns out you were wrong about them being weak. (But it won't turn out that way very often!)

There are also pro strategies for this. Playing thickly is perhaps the main one. But the value of thickness is not necessarily the one amateurs usually attach to it. The usual ideas are to make a wall so you can attack the opponent, or to build a moyo. Not bad ideas in themselves, though they do have nuances of being a greedy warmonger and imperialist. If you want to improve, I suggest trying the statesman-like or diplomatic approach instead - playing thick moves with the main and explicit idea of doing so as prophylactic defence. Or, better, look at a pro game and when a thick move is played, assume it was primarily for defensive purposes. Then see how that defensive posture pays off later in the game (e.g. zero being bullied). I would suggest this will be evident more often than not.

Another strategy is to play honte type moves. This is sometimes related to playing thickly, but in fact there are quite a few ways of securing the necessary prophylaxis. Making a base is an obvious one but on the basis of counting up the various ways old Chinese commentators expressed the same idea in its various guises, I'd suggest there are at least 10 ways - all easily achievable by amateurs. If I can spot them, you can.

Yet another and subtle way of defending weak groups prophylactically is one I mentioned in another thread recently. That is to play a move that lends a helping hand to one's other groups. The Chinese term 照应 for this seems to have been introduced by Wu Ruizheng in the late 17th century. You could translate it as "call and response" but I prefer to view it as setting up a lighthouse for ships that may (or may not) fall into distress. I like this also because I can imagine the opponent setting himself up as a wrecking crew, by waving a false lantern during a storm so as to lure the poor sailors onto the rocks - I happen to think another important step-change for amateurs to make is to learn to remember at all times that go is a two-player game. There is an opponent.

This lighthouse building concept is of the type that can sound too obvious to even think about. But that's a kind of familiarity breeds contempt trap. Wu in his book of 66 commentaries used the lighthouse concept 13 times. In the same way that I'm not trained to think of all groups as weak, I was inclined to think of some of his examples as too obvious (though with the major caution that I did not actually think of or notice the concept myself). And it also It took me a while to realise I kept looking at the lighthouse when I should have been looking at the ships and wondering whether their sailors were in peril. There were also instances that were not obvious. They were a bit crushing once I applied due thought. In other words, what seeped into my mind was: "Oh, so that group as well is weak!"

As a concept rather than a passing thought, this lighthouse idea was new to me. But by a fluke I got a chance to make use of it for myself almost straightaway this morning. I was transcribing a recent game between Shibano Toramaru and Gu Zhihao in the Chinese Weiqi League post-season.

Having reached the position where White was about to play the triangled move, I initially entered it one point to the right - a knee-jerk reaction (ikken-tobi is never bad). When I realised my mistake, my next reaction was to think, "That's a duff move, surely. It leaves a bad weakness." Then the lighthouse keeper kindly turned on the lamp and I realised - there were ships in peril. This was the helping hand. It was shining in all directions.

Although I gave myself a mental lollipop for that, just another couple of moves further on I realised I had been a little shallow. White launched another ship at A. The triangled move shone a beacon for that, too! But that move A also made me realise I had glibly assumed Black's upper-left group was a string group. It was in fact a weak group. It was strong (or not weak) if you limit your definition to "it can live if attacked." But move A exposed the fact that it had a weakness. A group that contains a weakness must be weak, no? I immediately resolved to use that as my working definition for a while. I don't know how that will work out but I'm very optimistic now of being likely to label more groups a weak! I also now see many AI-type moves as lighthouses, incidentally (including the notorious shoulder hits). It's been a very useful concept for me.

As these thoughts about the "strong" upper-left group were swirling through my brain, I chuckled to myself that even the White group at the top was weak, too. Oh, yeah! Blow me, when the dust of the full game had settled. White had lost almost that entire group. The weakness, which I did not spot, was a latent empty triangle. Clearly that group could have lived if attacked directly. But it was still weak - because it had a weakness.

Whether my analysis is accurate in detail or not, I think all this is a valid broad example of how, as knotwilg says, a step-change in thinking rather than slow, incremental change by refining current theoretical models is needed.

Knotwilg · **#195**

I'll reply to John later because I believe the 6 moves that went over my head were precisely examples where AI saw scope for bullying and I didn't.

Now for the connection/cutting issues:

1) In this game I disconnected White's top stones and connected mine. As we can easily understand :b3:

is a low value connection while :w4:

grabs the life with territory in the corner.

Click Here To Show Diagram Code: [go]$$B Moves 53 to 56 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . 4 . . | $$ | . . . . X . . . . X . O 2 O O . . . . | $$ | . . . X . . . . . X . 3 1 X X X . . . | $$ | . . . . . . . . . . a O O . . . X . . | $$ | . . X . . . . . . X . . . O . O X . . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X X . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . O O . O O . | $$ | . . . O . . . . . , O . . O X O O X . | $$ | . . . . . O . . . X O O O X X X X X . | $$ | . . . . . . . . . . O X X X . . . . . | $$ | . . . . . . . . . . . O . . . . . . . | $$ ---------------------------------------[/go]

The mistake is only partially due to a lack of reading. When I played this earlier move, I had read it out:

