Teaching beginners oldschool or new AI joseki?

schrody · #1

This is something I've been wondering for a while now. Back in the day, everything was much simpler. The first two joseki a beginner learned were usually the oldschool 3-3 invasion and the keima approach to a 4-4 stone followed by a slide. Now these two joseki are considered subpar and the new AI-approved variations are significantly more complex, so what do we do? I'm still leaning towards teaching the older variations because they are simpler and, for the most part, don't branch out as much as the newer ones. They also have fewer follow-ups. Still, I can't quite shake off that nagging feeling that I'm teaching people wrong things.

Bill Spight · #2

schrody wrote:

This is something I've been wondering for a while now. Back in the day, everything was much simpler. The first two joseki a beginner learned were usually the oldschool 3-3 invasion and the keima approach to a 4-4 stone followed by a slide. Now these two joseki are considered subpar and the new AI-approved variations are significantly more complex, so what do we do? I'm still leaning towards teaching the older variations because they are simpler and, for the most part, don't branch out as much as the newer ones. They also have fewer follow-ups. Still, I can't quite shake off that nagging feeling that I'm teaching people wrong things.

Why teach beginners joseki at all? Do we teach babies to walk with crutches? I probably waited too late to study joseki, because of the proverb. I started at 4 kyu, but I didn't really study joseki until I was a dan player. Joseki narrow your vision, something that can hamper SDKs. OC, I picked up some joseki along the way, which happens naturally.

Moi, I would rather teach basic tesuji, basic life and death, and how to fill dame. Up to 10 kyu or so, beginners will often have the chance to turn the tables at the dame filling stage. Too often beginners learn not to fill the dame, and that it is impolite to invade at the end of the game.

OC, beginners should learn how to review their own games with AI.

jlt · #3

The old school josekis have been played by pros until recently, so they can't be really bad.

Boidhre · #4

schrody wrote:

This is something I've been wondering for a while now. Back in the day, everything was much simpler. The first two joseki a beginner learned were usually the oldschool 3-3 invasion and the keima approach to a 4-4 stone followed by a slide. Now these two joseki are considered subpar and the new AI-approved variations are significantly more complex, so what do we do? I'm still leaning towards teaching the older variations because they are simpler and, for the most part, don't branch out as much as the newer ones. They also have fewer follow-ups. Still, I can't quite shake off that nagging feeling that I'm teaching people wrong things.

I think the same logic applies as before, well I'm kinda interested in this and open to being wrong here. You teach them what they can understand and work with. The old 4-4 slide joseki wasn't really about teaching someone the joseki but the idea of making a base, splitting the corner, threatening the various follow ups etc. The meaning behind the moves was what you were teaching rather than the joseki itself per se. If you show a joseki to someone it's because it has some useful general concept to impart in it rather than this is something they should memorise.

E.g. When we put this on the board is our intention to get them to memorise the joseki or understand why White plays a? Memorising joseki is of dubious benefit for beginners but playing b instead of a here and asking them how they should respond is probably useful (ok maybe not at 20k but you get what I mean).

Click Here To Show Diagram Code: [go]$$B $$ . . . . . . . | $$ . . . . . . . | $$ . . . . b X . | $$ . . . . X O . | $$ . . . . . a . | $$ . . . . O . . | $$ . . . . . . . | $$ . . . . . . . | $$ -------------[/go]

If you can find some good teachable sequences in the new AI joseki you use them. Otherwise you don't introduce them to beginners for the same reason you didn't more complicated josekis before, the student won't be able to get much that is useful out of it.

schrody · #5

I agree with Boidhre. I always teach them the rationale behind the moves in the hopes that they'll be able to apply such moves/rationale in other situations as well. The other reason is that at around 20k, the players usually find the 19X19 board overwhelmingly large. Providing them with a few set patterns helps them narrow down the board and makes them more confident.

I'm certainly not a fan of memorizing complex joseki and I try to make sure that the ones I choose to teach are appropriate for the strength of the player. For example, at 20k+ I often just introduce the idea of pincers and their function but don't really teach any specific joseki.

Bill Spight wrote:

Joseki narrow your vision, something that can hamper SDKs.

