[Welcome to Sensei's Library!]

StartingPoints
ReferenceSection
About


Referenced by
GoBlogs

Homepages
Holigor

 

Holigor's log
   
  • 25 August 2003

I did not play go during vacations. I played chess with my brother and my cousin. To my astonisment I won most of the games. Chess is a strange dynamic game. Usually I don't get good attacks from the openning. Just like with go, the win or loss is decided in the midgame. And not tactically. Just by seeing the main points of the position. It is strange when your opponent estimates it wrong and chooses self-damaging continuation.

This happens in go as well. If both players make moves with positive value the score is close. The score becomes large in the case when one of the players starts to damage his own army.

I played two games on KGS. Both with 2d players. One lost by 4.5 points (I wish I did not play the edge of the board ko, but just pooled back) and another won by 28.5 points. I did not kill anything in that game. I just built 28.5 points more territory. The opponent got all four corners. I know the result. I tested it several times. It is about 30 points loss.

  • 7 July 2003

SL needs more people to implement my ideas. Besides the snapback workshop we need geta workshop and sqeeze workshop and connect-and-die workshop, honte library etc. We do not need high level people for that. 1k players like me would do OK. The dan players could oversee this work in order to avoid mistakes and teach more difficult things like fuseki, side formations, frameworks, game flow etc.

Some work on the miai value lists for the frequent positions should continue as well.

  • 2 July 2003

I have a vague feeling that all yose moves have to be considered as gote. The principle of cooling works only in the case when a player has to pay tax. If the opponent replies there is no tax.

In this case the sente follow up has to be added to the value of the move right when it is performed.

But I have to admit that the double sente situations don't go this way. In these situations the first player wins without a tax. The only way to counter this is to play another double sente (mutual damage).

Another consideration that has to be taken into account is the conservation of the ko-threads. There are obviously zero value ko-threat, but there are ko-threats with positive value the conditions under which they have to be kept are not clear.

Another obscure thought. The game theory builds a matrix of strategies and the win is simple: just choose the strategy that gives the best score against the toughest opposition. There are also games with nature in which it is wrong to choose this safest strategy, because the opponent would not choose the toughest line and the safe approach loses to the line that involves some risk.

The perfect opponent and the nature are two abstraction, of course. There could be an opponent that has some finite ability to find the correct strategy.

  • 27 June 2003

In the last Honinbo game (5th in the title match) Cho U played greedy go. Just got enough territory and survived with his weak groups. The ko for the life of a big white dragon did not give black enough compensation. As the result white won by comfortable 5.5 points.

This is an issue in my games. It seems too risky to decide that my group is alive already. Sometimes I decide wrong and have to pay for it. The risk is very high.

  • 26 June 2003

As before I can't win. Several last wins went along the same path. First, my position becomes desparate. I have to play semeai and see how the opponent could win it. The opponent makes a mistake and it is me who kills. As the result a win with a comfortable score. I play too slow. Opponents play faster, scatter the stones on the board freely and I have difficult time then to find openings. I can make some kind of breakthrough against the players of my level. But the strong player would not make mistakes anymore.


  • 23 June 2003

The great experiment with the kogeima shimari invasion ended as I aught to end. Let's hope that several defeats will make me wiser.

So, I started with an empty board, immediate kakari and invasion, for the purity of experiment. There were no stones on the sides at the moment, so the opponent had to choose the variation I wanted to explore. The problem was that if I followed joseki here I would get a horrible position.

[Diagram]
joseki

On the empty board this has to be a disaster for white. White got only 7 points of territory in exchange for the iron wall. White had been unreasonable.


[Diagram]
19x19 diagram

The folowing result was actually comparatively good for white (white lost by 1.5 points in the game). Though the group in the corner is dead, black's second line crawling gives white good compendation.

But in the second game, the opponent did not go for the corner kill, just prevented being pressed to the second line, and white had to return back to protect the corner. That was a burden and game the result comparable to the diagram above. 20 point loss for white.

So, the conclusion is that returning to protect the corner is a burden. The cut is not good without some help, but if that help is available, the opponent would not chose to block the slide. The book was right, this is bad.



  • 19 June 2003

It is remarkable that a large share of the go games are close. And that is true about the games at any levels. The games between 9p are close and the games between 1k are often half pointers. That's because the moves and decisions we make are reasonable. If each of the players makes moves in the range close to the value of the optiomal move in the position, say, say in the range of five points then it is a problem of random walks and we know that with the large number of steps the random walk would not go far from zero. Yet, if the probability of one player making a suboptimal move if higher the result shifts strongly in the favour of the opponent.

The things are changed drastically when one of the opponents makes an incorrect judgement and produces a move of negative value.


Counting is a problem. In a recent game when my opponent pondered the next largest move I counted the game and obtained the sad result. I was losing by about 4 points. This I consider close, so we finished it and I was surprised with 3.5 win. To test my ability I recounted in mind and got 1.5 win this time. Shame.


Lost a game by 2.5 points. And this is one of the reasons.

[Diagram]
An overplay

White 1 is another double digit kyu misread which IGS 1k* players commit all over the place. Crawling to 'a' was a good 1 point miai value move. I jumped to W1 making a move with negative value. Minus 1 point (miai) sente. It would work if white had a stone at b. But alas...

Problem for double digit kyu players. How can black punish white?

Holigor's Log problem 1 answer



  • 18 June 2003.

There are several things that are difficult to understand for the student of go. Some of them go against intuition. For example, it is better to have a fight in your own sphere of influence. A weak player tries to defend what he thinks is his territory and as the result the territory is erased and the more the game progresses the more it shrinks. Yet, the correct approach in most case is just to keep your groups alive and leave the matter of collecting the territory for the yose stage. Instead, one has to think more in the terms of getting some advantage from attacking invaders, to build a support for your own invasion, for example.

Another thing that hinders the progress of many players is the belief that the centre is small, that there is no territory in the center. Yes, true, there is no territory there, but one has to fight for the center anyway. Otherwise the opponent will be able to form high and wide rectangles of territory and that is not small. Beginners are taught that the centre is not important and the understanding to what extent this claim is true takes a lot of time and lost games.


  • 13 Jun 2003. The SL's mistake page is awesome. What a bounty of different mistakes one can make in the game. It is important though to get rid of the mistakes that are several levels below your strength. I believe that if you are able to solve all 5 kyu problems in your game you are at least a shodan.

[Diagram]
Something from yesterday's game

Fighting the last half-point ko I made a ko-threat. I am sure that any player of IGS 1k* level can solve this in 95% of cases, but my opponent decided that I was bluffing. Perhaps, he felt that he was losing by a small margin. He connected the ko and resigned several moves later.

This can be a problem of double digit kyu players. What White is going to do if Black plays tenuki?



Games against the opponents of the same level don't give the proper feedback. Your moves are not necessarily good even if you did well with them.

HolIgor's Log 2003 June 13 Discussion.



This is a copy of the living page "Holigor's log" at Sensei's Library.
(OC) 2003 the Authors, published under the OpenContent License V1.0.