19x19 FOR BEGINNERS SERIES: Part 4: Joseki basics




[A quick disclaimer for any readers who are SDK and up: this series was written by a DDK for TPKs, with the goal of making the process of “losing their first 100 games quickly” less confusing and more productive. I say this because I know this article will barely scratch the surface of the complex, tangled world of joseki. My goal here is to provide beginners with some focussed, functional tips rather than do a deep dive into bottomless variations. I am also prioritizing joseki that allow players to settle their local exchanges quickly and get back to playing a balanced Opening, rather than starting lengthy life-or-death fights at move 7 (I’m looking at you Flying Knife joseki!)

If you would like to fill in those gaps and expand on the topic, I welcome your input. However, I humbly ask you to take the time to craft your explanation with the 25 kyu player in mind. Chances are your contribution may include elements that are obvious to you, but far from obvious to the beginning player. If you can break your explanation into parts that TPKs can understand, you will be doing both them and me a great favor.]

Welcome back! In my last article, I focused on using whole board play to maximize efficiency in the opening. I thought about doing the next article on the various Stages of the Game, or the Opening / Midgame Transition. But when I really considered my original motivation for these articles - to save beginners months of hitting their head against the same brick wall - I realized I needed to circle back and say a few words about joseki.

The reason I feel strongly about this is I spent the first year of my Go education completely screwing up every Opening I tried. Each response from my opponent was scary and new. I didn’t know about settling my stones, or creating the potential for living shapes. I was playing my stones too far apart and leaving lethal cut points all over the place. Take a look at one of my early games when I was ~20 kyu playing against a 16 kyu opponent:

I am Black - you see those stones in the lower right? Most of them are dead. I spaced my initial stones too far apart, and foolishly ignored White’s threats right up to the moment when I realized I’d lost the whole corner. The rest of my moves across the board are slapdash and scattershot. I was attacking White’s upper left group from the outside for some reason, when I should have been invading that corner. What a mess!

The problem was - I was trying to reinvent the wheel every time. With so much open space between stones, it was very difficult for me to guess where my opponent would respond. I knew that studying joseki might help - but trying to make sense of the variations (upon variations! upon variations!) felt like navigating a maze without an exit. If I had only known then what I know now, I wouldn’t have spent my first year starting all my games down by 20 points.

So let me give you the good news. Joseki don’t have to be complicated! If you can narrow your focus and learn a small handful of variations they will do (pretty much) everything you need them to do until you level up to 15-17 kyu, and it won’t feel like memorizing the encyclopedia.

Think of joseki as a vast Swiss Army knife with specialized blades for every occasion. Even though it comes equipped with a fish scaler, wood saw, hoof cleaner, and cuticle pusher - on a daily basis - you’ll probably keep using the same handful of basic blades over and over (i.e. knife, screwdriver, scissors, etc).

If I can paraphrase - the underlying process of joseki boils down to being politely greedy. All those myriad variations are the result of countless interactions all trying to solve the same problem:

How can I divide something so that it seems like a fair trade, but actually gives me the advantage?

This is necessary because getting a decisive advantage in Go is actually quite difficult. Oh sure - if one player is much stronger than the other - they can devastate their opponent in many ways. But when two players are evenly matched, they know all each other’s tricks. Because so much of the board is there for the taking, any land grab they make over here can be balanced by a different gambit over there.

Go players have spent thousands of years solving these puzzles, and joseki are the distilled results of those interactions - they are the trusted, well-worn path. They give you the comfort of knowing that if you follow some simple steps, you’ll either walk away with about half of the corner you’re fighting over or get something else in exchange. If you combine that with your knowledge of whole board play, you can choose joseki that help your groups work together across the board.

There’s another reason that having a bit of joseki knowledge can give the beginning player a nice advantage - chances are, if you’re playing against another beginner, they won’t understand the right time to tenuki (play somewhere else).

During the Opening - when so much unclaimed potential is free for the taking - it’s in your best interest to approach as many corners as possible. The thing is - if you tenuki before you’ve settled your stones, you might leave lethal cut points that could wipe out your progress in that area. Conversely, if you over-focus on one corner and play another extension from an already settled group, you squander the fleeting advantage of approaching a different corner. Knowing when a joseki sequence is complete allows you the confidence to tenuki appropriately, and gives you a better chance of grabbing unclaimed potential.