Agreed to an extent. As a SDK I often played a joseki variation just because it was the only one I knew, even if it didn't fit the whole-board situation. However, I'm not sure if ignoring studying joseki altogether is the right solution. I'd be more in favour of helping players focus more on whole-board thinking and encouraging them to explore a variety of different moves.

Bill Spight wrote:

Moi, I would rather teach basic tesuji, basic life and death, and how to fill dame. Up to 10 kyu or so, beginners will often have the chance to turn the tables at the dame filling stage. Too often beginners learn not to fill the dame, and that it is impolite to invade at the end of the game.

I agree that there's other important parts of the game that also need to be taught, including how to fill dame.

I find that in person pretty much everyone gets taught that, but that's often not the case for players who only play online.

Bill Spight · #6

If I were teaching the following joseki to DDKS, I would stop after :b3:

.

Why? Take a look. There are (at least) 15 possible followups, including tenuki, each of which probably falls within the margin of error, certainly the margin of error for a DDK, any of which might be a good play in a particular whole board situation. This is a good point to allow the beginner's imagination to operate, and to let her start to develop her judgement.

Edit: I don't mean that I would restrict replies to the keima, either. As always, the personal factor is very important.

bogiesan · #7

Teach them the new stuff; it will be old in a few months.

MikeKyle · #8

Re 3-3: I really like this pattern as a starting point:

Click Here To Show Diagram Code: [go]$$B $$ . . . . . . . | $$ . . . . . . . | $$ . . . . . . . | $$ . . . . . . . | $$ . . 5 1 3 . . | $$ . . . 4 2 . . | $$ . 6 . . . . . | $$ . . . . . . . | $$ -------------[/go]

Has a lot of interesting talking points around the possible follow ups, why the wall isn't so weak as you might think, the value of sente, pushing from behind/getting ahead..
I wouldn't go into much detail beyond this pattern unless it came up naturally.
If 5 is a hane, I would never say that that's wrong in a teaching game. I would say that it's a choice you can make and it has a lot going for it but it also leaves behind a small weakness in the shape of the wall that you should be aware of.
You get something but you also pay something - I think I say that a lot when discussing patterns with ddks.

gennan · #9

I think the apparent simplicity of that variation is quite deceptive. It has so many subtleties, post-joseki tactics and large scale considerations. I feel it's not really suitable for weaker players.
So until about 15k, I usually focus on the double hane variation. It is easy to explain and understand in the local context, because it is pretty much a one-way street and it leads to both colors ending up with a settled group in a fairly clear and mostly even local result.

gowan · **#10**

After how to capture and how to live, I think I would tell a real beginner about spreading out on the board and staying connected. It is painful to watch a real beginner playing nobi-nobi when the board is mostly empty. We know the saying that there is no joseki for a meijin and maybe we could amend that to include there is no joseki for beginners

gennan · **#11**

The beginner range is quite large and diverse. I don't think one should teach joseki to novices at all.

I think there are several subcategories in the beginner category that shouldn't be conflated:

1. novice (45k - 35k): playing on 9x9, learning the rules, surrounding territory, connecting & cutting, 2 eyes, scoring the game
2. beginner (35k - 25k): playing on 13x13, defending territory, basic tactics: double atari, snap-back, net, ladder, false eyes, simple semeai, throw-in, nakade, seki
3. advanced beginner (25k - 15k): playing on 19x19, more tactics (tesuji), some strategy, mapping out territorial frameworks, attacking and defending groups, invading, heangma, simple common joseki

I don't think one should try to teach topics for group 3 to group 1.

Knotwilg · **#12**

Here's a conversation between Jose Kiteacher, an amateur dan teaching at a club and Davy L'Sadvocate, a beginner with a sharp critical mind.

Jose: "The most common opening move is on the star point. Black plays here, on 4-4".
Davy: "What about playing deeper in the corner: on that ... 3-3 point"
Jose: "That's considered a little slow."
(Davy's critical mind is activated, taking no BS from anyone)
Jose: "The most common next move in that corner is White's 3-3 invasion"
(Jose shows modern joseki)
Davy: "So why is that a good move for White?"
Jose: "Because White gets territory, and Black influence, which is a little harder to play with. And White can do that and be the first to play in another area of the board."
Davy: "So if Black plays on 3-3 to start with, White can't go there anymore, to get this good result?"
Jose: "Err ... yes."
Davy: "So what happens if Black plays 3-3? What's White's answer?"
Jose: "Eh ... White can shoulder hit, for example" (shows traditional joseki)
Davy: "How can that be good? Black can ignore it and play in another area of the board, and still have that good result you just showed me?"
Jose: "Eh ..."