Click Here To Show Diagram Code: [go]$$B Moves 49 to 57 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . 1 . . a . . | $$ | . . . . X . . . . X . O 6 4 2 . . . . | $$ | . . . X . . . . . X . 7 5 X 3 X . . . | $$ | . . . . . . . . . . b O O . . . X . . | $$ | . . X . . . . . . X . . . O . O X . . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X X . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . O O . O O . | $$ | . . . O . . . . . , O . . O X O O X . | $$ | . . . . . O . . . X O O O X X X X X . | $$ | . . . . . . . . . . O X X X . . . . . | $$ | . . . . . . . . . . . O . . . . . . . | $$ ---------------------------------------[/go]

I had read this far and concluded White could not cut me. I had not read any further because I didn't need to: commection achieved. I hadn't considered A next nor the fact that B would next ruin my shape in sente

On the other hand, the fact that White played it regardless, should have made me re-evaluate. I didn't. I was confident in my reading and evaluatiom skills. So there was definitely an element of vanity at play.

jlt · **#196**

John Fairbairn wrote:

"Knowing if a group needs a defensive move": I think the writer means us to infer that such a group would be a "weak" group.

It depends on your definition of "weak". Let's say a group is "weak" if it has weaknesses (=will need a defensive move some time in the future), and is "very weak" if it urgently (=now) needs a defensive move. Then my point was that knowing if a group is "very weak" or not requires reading skills. But I also agree with the main idea of your post: professionals see many weaknesses that amateurs are generally unable to see.

Knotwilg · **#197**

Here's a second example

Click Here To Show Diagram Code: [go]$$B Moves 47 to 47 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O O X O . . . . . . . . . . . . | $$ | . . . O X O O X X . O . . . . . . . . | $$ | . X O O X X X X . , . . . . . O . . . | $$ | . X X O O a X O X b c . . . . . . . . | $$ | . . . . . 1 O O O . . . . . . . . . . | $$ | . . X . . X . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . X O . . . . . . . . . . . . . . . . | $$ | . X . . . . . . . . . . . . . . . . . | $$ | . O X X X X . . . , . . . . . X . . . | $$ | . O O O . O O . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

In the past I would have cut at A without even thinking. Interestingly I paused to consider B, taking influence at the top. I concluded this would be too soft and still went for the cut. KataGo recommends C and thinks A is a grave mistake.

Here too it was not a matter of reading but a decision not to read any further because I might conclude against my desire to cut. This desire is very strong.

Knotwilg · **#198**

3rd example

Click Here To Show Diagram Code: [go]$$W Moves 88 to 88 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . X . X X X . X . O O O . . . . | $$ | . . X , . . O O . , . X O X X O O . . | $$ | . . . . . . . . . O O O X . . X O . . | $$ | . . X . . X . X . O X X . X . X X . . | $$ | . . . . . . . . . 1 O X . . . . . . . | $$ | . . O . O . O . . . . . X . X . . O . | $$ | . . . . . . . . . . O O O . O X X O . | $$ | . . . X . . . . . , . X O . O , O . . | $$ | . . . . . . . . . . . X . . O . O . . | $$ | . . . . . . . . . . . . X . X X O . . | $$ | . . . . . . . . . X . . . . O X O . . | $$ | . . X . . . . . O X . . . X X X . . . | $$ | . . . . . . . O . O X . X . . . X . . | $$ | . . . X . . . . . O X . X O . O X . . | $$ | . . . . . O . . . . O O O . . . O O . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

This was a slow connection. In this case a lack of reading was definitely due.

Click Here To Show Diagram Code: [go]$$W Moves 88 to 88 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . X . X X X . X . O O O . . . . | $$ | . . X , . . O O . , . X O X X O O . . | $$ | . . . . . a 8 b . O O O X . . X O . . | $$ | . . X . . X 7 X 6 O X X . X . X X . . | $$ | . . . . . . 9 5 3 2 O X . . . . . . . | $$ | . . O . O . O . . . 4 . X . X . . O . | $$ | . . . . . . . . . . O O O . O X X O . | $$ | . . . X . . . . . , . X O . O , O . . | $$ | . . . . . . . . . . . X . . O . O . . | $$ | . . . . . . . . . . . . X . X X O . . | $$ | . . . . . . . . . X . . . . O X O . . | $$ | . . X . . . . . O X . . . X X X . . . | $$ | . . . . . . . O . O X . X . . . X . . | $$ | . . . X . . . . . O X . X O . O X . . | $$ | . . 1 . . O . . . . O O O . . . O O . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Uberdude · **#199**

Knotwilg, in second you should be overjoyed to make your opponent connect on dame, rather than cutting on dame, particularly with the heaviness of the 4 stones.

Knotwilg · **#200**

Played two practice games with focus on not cutting or connecting when the cut is dubious or the connection has little value. First game was funny as the opponent ended up with 7 groups. Second game I postponed a cut until I saw its value and then went for it. Easy victory again.

I also refrained from attacking a weak group from too close too soon in the opening. That also worked out well. Same with choosing to defend my own central group even if it allowed the opponent to make some territory.

1k-2k on KGS is probably too easy for me even with handicap. So now I want a couple of 1d-2d games in with the same focus.

Knotwilg's practice

Who is online