As you can see, I’m writing these 19x19 FOR BEGINNERS articles in such a way that each one builds on the ones before it. If the concepts I’m using are confusing to you, you may want to back up and read the articles on

  • Sente and Gote
  • Settling your Stones
  • Playing a Balanced Opening

because I’m going to be referencing a lot of elements from those articles assuming you’re familiar with them. I’m also going to be relying on another useful tool - the OGS JOSEKI EXPLORER.
Hopefully my brief guided tour will dispel some of the confusion and mystery about this Swiss Army Knife and allow you to become more comfortable with using it at your own pace.

Most of the time - you’re going to be using joseki during the Opening. While keeping those classic Opening priorities in mind (Corners > Sides > Middle) you want to make sure you choose joseki which help you accomplish the following goals:

  1. Retain half or more of the corners you claim
  2. Try to get half or more of your opponent’s corners
  3. Avoid getting chased out of that corner or forced into tight enclosures
  4. Help your groups work together across the board to create / prevent extensions
  5. Watch for opportunities to break the order of sente and gote to get an extra advantage

So, let’s start again with that empty board.

If both players are savvy, they’ll use their first 4 moves to claim the available corners. While there are many options, I’m only going to focus on two of them: the 4-4 point and the 3-4 point - just so we can keep the number of joseki manageable.

THE 4-4- POINT: aka “Hoshi” (Amazing how knowing just a few Japanese terms can help you feel like an experienced Go player right?)
When beginners place their first stones, they often cling to an unspoken assumption that “This is my part of the board and my opponent should stay away from it.” They may even get flustered or annoyed when their opponent plays on what they believe to be their corner.

Experienced players assume the exact opposite - by playing in an unclaimed corner you are inviting your opponent to approach it. It’s literally the 2nd most important priority in the Opening - it’s what both players should be thinking about as soon as all 4 corners are claimed. As such, when you claim a corner, it’s good to know which approach moves you are inviting. Here you see the most common ones for the Hoshi, and I will discuss them further below. My point is to help you see that - when you play your first moves - you are already inviting certain joseki sequences, and to help you do so selectively.

The main benefit of the Hoshi is its flexibility - no matter how your opponent responds, you will either retain some part of that corner, or trade it for influence towards the middle. Whatever happens, you will be ready.

THE 3-4 POINT: aka “Komoku”

I’ve marked the most common approach points with O’s. The 3-4 move trades the flexibility to build in any direction for a better chance to retain more than half of that corner. It also has several other strengths going for it:

  • Unless I’m playing a very high-level opponent, I don’t need to worry about a 3-3 invasion at R17 (it’s pretty rare - sometimes used as a ko threat)
  • If my opponent approaches at P17, I can temporarily ignore that threat and play one more sente move elsewhere, before coming back to deal with any additional threatening stones my opponent plays in that corner (I’ll explain why below).

Basically, while you’re making the climb from 25 to 20 kyu, I would advocate that you stick with the trusty old 4-4 and 3-4 moves, because they will help keep the complexity of the Opening more manageable. Take a look at these 3 example boards below:
3 4 ax
Even though Black and White have laid claim to different corners, they all used some combination of Hoshi and Komoku moves, and played in such a way that both can try and develop one side of the board (see my last article for why I’m cautioning beginners away from cross-cut games). Now that we’ve laid the ground-work, let’s dive in.

Ok, so Black and White have claimed all the available corners. Now Black has sente, and our Opening priorities say the next most important step is to enclose your own corner or approach your opponent’s corner. In my last article, I discussed various reasons why beginners may want to avoid enclosing their own corners or playing the Chinese Opening with Move 5 - you’re welcome to circle back and read those if you want. For now we are going to focus on using the sente in Move 5 to approach a Hoshi corner using a knight’s move approach (I’ll discuss the 3-3 invasion at the end).
In my last article on the Opening, I used an example where Black prevents White’s large extension on the left side. Here, I’m going to have Black try and create a large extension on the top side by playing at A. You’ll see this for Move 5 in many games on OGS.
I’m going to show you two short joseki variations, and one long variation, to give you some idea of how you can pick through those Swiss-Army-blades more selectively.

