The below is for newcomers to Go. It is intended to illuminate some best practices for learning.
Before I get into it, a bit about me: I started to learn Go in May 2020 as an alternative to chess. I was drawn to it for a few reasons:
First, Go is a Chinese game. Since spending my early 20s in China, I have had limited exposure to China and Chinese. Learning Go has been a great way to reconnect with China and exercise my brain in one swoop. Much more on this below.
Second, Go students experience the same satisfying––if arduous––path of continuous challenge and improvement that students experience training a martial art, like Brazilian jiujitsu. I love the Sisyphean journey of learning new techniques, only to be thwarted by new sets of challenges.
Third, learning Go is comparable to learning a foreign language. I’ve enjoyed learning languages in my life: Spanish and Mandarin. Go exercises the same part of the brain.
Learning to Read: Life, Death, and 9×9
Part of the difficulty of writing about learning Go as someone who knows how to play the game is that it is easy to forget how the game looks to a beginner.
An ignorant onlooker can easily reverse engineer the rules of certain games. Take soccer, for example. Besides “offsides,” the rules are fairly transparent.
Other games’ rules are difficult to reverse engineer. My German stepmother sat through countless baseball games, but America’s pastime remains a mystery to her.
Go is baseball. The ignorant onlooker staring at a game of Go is hopeless. Even looking at a nearly completed game of Go, the onlooker won’t be able to tell who is winning (except, perhaps, by looking at the relative strain in the players’ facial expressions).
I remember those first days of playing Go the same way that I remember learning to read Chinese characters. You cannot learn Chinese by staring at the page. You must learn the characters one by one. In Go, a new player slowly learns to recognize patterns of stones. Is it good shape? The new Go player has no idea.
Those first games of Go are painful. In fact, if I had to identify a reason that Go is not a popular game, it is because of the pain of the first few dozen games.
Imagine being put into a game of baseball (in right field, of course, where Little League coaches across America for a hundred years have put the least talented players), except that no one has told you the rules of the game. That is how the first few games of Go feel. And there’s nothing you can do but play through––or give up and play chess.
And so, regrettably, the newcomer must play 9×9 with someone better than himself. Someone who can explain and review games. This means losing––over and over again. And not even necessarily understanding why you’ve lost nor by how much you’ve lost. You must simply learn to recognize shapes. And after a few dozen games, when asked “Is this good shape,” the new player finally comes to an answer.
I remember in those early days of learning Go I would often think of the Subreddit Awful Taste But Great Execution. A sign that you are learning Go is that your “taste” (recognition of good shape) gets better but your execution lags behind. Be diligent and execution will follow!
My Secret Weapon: Hai Li and His Juvenile Fighters
Just weeks after I began to play Go, I decided to find a teacher. I approached the San Diego Go Club to see if they could recommend someone. My unusual request was that I wanted someone who would teach me in Chinese.
Ted Terpstra, President of the San Diego Go Club and admirably dedicated Go educator, referred me to Hai Li. Hai is a former 5-dan professional player. One of his students is a 9-dan professional international champion. Given the level of his mastery of the game and the prestige of his students, it was kind of awkward to ask to learn from him. Nevertheless, I asked.
Hai welcomed me into his beginner-level class. (I’ve since graduated to his mid-level class.)
Class is conducted in a suite of Chinese software known as Xinbo. It’s outstanding, comprehensive instructional software. My favorite part of it is that it is usable on the iPad, which is the best hardware for digital Go. My other favorite part is its mascot, this adorable smiling chipmunk-reindeer hybrid. Don’t ask me how they were bred.
Hai holds class a few times every week. Having a job and commitments, I join once per week.
My classmates in the beginner class were all Chinese Americans between 4 and 10 years old. Some of them, no doubt, are future high-level dan, even professional players.
Each class is broken down into three sections: Killing patterns and joseki, a game, and a review.
Hai’s curated killing patterns are a big reason my rating has risen as fast as it has. By curated killing patterns, I mean that Hai gives his students a set of puzzles to solve. The solution to each puzzle of the set requires the use of the same killing tactic, a “snapback” could be one such example. Solving fifteen puzzles using the “snapback” tactic is a much more effective way to develop tactical skills than solving random tsumego, in my opinion.
I would highly recommend that a new player search for a group of curated killing patterns like those that Hai creates. I’m not sure of a good online resource. Perhaps someday I’ll create one!
Sometimes in lieu of or in addition to a set of curated puzzles, Hai introduces one or two common joseki. (This doesn’t begin until his mid-level class when students are playing on an empty 19×19 board.)
I think the utility of Hai’s approach to joseki is that, as a beginner, there appear to be a daunting number of options for where to play your stones. Learning a handful of the most common joseki reduces the number of options and simplifies the game. This post on Sensei’s Library covers some critical beginner joseki.
After solving a set of puzzles and memorizing joseki, students break into pairs and play a game of Go. Whether students play 13×13 or 19×19 depends on the students’ rating. The ratings I’ve used below demonstrate how my rating on OGS corresponded the board size I played on in Hai’s class.
25-20k students play 13×13 with stones on the board to encourage fighting.
20k-15k students play 13×13.
15k and beyond play 19×19.
By “stones on the board” I mean that students play a regular game of Go, but play begins with the following pattern:
The point of the stones on the board is so that students rapidly learn basic fighting tactics. In my own experience, playing with “stones on the board” quickly becomes uninteresting. I want to choose where all my stones go!
