Over the last few months I've returned to playing go after a hiatus of a few years, and as I've been taking it more seriously recently I thought it might be nice to start a study journal as a way of seeing how my thinking and play changes over time.

I had another study journal here once, five or so years ago, but it's been long enough that I think it makes sense to start a new one.

Where I am these days: I started playing go again this summer after some old go friends convinced me to try out the AYD Summer Stage, which consisted of three games and a teaching game over the course of two weeks. When I played a few games on KGS to shake out some of the cobwebs, to my great surprise I went on a considerable winning streak and briefly attained a rank of 5k, which was several stones stronger than I had attained previously. I even beat an old friend and rival who's about 5k in an even game. Maybe I forgot more bad habits than good ones. Naturally I went on to go 0 for 3 in the Summer Stage games and my KGS rank soon dropped from that questionable height back down to about 8k, but playing serious games was such fun that I decided to sign up for the AYD fall season, which started in September.

I'm interesting in improving, but I also don't want to become rank-obsessed to the point where it discourages me from playing online, which has happened to me in the past. So in this thread I'll try to focus on other aspects: understanding mistakes, new patterns learned, tsumego, etc.

To start things off, here's a game I played in AYD last week (comments welcome!):



I'm still trying to figure out the best way to do these self-reviews. With the new-to-me AI tools available now it's easy to just click through and see what KataGo thinks were the biggest mistakes, but it's probably useful to go through the game without AI first and try to identify a few mistakes I can understand. If anyone has advice on the best self-review methods for kyu-level players, let me know!

This game leaves me with two takeaways, I think: first, there were a couple of situations where I wasn't sure about reading the result of a cut, but could see clearly that not cutting was bad for me -- I'm thinking of the reduction on the top and my failure to connect on the left side. Probably I should just cut in those situations and find out what will happen that way.

Second, there were a couple of big moves (the 3-3 point and the turn at P7) which went unplayed for both of us for a long time. I should probably try to value moves like that more highly.

When I review my games, I always run Katago to help me

1. Answer questions I asked myself during the game.
2. Try to understand at least one mistake.

There is no need to understand all Katago's suggestions.

In this game, it seems you were too defensive. In general, defensive moves are urgent, but only when necessary. If you don't need to defend, then attack (if there is a good attacking move available) or take a big point. As you pointed out, moves 44, 46 and 48 are overly defensive. For move 48, the point of making the hane is to try to split Black's top side into two groups. But maybe L16 wasn't the best place to start with, and a move like M17 or K17 would have been better (a knight's move on the third line relatively to an opponent's fourth line stone is often a good invasion point).

That makes sense! I like the idea of having questions in mind about the game before analyzing with KataGo.

Here is a fast game I played on Fox (5:00 plus 30 seconds/move). The main takeaway I have from it is the quite bad intuition I had about handling the invasion on the top (starts at move 123) -- all I needed to do was prevent the black group from creating a base, but I kept also trying to cap it and cut it apart, which resulted in the group living and probably decided the game.



I've often wondered whether fast games are worth anything for improving, but I think that at least at my level they definitely are, as long as I'm able to review them afterwards and see some ways in which my intuition is flawed. I think there's probably a lot of value in mileage, too -- I was curious recently about how many games of go I've played total since I learned (~10 years ago), and I think the number of serious games (i.e. 19x19 games against a human opponent) is probably in the ballpark of 200 or so, which is lower than I would have guessed.

5 minutes + 5x30 seconds byo-yomi is fast, but not unreasonably fast if you use the 30 seconds. I don't like this time setting because I find time pressure too stressful, but if you can handle it then why not?

I think you identified by yourself a few shape problems in the sequence starting from move 124. Here are other remarks:

should be at N4. Most of the time, don't allow your opponent to make hane at the head of two stones or, even worse, hane at the head and the bottom of two stones. In addition, at S6 doesn't threaten anything since the corner stones are already pretty safe, and doesn't harm R8 much since it was already separated from R5. So at S6 is just an endgame move, this is too early at this stage of the game and your opponent was right not to respond (although Black should have played at N4).

is strange, it shouldn't have worked so well because should have been at E4 and you would have had a hard time living. at C3 would have been more normal.

Move 102 should obviously have been at B9.

After move 170 you made several endgame moves that were too small but it wasn't game-deciding. At move 170 you apparently have a bit less than 50 points + 9 prisoners and your opponent a bit more than 60 points + 7 prisoners so you can't win unless you kill something.

Thanks for the comments! Playing a move to disconnect already-disconnected stones is something I've noticed myself doing a few times lately.

Here's a game I played on KGS. I made a couple of interesting mistakes. In particular I was curious about the corner invasion in the lower right. In the game I thought this should be killable, but reviewing with KataGo it seems like maybe the white stones on the outside allow White to live or at least do a lot of damage, so the mistake was less how I handled it and more not playing the big move to protect the corner earlier by descending.

I also missed that for a long time, the five white stones in the group around G4 were capturable in a snapback, which was much bigger than any of the other endgame moves on the board, I think.