Option A
This is the simplest and quickest variation. It only uses 3 moves and it’s the easiest to remember. Black approaches at F17 - White plays a gote move at C14 to claim the left portion of that corner - and Black settles their stones with J16 to claim the top portion. According to the OGS Joseki Explorer, you don’t have to play J16 to settle - J17, K17, and K16 are all valid options depending on how conservative or risky you’d like to be. https://online-go.com/joseki/14

By playing so few stones, both players are saying “We’ve sketched out some loose connections and we’ll come back to fight for this corner later. Right now, we want to focus on grabbing as much unclaimed potential as we can while it’s there for the taking.”

Option B
In this variation, Black uses one additional move at D18 to push further into White’s left corner - keeping in mind that they’re trying to create a large extension across the top. White is a bit of a captive audience here - playing elsewhere might end with Black completely invading that corner, so White has to respond at C17. Black settles their stones with J16 (or J17, etc) https://online-go.com/joseki/13

Option C
Here is the more modern variation - something experienced players might play. Where Options A and B avoided contact fighting (i.e. stones didn’t touch), here both players attach their stones to provoke an immediate response. The back-and-forth of sente attacks and atari threats solidifies the structures that players in Options A and B were saving for later in the game.

While A and B were intuitive and easy to remember, Variation C requires a much deeper understanding of sente and gote, and the decision trees have more branches per step. In the Joseki explorer, these are your options for Move 6 of the above sequence:

Each of those decisions at Move 6 takes you down a different fork in the road, requiring many other decisions.

Why did I show you these 3 variations when they accomplish almost the same goal? The main thing I hope you take away from this, is that while joseki can be complicated - they don’t have to be. If Black plays Option A and then approaches another corner, they can always come back the next time they have sente and play D18. Or if White gets sente down the road, they can always play E17 to maximize that corner.

While there are still another three corners that can be approached or enclosed, there are many reasons why you might want to wrap up the first corner fight in 3 to 5 moves rather than 11 moves. Why you ask?

I’ve said it before, but it bears repeating - while there is unclaimed potential on the board, it’s in your best interest to settle each corner as quickly as possible so that you can try to grab part of another. As you begin to play more advanced opponents, you may see them start to omit their settling moves, and tenuki before the joseki is finished. Take a look at this fairly typical opening in a game between a 2 dan and a 1 dan.
After the first 4 moves,
5: Black approaches at F17, but White is confident that they can do something in the upper left even if they get invaded (C17) or double-approached (C14).
6: White tenukis at R5 because what they can gain there is bigger than either of those risks.
7-8: Both sides spend one stone to settle, then
9: Black tenukis at F3.
10: White tenukis again at O17, and Black settles with R14.

Would I encourage 20-25 kyu players to play like this? No! These dan players are taking big risks because they have the experience to deal with the consequences. I would encourage beginning players to settle their stones in each corner fight before playing elsewhere. The point of this demonstration was - the more advanced your opponent is, the more they will risk leaving their stones un-settled, and tenuki early because they’re confident they can settle those stones later.

As you study the world of joseki, and make decisions about which ones are best for which situation, please keep in mind that you need to balance the conflicting priorities of getting the most of one corner versus claiming your portion of another.

DEFENDING THE 4-4 CORNER (OK let’s get back on topic!)
It’s useful to bear in mind that joseki have options for White’s defense as well as Black’s offense. Responding to a knight’s move approach by calmly claiming the other side of the Hoshi is one option for White. The slightly more territorial option - if White doesn’t want to divide that corner - is The Kick:
When Black approaches White’s corner with the knight’s move at F17, White sends out a connected stone at E17 to take away one of Black’s liberties. Black extends at F16, and White settles their stones with D14 - giving them a good chance of securing that entire corner. Black settles their stones at the top with K16, which trades a bigger extension for the risk of creating cut points (safer options marked with O’s). In exchange for not getting half of that corner, Black gains a better chance of making a large extension across the top, and - if they get sente and play F14 before White - expanding that potential into the middle.

Let’s take a look at a different board. Here Black and White played on opposite Komoku points in their corners. Black wants to make the biggest extension at the top, so they will try and approach White’s upper left corner at A or B. Which one to choose?