We can't confidently teach traditional or modern joseki to beginners. We hardly understand today's basic patterns ourselves, and since they've been evaluated as superior to old patterns, we didn't really understand those either.

gennan · **#13**

From that conversation, I would guess that Davy is an advanced beginner, but that would depend on the number of games they have played.
Davy could also be an overteached novice, being exposed to more theory than they can really grasp and collapsing to any aggressive beginner they encounter in an actual game.

Boidhre · **#14**

I can't really fathom using a term like "slow" with an opening move for a beginner. There's just *so* much in that single word that presumes so much more familiarity with the game than someone would have who is being showing the 4-4 move for the first time in a lesson.

John Fairbairn · **#15**

Quote:

We hardly understand today's basic patterns ourselves, and since they've been evaluated as superior to old patterns, we didn't really understand those either.

Go is usually touted as a game of pattern recognition. But over recent years I've become increasingly convinced that it's not. We play it as a game of pattern recognition because we superimpose patterns on it, and then of course we recognise those patterns. How very circular.

This is especially fatuous in the case of shapes/patterns like horse's neck. What possible use can a scene from the Godfather have on a go board? And it all applies as much to patterns that don't have a name.

It's all just pareidolia.

I've been struck by the fact that the old Chinese masters, who wrote extensively about go, had virtually zero named patterns (not even bamboo joints), and those that I can think of are just names of opening formations - just labels as opposed to go wisdom.

The old Japanese masters wrote very little, so we don't know how many pattern terms they really had, but it's pound to a penny that virtually every pattern term came from the 1920s when the Nihon Ki-in started publishing books for beginners (of whatever stamp), and so the game was dumbed down.

AI is a wake-up call. We need to dumb back up.

lightvector · **#16**

Knotwilg wrote:

Here's a conversation between Jose Kiteacher, an amateur dan teaching at a club and Davy L'Sadvocate, a beginner with a sharp critical mind.

Jose: "The most common opening move is on the star point. Black plays here, on 4-4".
Davy: "What about playing deeper in the corner: on that ... 3-3 point"
Jose: "That's considered a little slow."
(Davy's critical mind is activated, taking no BS from anyone)
Jose: "The most common next move in that corner is White's 3-3 invasion"
(Jose shows modern joseki)
Davy: "So why is that a good move for White?"
Jose: "Because White gets territory, and Black influence, which is a little harder to play with. And White can do that and be the first to play in another area of the board."
Davy: "So if Black plays on 3-3 to start with, White can't go there anymore, to get this good result?"
Jose: "Err ... yes."
Davy: "So what happens if Black plays 3-3? What's White's answer?"
Jose: "Eh ... White can shoulder hit, for example" (shows traditional joseki)
Davy: "How can that be good? Black can ignore it and play in another area of the board, and still have that good result you just showed me?"
Jose: "Eh ..."

We can't confidently teach traditional or modern joseki to beginners. We hardly understand today's basic patterns ourselves, and since they've been evaluated as superior to old patterns, we didn't really understand those either.

For how it can simultaneously be bad to ignore 4-4 shoulder hit a 3-3, yet also be good to 3-3 invade under a 4-4, well it's not surprising if the person who plays first in the corner has a slight edge in that corner.

* 3-3 invade under a 4-4 is a local slight advantage for the 4-4, but still a good move because it's only an advantage for the 4-4 by about the amount that the 4-4 expected anyways due to taking the corner first.
* 4-4 shoulder hit on a 3-3 is local slight advantage for the 3-3, but can be still a good move because it's only an advantage for the 3-3 by about what the 3-3 expected anyways due to taking the corner first. (And in this case IIRC it's not automatically a good move, you want to delay it until it reduces a moyo or has some other purpose).
* Ignoring 4-4 shoulder hit on a 3-3 is often a loss since you flip from a 3-3-advantaged result to a 4-4-advantaged result, or said another way, you effectively flip from getting a corner-first-move-advantage level of result to getting the opposite.

