Monday, August 10, 2020

Tracking the Time I Spend on Go Study

About a month ago I started using a program called Now Then Time Tracking Pro. It is an IOS app designed to track the amount of time one spends on various tasks. I thought this might have some value in focusing my attention on how I use my time for go studies.

This chart represents my first month of activities.

Last month I spent 62.38 hours on go activities, which is roughly two hours per day on average. That is a modest amount of time to spend studying go. I should be spending more time than that. 

I have created twelve different categories to track my study time.

Since it is difficult to see all of the details in the full screen capture of the app I will show each activity separately starting with the activity that got the most attention, and ending with the activity that got the least attention.

Correspondence Play accounted for 11.44 hours, or 18.35%, of my time. I play correspondence on OGS and the majority of that time is spent pondering variations. I spend way too much time doing so, and should probably play more live go than correspondence go.

Twitch and YouTube accounted for 10.49 hours, or 16.81%, of my time. I watch Clossius on Twitch, and also a few go related videos on YouTube. I also recently watched live commentary on the AGA channel on Twitch during the e-Go Congress. This is probably the category that is the least personally engaging. The games are not my own, so my attention wavers and the experience probably offers the least value for the time spent. It is more along the lines of entertainment than study.

Yunguseng Dojang activities accounted for 8.63 hours, or 13.83% of my time. These activities include live play. It was difficult to decide whether to separate out the time spent playing go, but I decided to include it because then the results for playing Live Go show the amount of time I spent playing games that I was not being forced to play as part of the Dojang. That is an important distinction. In addition to live play Yunguseng Dojang time includes game reviews and lectures which I can watch as they happen, but I can also review them later. These game reviews are engaging because they are for games of people I know I will be playing again. They are people who have become friends. I concentrate on the reviews of players in my own league, which is the C league, the lowest one in the Dojang.

Watching Go Games accounted for 5.25 hours, or 8.42% of my time. These are mostly games of my friends in the Yunguseng  Dojang, though some are games of friends I met through the AGA at either Go Congresses, Tournaments, Workshops, or local go clubs. I am not sure if this should even be counted as go study as it is much more like entertainment than study, but I believe it has some value.

Spaced Repetition Problems accounted for 5.13 hours, or 8.22% of my time. These problems are derived from lectures taught by Guo Juan as part of her Internet Go School. I’ve been a member of the school since near its inception, with a few breaks because I was taking sabbaticals from go study. I usually do 30 problems a day which takes me about ten minutes. I am thinking of doubling my problems per day because I have a backlog of over 500 problems which I would like to catch up on.

Yang Lesson Review accounts fro 5.08 hours, or 8.14% of my time. I studied with Yilun Yang off and on from 1998 until 2015 with a four year break in the middle during my last four years of teaching. I felt as if I was not getting the most from my lessons at that time because my job took up way too much of my time. In 2015 I decided to take a break from go and stopped playing altogether except at Go Congresses, which I continued to attend. I have recently started to go back and review my previous lessons with Yang using Smart Go Kifu. I am currently working on lessons from 2010.

Studying Go Books accounts for 4.82 hours, or 7.73% of my time. In spite of the fact that I have an extensive library of go books, my current go book study consists of reading books in GoBooks, the app from SmartGo created by Anders Kierulf. I recently downloaded the entire Learn to Play Go series by Janice Kim, and am working my way through it again. I have the books, but I much prefer to read go books digitally because I always felt the need to study go books with a board and stones which slows things down considerably.

Tsumego accounts for 4.67 hours, or 7.49% of my time. Up until recently my tsumgo study has consisted of doing problems on I just downloaded the BadukPop app at the suggestion of another student of the American Yunguseng Dojang. I also study the Graded Go Problems books, but I count that as reading go books.

Playing Live Go accounts for 3.15 hours, or 5.06% of my time. This is the category that most needs to be increased. Next month will show an increase due to the e-Go Congress and the Yunguseng Summer Stage. It will also increase because I am beginning to play more games with friends from the Yunguseng Dojang.