E16 requires a more urgent response from White. If White ignores E16 and plays elsewhere, then Black can follow up with C17 or C15 and put White’s lonely stone on the run. Instead, White can calmly play this joseki variation and secure their corner:
Similar to The Kick, White protects their corner by attacking Black’s E16 stone with a sente move at E17, taking one of Black’s liberties. Black counter-attacks with their own sente move at F17, and White pulls back with a gote move and connects with D17. Black secures the cut point with a gote move at F16. White settles their stones with C14 (I’ve marked alternate variations with X’s) and Black settles with K16 (I’ve marked alternate variations with O’s).

As I said before, playing the komoku gives White a better chance of keeping half or more of that corner, so Black was satisfied with receiving an immediate local response, and getting a good start on a large extension at the top.

If Black plays the knight’s move approach to the komoku, then the choose-your-own-adventure story gets a bit more tricky. Basically

  • White can choose an immediate response, and secure more of the upper left corner, or
  • White can choose to tenuki, approach/settle a different corner, and be fairly confident of retaining some of the upper left corner even if Black counter-attacks.

Let me walk you through the options:

This is exactly like the Kick, except everything is moved over towards the left by one space. I’ve indicated other options for White to settle the stones with X’s and for Black to settle with O’s.

This little gambit completely blew me away when I learned about it. Basically, when White’s left komoku corner is threatened by Black with a knight’s approach, they can still survive and crawl towards reinforcements on the 3rd line, even if Black gets two moves on the left. Here’s an example to show you how it works.
Black starts by approaching White’s upper left corner with a knight’s move at E17, aiming at creating a large extension back to their upper right stone at R16. Instead of responding locally, White chooses to tenuki and prevents Black’s extension at the top by approaching the right side with P16. Black responds locally, and comes out of the sequence in sente (Moves 2-8).

Black then returns to the left side and starts putting pressure on White with D15 (Move 9). White completes this simple joseki on the left and settles with B12. At this point - if there are still corners that can be enclosed or approached, Black can gain more profit by playing elsewhere rather than nickel-and-diming White for the upper left corner. However, once all the other available corners are approached/enclosed, then either side can come back and play in the upper left - Black can try and grab more of the corner with C18, or White can secure the corner with D17. Also, if black plays C18, then White can counter with D11 and get something in exchange.

Unlike the other examples, I do not have an OGS joseki dictionary link for you, because White tenukis for Move 2 of the sequence. However, if you watch live games on OGS, you’ll see people play variations on this theme all the time. Once you become more familiar with how individual joseki are supposed to work, you’ll begin to notice the times that people choose to omit the moves to settle their stones, or to tenuki rather than respond locally. The more you can absorb these types of lessons, the more confident you’ll feel trying the same gambits in your own games.

Before strong AI Go engines hit the scene, playing the 3-3 in response to the Hoshi was seen as a “beginner’s move” - the rationale being that it was too small. The handful of points gained in that corner were seen as a poor trade for giving one’s opponent a thick wall with lots of influence into the middle and up both sides.

After AlphaGo (and all the other AI that came after), the 3-3 invasion became much more popular, and the actual variations for playing those joseki were updated with the more efficient sequences that humans picked up from AI. I will give you links to those more modern 3-3 sequences at the end, but I’m not going to discuss them here. Instead, I’m going to show you the old, simple 3-3 joseki sequences, because they are easier to remember, and you are more likely to encounter them when you play other beginners.

Here is the classic (pre-AI) 3-3 invasion joseki:
Instead of approaching from the outside,

1: Black invades White’s corner at C17
2: White extends at C16 and takes one of Black’s liberties away.
3: Black extends at D17 matching White’s 2- stone wall with their own
4: White plays E17 to hane at the head of 2 stones
5: Black protects with E18, threatening White’s E17 stone
6: White extends with F17
7: Black matches with F18
8: White extends with G17 to keep black from playing there
9: Black plays B16 to hane at the head of 2 stones
10: White plays B15 to protect
11: Black plays B17 to secure the cut point and secure the corner
12: White plays C14 to secure the cut point.

It might feel satisfying for Black to take corner territory from White, but it’s a lot of work for 8-10 points. Plus White gets a nice, thick wall facing into the middle and has a better chance of building a framework down the left side or across the top.

