Life In 19x19

I hope that this advice could help a beginner to improve to the level of western (AGA-OGS-EGF) shodan. It's only my approach to the topic, and commentary is welcomed, especially from dans. Some of my suggestions might be controversial.

Playing
----------

* Finding a teacher is advantageous but not essential. Both commercial and free leagues, as well as clubs, offer an opportunity to play games that are more serious than server automatches. Try to attend tournaments, in order to play games that are even more serious.

* Refuse to tolerate rank anxiety. If need be, use OGS' option to hide ranks, but that only evades the problem instead of addressing it.

* Play on at least one western and eastern server (eg. OGS and Fox), for style contrast.
* Play one or two games a day. Clossius recommends ten a week. If time is an issue, try to play at least every other day.
* Play at least a quarter of your games with stronger players and no more than a quarter with weaker ones. Be prepared to give and take handicap.

* Don't use in-game analysis, which is a crutch that will keep your reading weak.
* Don't use in-game score estimation tools, which are a crutch that will keep your counting weak.
* Don't use in-game AI, which is (ethics aside) a crutch that will keep your whole game weak.

* Only play 'live'-timed games (20-60s / move), so that you can think but still experience time pressure.

* Focus on the game. Don't listen to a podcast or keep alt-tabbing to other windows.

* Ignore etiquette. Never be offended by any legal move. Never worry that any of your moves are rude. This is very important. Any legal move, including in the opening, by either you or your opponent, is valid. (Extreme timewasting at the end of the game is an exception.)

* Don't resign before Move 150. Don't resign more than half your losses after that point. You need to steel yourself against defeatism.

Studying
----------

* Review all your games, even if they were painful losses or easy wins. Review them yourself, with your peers, or with your teacher or stronger players. Don't use AI: this is a crutch that will keep your review skill weak. Kevin Yang calls AI review unhelpful until AGA 5d. If you must do, use AI Sensei and set the level of mistake notability to your strength or a few ranks above, and make sure you understand the meaning of the results. Reviewing your games is more important than playing many without doing so.

* Don't learn joseki en masse. Each time you experience a novel corner situation, look it up in the Waltheri professional games database and / or consult a stronger player, and learn some continuations or related lines. Association with playing memory will help the information stay in your brain.

* Ask stronger players to explain positions and sequences from their games. Twitch is the perfect place for this.

* If you must buy books, only get a few classics. Try to study them through book studies with stronger players, or at least ask them questions about content you don't understand.

* Try to learn to count the board, if just to a basic level at first. Even a ten-point margin of error is a start.

* At SDK, especially after 5k, begin learning basic endgame values. The Guanzi Jiazhi article on Sensei's Library is a good foundation.

Go problems
----------

* Solve / attempt a a hundred or more problems per day. At least two thirds should be repetitions; thus every problem should be repeated at least twice. The BenKyo tsumego club (not run by Ben) works through twenty new problems per day and eighty repetitions.

* Try to read the solution to each problem before playing the first stone.

* Do guan (timed tsumego) on 101Weiqi every day. Remember that life and death situations in your games will often occur when you're in overtime.

Just a few comments if beginners are reading this. I'll be brief since I'm not a dan on the EGF scale.

• Other dan players may have taken very different paths, so you don't need to follow each of bugcat's recommendations. Compare for instance with SoDesuNe's guide.
• Whatever the method you choose, becoming 1d will require to put in a lot of mental effort.
• Becoming 1d is not necessary to enjoy the game. I've met people of any level between 10k and 6d, who have been participating in tournaments for many years.
• Personally I don't have the mental discipline to follow the path of any dan player I know. I don't practice tsumegos daily, often look at solutions too early, often play online games without thinking enough and I review my games with the AI crutch without spending a lot of time exploring variations. If I ever reach 1d by accident it will be an unexpected bonus, and if I stay at my current level (FFG 2k, and "dan" level on some servers) that would be perfectly fine.

Lots of good material on Sensei library for endgame.