Having useful knowledge about something and being able to apply it accurately oneself is a fairly distinct skill from being able to verbalize that knowledge, which is itself a fairly distinct skill from being able to introspect and then verbalize the reasons and principles behind that knowledge to justify it. (*)

As an aside, something that annoys me about some "take no BS" people is that some of them occasionally seem to end up in a mode where their goal is to outlogic and prove the other person wrong, rather than having a conversation in good faith. But the ones who are aware of the above (*) and who do consistently go out of their way to repeatedly reinforce and signal good faith even as they challenge, or even to turn their critical thinking towards finding ways your own knowledge can be made justified and consistent rather than only poking holes in it - can be a pleasure to teach or talk with. Those are the ones who can help you even clarify your own understanding, the kind of person who at might at minimum say "What you've said doesn't make any sense to me, but if you're good at the game I can accept that these are probably all still great patterns to get an initial feel for what good moves look like and maybe why they're good is something that's not easy to verbally explain."

Boidhre · **#17**

lightvector wrote:

As an aside, something that annoys me about some "take no BS" people is that some of them occasionally seem to end up in a mode where their goal is to outlogic and prove the other person wrong, rather than having a conversation in good faith. But the ones who are aware of the above (*) and who do consistently go out of their way to repeatedly reinforce and signal good faith even as they challenge, or even to turn their critical thinking towards finding ways your own knowledge can be made justified and consistent rather than only poking holes in it - can be a pleasure to teach or talk with. Those are the ones who can help you even clarify your own understanding, the kind of person who at might at minimum say "What you've said doesn't make any sense to me, but if you're good at the game I can accept that these are probably all still great patterns to get an initial feel for what good moves look like and maybe why they're good is something that's not easy to verbally explain."

For me the issue very often with this (and beginners/casual players/whatever) is that they're trying to get you to explain to them how to run when they don't know how to walk yet. So you cannot ever really give them the logical explanation they're demanding as they lack the experience or knowledge to understand such and these are not things you can easily or quickly impart and are they are also often blind to the need for this experience or knowledge.

Knotwilg · **#18**

In the conversation above, Davy is asking genuine questions. I object to dismissing Davy (it is I, Dieter) as an obnoxious person who refuses to learn and only wants to outlogic the benevolent teacher. I equally object to Jose being portrayed as someone who knows it all and is teaching Davy to walk before he can run. That's the old paradigm of teaching Go, which I'm arguing AI should have humbled us out of. Jose is better at running than Davy but perhaps not because of his deep understanding of how the knee works.

The logic holds if we tewari that the slower opening at 3-3 and the slower shoulder hit at 4-4 are canceling mistakes, reverting to a fast 4-4 opening and a good 3-3 invasion. That's a simple logic but not a trivial one. It needs whole board thinking, rather than local analysis.

This understanding, which is the (AI reinforced) idea that it is recommended to switch to an undisputed corner rather soon than late, is one that I would tutor earlier than any particular joseki.

Now to lightvector's answer and the tewari argument: consider these two openings

Click Here To Show Diagram Code: [go]$$B Opening 1 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 2 . . . . . , . . . . . 1 . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 4 . . . . . , . . . . . 3 . . . | $$ | . . 5 . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$B Opening 2 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 2 . . . . . , . . . . . 1 . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . 5 . . . . . , . . . . . 3 . . . | $$ | . . 4 . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Opening 2 is as if White played opening 1 as Black, then flipped stone color, except for one stone, took komi and kept sente. Flipping one stone = 2 moves. Komi and sente are one move, so the net result is equal.

For me it is very hard, from these diagrams, to argue that opening two has two small mistakes, which cancel each other out, rather than being two good moves. I know that :b5:

in this diagram is expected to be a 3-3 in the other white corner. But why, I can't tell.

gennan · **#19**

I think the problem is that you can't evaluate all positions statically (spatial aspect) without information about whose turn it is (temporal aspect).