After the students finish playing, Hai reviews the games of each pair of students so that they can see their mistakes.
Before I go on I just want to make a quick note about my classmates: There’s a big difference between how I play and how the kids play. The kids all like to fight fight fight fight fight. Knowing my own fighting skills to be subpar and that I am relatively wise (old) and can capitalize on their youthful mistakes, I have developed a style of play that avoids fighting: “Always be tenuki-ing” like Alec Baldwin. I have become much more focused on developing strong “macro”––playing stones in broad strokes and ducking fights.
In some ways I think this style has helped my rating and also helped me to develop Go fundamentals.
13×13, Joseki, and Corner Fighting
As soon as I understood concept of “making life” I began to play as much as I could on a blank 13×13 board on OGS. I recommend that new Go players do the same. That said, there are both benefits and drawbacks to playing on a 13×13 board, all related to the tautologic fact that it’s a smaller-than-standard board.
The principal drawback to playing on a 13×13 board is that players learn bad habits. Compared with 19×19 games, corner territories in 13×13 are worth more relative to the rest of the board. It incentivizes players to play very low. Looking back at my 13×13 games, I was able to win many games simply by playing at the 3-3 point. On a 19×19 board, a 3-3 point opening is taboo.
I won so many games by playing 3-3 point openings on 13×13 boards that my rating soared to 12kyu way back in August! I must have developed the hubris of a prodigy.
But I was in for a disheartening: When I finally switched to playing 19×19, my rating plummeted to 17kyu. It was a slow clawing back to 12k which I achieved again only in early November.
The other chief drawback to 13×13 is that it is less interesting than 19×19. Part of what makes Go interesting is the astounding number of permutations of the game. A smaller board allows for fewer permutations.
In my opinion, the good of 13×13 outweighs the bad for new players. There is a Go proverb: “Lose your first 50 games as quickly as possible.” And the chief benefit of the 13×13 board is that it makes losing 50 painful games happen fast! On OGS, 19×19 games are timed at 20 minutes for each player, whereas 13×13 games are half that.
Similarly important for beginning players is that more games means more iterations of joseki and more life-making in corners. By the end of fifty 13×13 games, a new player will at the very least know one or two variations of a 3-3 invasion. And that is critical to know once you are playing on a 19×19 board.
Climbing Rating 19×19: Fuseki
I suspect that each player’s road to single-digit kyu is different. Some players, like the Chinese American youths in my class, probably fight their way to SDK. But the idea of fighting like they do makes me queasy.
As I mentioned above, my strategy for dealing with my bloodthirsty juvenile classmates has been to out “macro” them, which starts with joseki and fuseki.
Hai’s class helped me to recognize which joseki were most critical to my development as a player. But learning joseki is just the beginning: In every game, each corner of the board will have its own joseki. And when one joseki interacts with three others, they engender unpredictable interactions.
Viewed together on the entire board, these various joseki combine to form a fuseki––the whole board opening or “macro” picture of the game. I’ve come to think of the fuseki as a kind of “scaffolding.”
The most critical point of a Go game comes immediately after the fuseki. It is then that each player makes his highest-leverage moves, and the permutations and possibilities of the game manifest. At my level, this is where games are often lost after egregeious kyu-level blunders.
To improve in double digit kyu you must learn tactics. Improvement looks like this: Learn a specific tactic (a joseki, an invasion, etc.); attempt to implement it; fail in implementation dozens of times; observe your rating fall as you fail; experience modest success against weak opponents; and finally watch your rating climb back to where it was and beyond. (This, by the way, is precisely the same road to success in Brazilian Jiujitsu.)
One particular invasion tactic propelled me through the final ranks of double-digit kyu: The 2-5 point invasion, illustrated below.
In my in-class reviews with Hai Li, I played several losing games in a row where he gave me the same tip: If I want to invade the 4-4 / 6-3 enclosure, I need to use the 2-5 invasion. Eventually, I actually listened to my 5-dan pro teacher, and began to implement it.
Part of the fun of Go is that players have autonomy to “force” certain situations. That means, if you’re practicing a tactic like the 2-5 point invasion, you can play it whenever you want! I find this contrary to chess, where the opportunity to use a certain tactic won’t come up unless you force it while playing white.
Next Stop: Dan!
Soon I will be entering Hai Li’s advanced class. That means stronger opponents––maybe even teenagers! Here is what I’m thinking about right now in terms of how to improve now that I’m a vaunted SDK player:
1. More advanced, complicated joseki. I don’t think the process of learning joseki will end until A.I. “solves” Go. Until then, I have joseki to learn. Here’s one that has been giving me trouble recently:
2. Reduction. In general reduction of opponents’ territory is befuddling me. When facing an opponent with a strong moyo or “macro structure,” a Go player must choose to invade or reduce the opponent’s territory. To date I basically still just do a gunslinging “YOLO” invasion. That said, I don’t have any rules for reduction just yet. It seems to be a matter of “feeling” where to play.
3. Tsumego. Though I’ve been solving tsumego for half a year, I still don’t think I am very good at solving them. I wonder if Ke Jie feels the same way. It’s like this in BJJ: Here’s a video of Roger Gracie saying that he considers himself to have a “pretty good closed guard.” Roger Gracie has the undisputed best closed guard of all time.
That’s it. See you on OGS!