There are lots of modern 3-3 variations that get really complicated, but here is a simple one that you will probably see on OGS.
Rather than playing the hane at the head of 2 stones for Move 4, White simply extends, and Black settles with F18. The players who play this variation want to settle the corner quickly and move on to battling for other corners.

If you ever find yourself annoyed that someone invaded your Hoshi with the 3-3, you can reclaim it with this lovely trick called the Flower Joseki. I’m going to show this one to you in stages, because it’s easy to lose track:
Up to move 5, everything follows the same order as the classic 3-3 invasion, except instead of White extending at F17 for Move 6, White plays the double hane at F18 to bait the trap.
7: Black takes advantage of the cut point and plays F17 to put White in atari
8: White secures the E17 stone with E16
9: Black plays G18 to put White’s F18 (bait stone) in atari
10: White plays their own atari attack at D18 (to set the actual trap)
11: Black captures at F19 only to realize that their original invading stones are now trapped
12: White completes the trap by putting Black’s 2 invading stones into atari with C18
13: Even if Black tries to wiggle out with B17,
14: White plays B18 and no matter what Black does, White can eventually capture those stones.

I’m sure by now, your brain feels full to bursting. The funny thing is - I feel like there’s so much I left out - so much more that I could have said. Part of being a savvy consumer of joseki is knowing when to use one variation over another, and combining your knowledge of what each individual joseki provides for you locally with how your various groups can cooperate across the board. Unfortunately, the experience and training needed to gain those perspectives will be neither brief nor simple. You will have to observe and play many games over many months to get a sense for the give-and-take that these joseki provide, and begin to choose the variations that work for you. Hopefully, this little tour has taken some of the dread and mystery out of those variations, and will help you find your footing so that you can explore more confidently.

I’ve been a big fan of Go Pro Yeonwoo’s YouTube channel. Even though her lectures are aimed at experienced players, her style of teaching is clear and simple enough for beginning players to understand. If you’re looking for some support in wrapping your head around the various joseki options, I would strongly recommend the following playlists:




Good luck and see you next time!


NEXT ARTICLE COMING SOON (or maybe not so soon given my track record?): THE STAGES OF THE GAME


I love this! The Swiss-army-knife metaphor will be forever etched into my memory. Like the others, this article is long, but the pacing is sublime and the time it takes is time well spent. Brilliant stuff! Every motivated beginner should do themselves a favor and read this and your earlier stuff right away.

A few picky comments:

I’d say sente-ish, rather than sente. According to the Joseki Explorer tags, move 3 gives the other player sente. In fact, that’s one big reason why option C became popular in the first place. But that shouldn’t detract from everything else you say. Beginners should still play there to settle their stones, because omitting it brings unneeded risk. Until you know what you’re doing, safety first!

Moves 8 and 10 should be flipped.

Most players save move 13 for a ko threat. The basic joseki ends at move 12.


Good eye!

Fixed the 8/10 flip


WRT the Flower Joseki - I know that experienced players would usually understand that the joseki ends at move 12, but TPKs might think that they can still escape such a trap by extending, and I wanted to lay those hopes to rest.

Your feedback is appreciated as always : )


Maybe you can just mention firmly the capture of the 2 stones, charge on them to check why… Well not a big deal whatever, nice job.

Great post!

I do have a comment, though.

[to approach a corner is] literally the 2nd most important priority in the Opening - it’s what both players should be thinking about as soon as all 4 corners are claimed.

This statement is a bit misleading. Making a corner enclosure has the same value as approaching an unenclosed corner – you can consider an enclosure to be the prevention of an approach, and an approach to be the prevention of an enclosure.


eg. in this diagram, I like to play none of A – D, and instead P17. This is also an opening system that World #1 Shin Jinseo played recently, and the most professionally played move after A.

Enclosing the top right with P17 or O17 achieves one purpose of A, which is to play first on the widest side, but also takes points in the corner. If Black P17, White A, Black is happy to continue by enclosing the lower right or approaching at C.


This made me rethink many moves, in retrospect. :flushed:

The approach of “knowing the joseki helps you tell when it’s time to tenuki” helps me understand joseki a little bit better now.


Thanks - this type of feedback is very helpful.

Whenever I write these articles, I always try to remember how I viewed Go as a 20-25kyu, and think “what would have helped me get over that hurdle and understand it better?”

Comments like this help me feel like I’m on the right track - thanks : )