Guo Juan Lectures account for 2.15 hours, or 3.45% of my time. I decided that it made sense to separate the time I spend in The Internet Go School on go problems and lectures. Currently since I have a backlog of problems I am not watching new lectures unless they are beginner lectures which I really enjoy. Guo Juan does such a great job providing materials for lower level players.

Reviewing Games accounts for 0.92 hours, or 1.48% of my time. This is a big weakness. These are the games I play outside of the Yunguseng Dojang. They mostly don’t get reviewed, and that is going to change. I am a supporter of OGS, so I have access to AI reviews of all of my games there, as well as anything I see fit to upload, but I do not take full advantage of that. I recently subscribed to a basic membership at AI Sensei which I like a lot more. I intend to discipline myself to review my games using AI Sensei.

Replaying Pro Games accounts for 0.63%, or 1.02% of my time. I reviewed one Shusaku game using SmartGo Kifu last month. I am not sure that there is much value in reviewing pro games for me, but I will keep it as a category.

Initially I thought I would include a category for blogging, but I rejected that idea after I realized that it would dwarf the other categories, and not provide much real value for go study.

Friday, August 07, 2020

Ratings Drift In Action


I fear playing rated games, and I have written about this before. One of the bad things about not usually playing rated games is that when you finally play only a few games you will become the victim of ratings drift. This is when your rating goes up without you playing any games. The image above is an example of classic ratings drift.

I am not entirely sure why this happens, but your rating can go up pretty high before eventually falling off altogether.

I played two rated games in March in an attempt to get a rating for my buzzsaw account. I lost both of those games to a 9k player to get a rating of 10k?

Over the last five months I have noticed my rating rise all the while keeping the question mark because the server is unsure of my rating. 

Today I was observing Guo Juan reviewing games for the e-Go Congress. She complimented me on my rank. I heard her in Discord as I was exiting the room. I came back to the room to disillusion her, but she had already left. So now I need to write an email to tell her that I am a victim of ratings drift.

Last night I played a game with my other KGS account, goddess. It was a rated game with a 9k player I know to be my strength. Sometimes he beats me and sometimes I beat him. I had decided to make buzzsaw NR, and start playing rated with goddess. I used to have a rated account reserved for reckless play named fearless. That account would have long been inactive, and I do know if it is available.

I won the game against the 9k last night, and now I am 8k? I need to play rated with goddess before she drifts to 5k? :-)

Monday, August 03, 2020

Go Congress Trouble Master Problem

This is the third round of the Open in the e-Go Congress.

This game was clearly lost, but I found something at the end of the game that could win it for me.

I am white. Here we have a clearly lost game. I played poorly and there is nothing left but dame. Is that true? Is there some trouble for white to find. Hwang In-Seong has a lecture series he calls “Trouble Master” in which he asks students to find situations near the end of the game that can result in an upset of some sort.

At this point in the game I realized that I could capture a great many stones with an atari at s11, but only after the liberties at o16 and p16 are filled. AGA rules provide a great opportunity to pull off this type of sneaky play because they require the filling of dame. If I filled those dame under Japanese rules my opponent would get suspicious. Even in an AGA game one would not want to fill them in succession.

I read out the whole sequence. I knew that I would be able to gain enough liberties by driving black to his stones at t6. I knew that I could then atari the q13 stones. If black connected I would then play at m19 to capture a shitload of stones, and win the game. Of course, black could choose to not connect and protect the larger portion of stones, but I hoped that would not happen.

Black cooperated with my plan up to this point though later in reviewing the game we determined that he could have sacrificed a couple stones if he had seen the larger danger.

I followed through with the atari at r11.

Black connected as expected.

M19 should have come next, and the game would have been over. But it didn’t.

A happy story became a tragedy because I saw something. I saw the throw in at r7. I got greedy and foolishly thought I could have them both. I got distracted from my plan. 

I played the throw in. 

Black forced the capture with r9 and then protected against future loss with j18. I suppose he could have just played j18. Once j18 was played I realized I had missed my chance.

Here is the final board position. I lost this game by 1.5 points. 

It was painful. What should have been an epic win became an epic loss. I do take great pride, however, in having found the weakness and having read it correctly. That is a step in the right direction.

