Knotwilg's practice

Knotwilg · Post by **Knotwilg** » Sat Feb 13, 2016 5:48 pm

I analyzed 6 more Gu - Lee games for the endgame and 4 other pro games. I will post the analyses here soon.

ere are my findings:

1) A player will first execute his major sente and then take the biggest gote
2) What is sente? A move which has a bigger follow-up than any move on the board.
3) when does a professional execute his sente? a bit before it becomes the biggest gote for the opponent; so in effect, the opportunity to play "reverse sente" in the sense that it prevents the opponent to execute his sente never occurs in a pro game
A reverse sente is only played if the remaining gote are equivalent
4) in particular, 1 point sente moves are played very close to the end and kept in reserve for ko fights
5) an endgame play that affects the L&D status of a group, even only in follow-up, is always a bit bigger than it seems
6) in confused areas they leave options open until they find the best move; you rarely see brutal point taking
7) endgame tesuji make a difference of various points per move
8) a large endgame move is not necessarily followed up immediately; a good example is the monkey jump
9) making miai is another common tactic: rather than executing trivial endgame, they will look for a move that makes profit in itself and creates a follow-up which leaves two such endgames as miai

6 of the 10 games I studies were between Gu Li and Lee Sedol. Nearly always when there is a small victory, it's Lee's.

Lee Sedol's endgame is a little different and harder to predict. He tends to pay even more attention to stability of groups and doesn't protect open territories if they're not critical eyespace. He is also not afraid to play ko in the endgame and will reinforce with a gote move in order to prepare for the ko.

Knotwilg · Post by **Knotwilg** » Sun Feb 14, 2016 12:27 pm

I present this game as an exercise. It starts at move 152.
I was able to correctly predict the last 40 moves (sometimes luckily so but still an achievement I was proud of).
I do question Black 169, so any comment is welcome there (all comments are welcome everywhere).
Eventually White resigns so this was not a close game but still one that I found interesting from an endgame perspective.

Bill Spight · Post by **Bill Spight** » Sun Feb 14, 2016 1:11 pm

There seems to be something screwy with your last post. I get an unresponsive script error and the game does not display any moves. Nor can I download the SGF file.

Knotwilg · Post by **Knotwilg** » Sun Feb 14, 2016 1:17 pm

I edited it but it still behaves oddly. I think you'll be able to browse it but indeed downloading doesn't work.
I sent it to you via email.

Jhyn · Post by **Jhyn** » Wed Feb 17, 2016 6:35 am

Knotwilg wrote:I present this game as an exercise. It starts at move 152.
I was able to correctly predict the last 40 moves (sometimes luckily so but still an achievement I was proud of).
I do question Black 169, so any comment is welcome there (all comments are welcome everywhere).
Eventually White resigns so this was not a close game but still one that I found interesting from an endgame perspective.

Are the A, B, C, D situations all your work or did you take inspiration from a pro commentary?
Eitherway, thanks for the great yose training.

Knotwilg · Post by **Knotwilg** » Wed Feb 17, 2016 7:34 am

They're my own work. So it's to be taken with a lower dan grain of salt.

Knotwilg · Post by **Knotwilg** » Fri Feb 26, 2016 3:04 am

Recently I did quite some analysis of professional games. I've always been sceptical of the activity as a way to improve because it's hard to understand a pro game and if you do, the time you invested in it could have been spent more efficiently. On the other hand, analyzing pro games is intellectually stimulating in itself.

I found it very useful to analyze a pro game yourself first and then look up a professional commentary of the same game. Of course you start the other way round: pick a commented pro game and first analyze it yourself without looking at the commentary. Your own analysis will create a sort of landscape in the game: things you've analyzed and understand, moves that you don't understand, trivial moves or sequences. When looking at the commentary, you are visiting a familiar landscape again, this time with a guide who sees things you don't.

1. You analyzed something and it is confirmed by the pro: reward
2. You analyzed something and the pro doesn't mention it: possibly trivial at pro level but not for you
3. You analyzed something and it is denied by the pro: thinking to be corrected (important!)
4. You did not understand a move and it is explained by the pro: new knowledge to ingest (not as important)
5. You did not understand a move and it is not explained by the pro: apparently it is trivial for the pro and not for you; investigate
6. You did not analyze a move but the pro says something about it you didn't consider: thinking to be corrected (important!)
7. you did not analyze a move and the pro says nothing about it: possibly trivial

This kind of analysis is much richer than merely reading a pro commentary, where all that happens is scenario 6, which can hide 1 & 4. A scenario that doesn't happen is 3, while it is probably the most important one.

Knotwilg · Post by **Knotwilg** » Mon Feb 29, 2016 2:29 am

Another endgame analysis of a pro game. Starts at 195.

Knotwilg · Post by **Knotwilg** » Fri Mar 04, 2016 10:27 am

Not many people take an interest in my ongoing endgame training, but anyway. Here's a blitz game where I wanted to see how well I intuitively play the endgame. I took a big advantage in the middle game and lost many points in the endgame, mostly due to poor reading (even some blindness) which caused me to play unnecessary gote and miss sente opportunities.

Any comments welcome, especially on the endgame

Bill Spight · Post by **Bill Spight** » Fri Mar 04, 2016 12:09 pm

Knotwilg wrote:Not many people take an interest in my ongoing endgame training, but anyway.

I have not commented much, but I applaud your effort and endeavor.

ez4u · Post by **ez4u** » Fri Mar 04, 2016 4:06 pm

With 2750+ views to date, I would not call your thread unpopular.

mitsun · Post by **mitsun** » Fri Mar 04, 2016 6:44 pm

Some endgame notes:

In your endgame start position, I think the correct move for B in the upper left is A, rather than the atari to the left. It looks like this reduces the W corner to only two points, while keeping sente. (And in a blitz game, it would be easy for W to make a mistake and die.)

Later on, the descent to R1 looks stylish, but leaves B with a 2-point gote play. Simply atari at S1 is better. Then W should also get R4 in sente, before it is profitable for B as reverse sente.

Knotwilg · Post by **Knotwilg** » Sun Apr 03, 2016 1:24 pm

Today's game fully analyzed.

Knotwilg · Post by **Knotwilg** » Wed May 04, 2016 1:47 pm

I have two questions in this game, at 38 and 96 but any comment is welcome

Knotwilg · Post by **Knotwilg** » Mon May 09, 2016 8:17 am

Encouraged by the heartwarming response to my posted games, I'm off to a new challenge. This week I'm on a business trip and I plan to fill the empty hotel hours with some games. I'll try to get to 10 games (1m+5x30s) with the usual focus:

1. Time management
- play opening without losing an overtime period (OT)
- consume 3 overtime periods on important L&D issues or a decisive ko
- consume 1 but last OT on counting the score at start of endgame and find 5 biggest endgame

2. Reading
- in the opening, keep "biggest area" in mind
- read 3 moves wide at every move
- read 3 moves deep for every alternative
- go as deep as possible in the challenging L&D positions
- during endgame, recalculate 5 biggest endgame whenever one is played

3. Attitude
- don't care about the result or the rank, only about the best move to win the game
- reduce risk when ahead
- seek risk when behind
- don't resign out of disappointment
- respect the opponent: he's there for a good reason

Life In 19x19

Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice

Re: Knotwilg's practice