In this example it is very urgent to play locally, so tempo matters greatly for the evaluation of the position:

Click Here To Show Diagram Code: [go]$$ whose turn is it? $$ . . . . . . . . . $$ . . . . . . . . . $$ . . . . O . . . . $$ . . . O X . . . . $$ . . . . O X . . . $$ . . . . X . . . . $$ . . . . . . . . . $$ . . . . . . . . .[/go]

Click Here To Show Diagram Code: [go]$$B black's turn -> great for black $$ . . . . . . . . . $$ . . . . . . . . . $$ . . . . O . . . . $$ . . . O X . . . . $$ . . . 1 O X . . . $$ . . . . X . . . . $$ . . . . . . . . . $$ . . . . . . . . .[/go]

Click Here To Show Diagram Code: [go]$$W white's turn -> great for white $$ . . . . . . . . . $$ . . . . . . . . . $$ . . . . O . . . . $$ . . . O X 1 . . . $$ . . . . O X . . . $$ . . . . X . . . . $$ . . . . . . . . . $$ . . . . . . . . .[/go]

In this example it is not urgent to play locally (capture 1 stone), so tempo doesn't matter much. Tenuki is best for either color in most cases, so this shape can be statically evaluated as even:

Click Here To Show Diagram Code: [go]$$ whose turn is it? $$ . . . . . . . . . $$ . . . . . . . . . $$ . . . O O O X . . $$ . . . O X . X . . $$ . . O . O X . . . $$ . . O X X X . . . $$ . . . . . . . . . $$ . . . . . . . . .[/go]

Another example where it is not very urgent to play locally.
Black a is a good local move. It gives white a secure group with a decent amount of territory, but black settles his group while making some influence.
White a is a good local move, but it may be a bit slow globally, because black can easily ignore it and tenuki (black is satisfied with preventing a shimari in sente).
So tenuki is an option for either color in many cases and the position can be statically evaluated as about even:

Click Here To Show Diagram Code: [go]$$ whose turn is it? $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . a . . . . . $$ | . . O . . . . . . $$ | . . . . X . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ -------------------[/go]

Click Here To Show Diagram Code: [go]$$B black's turn -> even $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . 4 . . . . . . $$ | . . . 3 . . . . . $$ | . . 2 1 . . . . . $$ | . . O . . . . . . $$ | . . . . X . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ -------------------[/go]

Click Here To Show Diagram Code: [go]$$W white's turn -> even $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . 1 . . . . . $$ | . . O . . . . . . $$ | . . . . X . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ -------------------[/go]

tenuki

In the 3-3/4-4 situation it is quite urgent to play locally, so tempo matters and the shape cannot be evaluated statically:

Click Here To Show Diagram Code: [go]$$ whose turn is it? $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . X . . . . . $$ | . . O . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ -------------------[/go]

Click Here To Show Diagram Code: [go]$$B black's turn -> even $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . 1 X 3 . . . . $$ | . . O 2 . . . . . $$ | . . . . . 4 . . . $$ | . . . . . . . . . $$ -------------------[/go]

Click Here To Show Diagram Code: [go]$$W white's turn -> good for white $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . 4 2 . . . . . $$ | . . 1 X . . . . . $$ | . . O . . . . . . $$ | . . . . 3 . . . . $$ | . . . . . . . . . $$ -------------------[/go]

jlt · **#20**

Both openings 1 and 2 have been played by pros, so Davy really shouldn't worry about whether starting with a 3-3 or a 4-4 is better. If Davy is really a beginner, I would expect him to consider moves like

a or b: "Jose, you told me it's easier to make territory near the corners than on the sides or in the center, so I am making territory".

c or d: "Jose, you told me to play where the space is wide open, so I followed your advice".

Click Here To Show Diagram Code: [go]$$Bc $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . a . . . | $$ | . . . . . . . . . . . . . O . b . . . | $$ | . . . O . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . c . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . d . . . | $$ | . . . . . . . . . . . . . . . . . . . |[/go]

I've seen moves like a or b in Fox 5k games, so learning josekis is really not a priority for a beginner. On the other hand, josekis are examples of good sequences which can be learnt from, so there are pros and cons about learning josekis as a beginner.

Teaching beginners oldschool or new AI joseki?

Who is online