My opponent told me in chat after the game that his AGA rating is outdated and that he is 3 or 4 kyu on KGS. So I don’t feel too bad about having been pushed around in this game.

The game can be found here.

Friday, July 17, 2020

It Should Be Ko

I played a game last night on KGS with a friend.

The game took us an hour and a half to play. We used time settings which are as generous as a typical face to face tournament game so we had plenty of time to think and play a serious game. 

I played white.

After black’s response at g3 I felt the need to address the lower left corner. I have been seeing such large corners in my opponent’s games, and they just seemed too large to me. They seem almost selfish. The question was how to invade, and what result to expect. I did not know the answers, but I decided to dive in and see what would happen.

I chose the attachment at d3, which did not work out well for me. I later came to find out that the correct invasion is at c3, and that the result should be ko. I discovered this from a lecture by Guo Juan on the Internet Go School. Details will follow, but let’s see the disaster that resulted for me in my game.

By move 45 it was pretty obvious to me that things were not going to go well here. It might have been prudent to leave the position alone and save it for ko threats later, but I struggled and tried to make seki with no success.

After the game was over I decided to look for a lecture in The Internet Go School which would answer my question about an appropriate corner invasion in the case of a small knight’s move and large knight’s move added to the 4-4 point.

I performed a search on the position

The second lecture in the search covers two corner positions, both of which can result in the position I was interested in studying.

The position came up within the first six minutes of the lecture. Jackpot!

The correct invasion point is at c3.

The result is ko.

Now I know.

Wednesday, July 15, 2020

I Forgot I Was Playing Ko

This game taught me a few things which I will enumerate.

However, first off let me say that I am not a fan of playing ko. I generally do what I can to avoid ko. I know that this is a weakness of mine that I should work on. I've played ko when it was forced upon me, but I don't think I have ever knowingly started a ko in the past.

I did knowingly start this ko out of necessity, and I believed that I could win it. I counted up my ko threats and I knew when I was going to finish the ko. I waited eagerly for my opponent's next move after my last ko threat was exhausted. I was able to be so sure of this because I had the time to think that a live game does not afford.

Before I show the particulars of the game, let me list the things I have learned:

  1. When playing correspondence games remember to look at the whole board before playing. Do not look only at the local situation. (Do not forget you are playing ko.)
  2. Do not invest emotionally in a game. It is just one of many.
  3. Don't get angry. The English speaking go world is small, and opponents are not disposable. Be nice.
  4. The special circumstances of any one game do not make it okay to ask for an undo if you are philosophically opposed to undo except in cases of obvious miss clicks which can happen in live play.

Let's see how this all started.

I am black.

My group on the left was still open toward the center, but it was on the small side. I decided to make at least one sente move with a13. I expected my opponent to play at a14. Instead he played at c12. After determining that I could endure the resulting atari I extended into the white territory in the upper right.

I eventually made two rooms with b10. This group would later offer white a ko threat. 

The game progressed. I made a grab for a big end game sente move by playing at m18.

My opponent ignored my move and played h8 to threaten the single stone at g8.  I knew that the single stone was small, and  I played k18 to jump into white’s territory. White answered at k17. Then I made a mistake as critical as white’s tenuki. I jumped to h18 instead of playing nobi to j18. I saw it immediately, but I wasn’t going to ask for an undo. We had each made a critical mistake, and a ko was definitely on the way.

White pushed at L18 as expected. I played under it to L19. White gave atari at j18. Before setting up the ko I played at h15 to make the two black stones at h14 and g15 useful. I might actually catch nine stones. The atari was indeed sente. White had to answer.

Edit: A comment from the Facebook Go group provided this variation that eliminates the need for ko altogether. This is beautiful.

Edit: Someone else suggested that since I missed the move at j16, I should have saved h15 as a ko threat, and should have started the ko immediately.

White did answer at j16. I set up the ko with j19, and white initiated the ko with k19.

By my estimation I had three ko threats, all of which were against the p13 group. I played to threaten to cut the group with q10.

White responded at q11. I retook the ko. White threatened my left side group with a10, which I had anticipated. 