This is a good practical approach without too much mathematics:
https://senseis.xmp.net/?DieterVerhofst ... calEndgame

Out of interest, does anyone know where to find the pdf mentioned here for endgame values?
https://senseis.xmp.net/?GuanziJiazhi

Personally I like to do tsumego that are not too easy. So 20-30 a day of those would be a lot. 1-2 hours.

Playing on a real board with lots of thinking time is good. Laying out pro kifu can also be good reading practice.

Most important is to like what you do to keep motivation.

https://senseis.xmp.net/?DieterVerhofst ... mprovement

The perennial balance is how to challenge yourself while remaining motivated. When things are too comfortable, you might enjoy yourself but you're not really improving. When things are too uncomfortable, you might be on the brink of improvement, but then lose your stamina.

What I've learnt in any craft is to enjoy the process, even when it's challenging.

Being chill about gaining ranks helps. It usually takes time. There is also a delayed feedback, if you study and practice in certain way it will take some time to see the fruits of that in your games. If you are impatient and change your ways all the time, chances are that you won't have given it long enough to see the benefits.

Often the smallest things are what has the greatest impact. For example simply doing tsumego every now and then probably has much more impact than going from that to doing 300 in a month, and doing 3000 in a month will again only have a marginal greater effect compared to doing 300 in a month. Therefore don't stress out or overdo it! If you are doing something it is already a lot. It is similar when it comes to playing games, there is a diminishing return with more quantity. If someone were really able to fix one of their typical errors in every single game over a long period like ten years, then they are probably destined to be a top professional At least that would be huge capacity to improve compared to the average Joe.

I think the improvement guide could be more chill, even if the plan is to study intensively everyday it should be easier to settle gradually into a routine. It is definitely possible to overwork and get too worked up about Go to the detriment of ability. Probably if your head is swimming with Go ideas it is time for a rest

kvasir wrote:There is also a delayed feedback, if you study and practice in certain way it will take some time to see the fruits of that in your games.

Another factor to bear in mind that each way of studying the game probably has a slightly distortive effect on your game in addition to the positive gains.

Eg:

- if you study lots of tsume go, you may start to focus on those opportunities in games and can lose the wider perspective for a while
- If you take lessons with a pro or strong player, you may learn to play some moves in the style that a pro or 6 dan would play - but feel cast adrift without the reading and fighting skills to back it up
- If you play against KataGo, your game may become too solid and slow and defensive
- If you review your games with kata go, you may adopt a style of over-tenuking your weak groups and making impossible moyo/territory invasions
- If you study pro games, you may enjoy the style and overall feel of games but find it difficult to play similar moves without understanding the reasons and strength behind the moves.

It takes time to assimilate study in a way that helps your game in terms of results.

I can't help wondering what would Bill say?

My recipe was a bit different. I did lots of studying, mostly reading through pro games, but also a fair bit of tsumego and some strategy books, but I didn't actually play that many games. Maybe one online game a week on average, a couple of games a month at the local club, and one or two in-person tournaments per year. But I'm probably a bad example to follow.

Learning to count made a big difference to me. Being aware of whether you're ahead or behind, when to take a risk and when to play safe. Realising that sometimes you don't need to invade, or don't need to try and capture an invader.

I think "don't use AI" (for reviews) is overstating things. Definitely be careful using AI. Don't trust it blindly. Figure out how to ask it "why did you say that?" and "what if I tried this move/what if my opponent tried this response?" That is, don't just stare at the numbers, actually play out variations and check the difference between good and bad variations. Pay attention to swings of 10 points. Don't obsess over 0.2 of a point in the opening. But AI has a lot to teach us. We shouldn't just ignore it. Learn to use it wisely.

I'm not sure about memorising endgame values. I still haven't done that, apart from the basics of hane-and-connect on the first and second line, a very approximate understanding of monkey jumps, and awareness that endgame moves around the 3-3 point can be much bigger than they first look. Maybe I'd have improved faster if I did actually study these things carefully. Or maybe it's not that useful below mid dan level. Hard to say. In my opinion, most endgame mistakes around SDK or low dan level are failing to notice tesuji moves (especially involving shortage of liberties), and losing sente when you don't need to.

Asa relative beginner, this is a fascinating discussion of the relative merits and flaws of different approaches to learning the game. Please continue!