Edit: It turns out that a10 is no threat at all. The worst that white can do is make seki, so I should have resolved the ko right then and there instead of answering white’s “threat” I was told about this in the Go group on Facebook.

I responded to white’s threat at a9. White retook the ko. I then used my second ko threat at s10 which threatens to destroy white’s second eye.

White answers at r10, and I retake the ko. White makes a threat at n4. If I did not have another good threat I think I could afford to ignore this threat because the two groups it would threaten could line independently. However, since I had another powerful threat I chose to answer.

I decided that the best response was at n3. White retook the ko and I played my last viable threat at t8.

White answered my threat by taking at t10. I retook the ko. Then the waiting began. I waited and waited and waited for what seemed like an eternity. In the mean time I was playing other correspondence games. I had every intention of ignoring any ko threat white would make, but that is not the way it went.

I remember vividly that I was sitting on my front porch enjoying the outdoors with my iPad in my lap. A blizzard of games came up. I responded to them quickly, in uncharacteristic fashion for me. This game was the last in the series of games. I looked at the local situation, evaluated it, and answered the threat unaware that the game I was playing was the one I was so anxiously awaiting. Big mistake. I answered at s14.

Emotions ran high. I realized my mistake immediately, and did something I never do. I asked for an undo, and explained why I was asking for it. I was ashamed to admit that I had forgotten the game I was playing. Apparently my opponent had taken his move before he saw my text. He explained later that he would have granted the undo if he had seen it in time, which was very nice of him.

His comment, however, was ambiguous as it was, “Sorry. Saw it too late”, which I interpreted to mean that I had seen my mistake too late, rather than that he had seen my request too late. I saw red, and responded with, “Fine. If that’s the way you want to be.” I then proceeded angrily with the rest of the game until we chatted at the end, and it became clear what he had actually meant by his comment. Chat is so easily misunderstood.

With the undo not granted I considered resigning, but I knew that was rash and immature. It was more immature than even asking for the undo. Besides, it was still a game, and it ended up being only a 7.5 loss. It turned out to be even on the board at the end. White won only by komi.

The move I played was at e17. (See above) It was not meant to be a ko threat. I knew he would finish the ko. I thought I might have a chance to live on the left side. AI likes N7 (above) and showed a variation I will not bother you with here. AI knew he would finish the ko too. These screen captures were taken after the game was over. No AI was consulted during the game.

Above AI told me what I already knew about my move at s14. I should have taken the ko. Yeah. Rub it in, AI.

AI did not like white’s response at m19 either, preferring the take at h19, probably because of the aji of my move at e17.

I struggled to live in the upper left, but failed. Here is the final position. AI recommends filling dame, probably because we are playing with AGA rules. I could have snatched three points I guess.

This was a loss by 7.5, which is an even result on the board. White wins by komi. It was a good game. I’m glad I finished it, and that I spoke with my opponent at the end.

When I went back to review the game I saw that my opponent had suggested a rematch but wanted to know if I wanted it with undos or without. I responded that I wanted it without undos because I am philosophically opposed to undos on principle, and that I had learned my lesson. I then told him that I was planning a blog post about our game. 

The game can be found here.

Any comments are appreciated.

Monday, July 13, 2020

OGS Game With Fred8 Reviewed by Cornel Burzo

You can find this review here.

One of the things I like about OGS correspondence games is that sometimes you end up chatting a bit with your opponent and making friends. Such is the case with Fred8. We have played three games together so far, the last of which I was able to win. This is that game.  It was a fairly close game, throughout,  but he appears to have lost a bit of concentration near the end. Therefore, I was able to take the lead.

We finished this game yesterday, and much to my surprise, the review was up on YouTube this morning. I found a link to it on Facebook and watched it with pleasure with my morning coffee.

I enjoy Cornel’s review style. He explains clearly, covers a lot, and goes slowly enough that I can follow without pausing the video. I like that. I will be watching it again along with others of Fred8 so I might get more familiar with his style. 

Fred8 and I are working on our fourth game together already. I fully expect to lose this next one, but I shall try my best to give him an interesting game.