My apologies to the OP if my explanation was confusing. Apparently I have a slightly different perspective on this. I remember when I was learning the game, finally understanding the difference between 3rd and 4th line was really eye-opening. For me personally, it was the most important thing for reaching 10kyu.
jts explained what I was thinking of really well. His 2nd diagram in particular is something I find very important. When you understand it, and see why the right side is wrong and the left side correct, then you'll understand why you put stones on the 3rd line and why you put them on the 4th.
Opening Theory Made Easy
-
billywoods
- Lives in gote
- Posts: 460
- Joined: Sun Jul 29, 2012 1:12 pm
- Rank: 3 kyu
- GD Posts: 0
- Universal go server handle: billywoods
- Has thanked: 149 times
- Been thanked: 101 times
Re: Opening Theory Made Easy
I also find this issue a little confusing. I have a tentative explanation, but I'm going to phrase it as a series of responses to jts's answer:
Absolutely - because the extension on the left is a 2-space extension, and on the right it's a 1-space extension. This is completely clear.
The same is also clear in this diagram, which might be more illustrative:
On the left hand side, white's shape was a little weaker - 'incomplete' - and now white has neither potential for development nor points, and black can still slide later in endgame. On the right, white doesn't care - he never wanted to develop further from that corner, and the shape was (and still is) complete. No points or development potential have been lost. Comparing to if white played first:
On the left, white develops nicely (albeit in a smallish way) and 'completes' his shape. On the right, white doesn't really develop at all, and just makes a couple of extra points on a shape that was already complete.
But I find this explanation spurious - if the sequence in question had been played in the lower left and right (and the board was not empty there, as your diagrams suggest), neither of these moves is necessary or even very big. The upper left is a little bigger than the upper right, but not by much. After all, for white to play his development/defence move is small and gote.
Well, anyway. Is this what you were getting at?
3 on the left is one of those moves that limits (and threatens to limit further) white's development, whereas such a move wouldn't do so on the right. This is clear, but all you've convinced us of is that it's better for black to play the big loose extension than the small tight one when white's shape is incomplete. If I understand correctly, cherryhill wanted to know sort of the opposite: when white's shape is complete, why is the small tight one better than the big loose one?
After all, just visually, it looks like it's making a lot more points.
And my tentative answer is: the few extra points it might make are not adequate compensation for the weakness it introduces in black's shape, whereas giving white a weakness is adequate compensation.
Firstly, how many points is it worth? In the second diagram, it's clear that both black and white are settled and strong and on the third line, so anything that happens on the left hand side is a matter of a very small number of points (maybe 10 at most); moreover, any moves there are small and slow and not likely to gain all the points there anyway. It will probably either turn neutral or split neatly in half. It's very non-urgent.
Secondly, does black actually make those points? Well, not really. White can allow that to happen and tenuki if there are big things elsewhere (if white is happy for black to spend 2 gote moves on 10-15 points and security of his group), but personally as white I'd usually play immediately as follows:
White can easily come inside, and some fighting sequence will ensue (above is the start of one potential simple continuation) in which white has a weak group (for obvious reasons), black has a weak group (because of the white group in the upper left!), white's upper left is completed (so won't be hurt by the ensuing fight), and black's lower left is incomplete (and is under attack from both sides). (Black connecting is just as bad:
And maybe something like this happens? Hardly satisfactory - black gets no points, and his lower left is still looking like it needs another move to get a solid base, and the thickness is a little useless because of white's strong-ish group on the lower side.)
Let's compare this to if white had an incomplete shape in the upper left:
Now black has no weak groups. (Well, the lower left is weak, but black has sente and can fix it immediately, then go on to attack the weak white group.)
Or here (or 2 at triangle), black's group is stronger than before because white's upper left corner is weaker - either white defends and black gets a free jump to the centre, or white doesn't defend and black can settle while taking away white's base by sliding. This is a messy fight for both sides because everything is weak, but at least it's a fight on roughly equal terms.
So, in practice, this invasion might not happen any more - white has nasty aji left behind and doesn't want black to be able to poke at it. And if white goes back to defend, black can defend the large extension, and black made a lot of points in one gote move (and one sente move), which is better than the more solid joseki, which only makes a few points in gote.
I'm unsure about all of this, and probably shouldn't be talking about go when I can't sleep, but these are my thoughts. I'm open to correction.
jts wrote:#1 The black extension on the left side is worth way more than the black extension on the right side.
Absolutely - because the extension on the left is a 2-space extension, and on the right it's a 1-space extension. This is completely clear.
The same is also clear in this diagram, which might be more illustrative:
On the left hand side, white's shape was a little weaker - 'incomplete' - and now white has neither potential for development nor points, and black can still slide later in endgame. On the right, white doesn't care - he never wanted to develop further from that corner, and the shape was (and still is) complete. No points or development potential have been lost. Comparing to if white played first:
On the left, white develops nicely (albeit in a smallish way) and 'completes' his shape. On the right, white doesn't really develop at all, and just makes a couple of extra points on a shape that was already complete.
But I find this explanation spurious - if the sequence in question had been played in the lower left and right (and the board was not empty there, as your diagrams suggest), neither of these moves is necessary or even very big. The upper left is a little bigger than the upper right, but not by much. After all, for white to play his development/defence move is small and gote.
Well, anyway. Is this what you were getting at?
jts wrote:#3 So therefore, in a position like the one on the left, the high value of the second extension should make B more eager to make the first extension, and so he should choose a joseki () that will make the first extension worth making (
).
But on the left, the low value of the second extension should make B less eager to make the first extension, so he should choose a joseki (
3 on the left is one of those moves that limits (and threatens to limit further) white's development, whereas such a move wouldn't do so on the right. This is clear, but all you've convinced us of is that it's better for black to play the big loose extension than the small tight one when white's shape is incomplete. If I understand correctly, cherryhill wanted to know sort of the opposite: when white's shape is complete, why is the small tight one better than the big loose one?
After all, just visually, it looks like it's making a lot more points.
And my tentative answer is: the few extra points it might make are not adequate compensation for the weakness it introduces in black's shape, whereas giving white a weakness is adequate compensation.
Firstly, how many points is it worth? In the second diagram, it's clear that both black and white are settled and strong and on the third line, so anything that happens on the left hand side is a matter of a very small number of points (maybe 10 at most); moreover, any moves there are small and slow and not likely to gain all the points there anyway. It will probably either turn neutral or split neatly in half. It's very non-urgent.
Secondly, does black actually make those points? Well, not really. White can allow that to happen and tenuki if there are big things elsewhere (if white is happy for black to spend 2 gote moves on 10-15 points and security of his group), but personally as white I'd usually play immediately as follows:
White can easily come inside, and some fighting sequence will ensue (above is the start of one potential simple continuation) in which white has a weak group (for obvious reasons), black has a weak group (because of the white group in the upper left!), white's upper left is completed (so won't be hurt by the ensuing fight), and black's lower left is incomplete (and is under attack from both sides). (Black connecting is just as bad:
And maybe something like this happens? Hardly satisfactory - black gets no points, and his lower left is still looking like it needs another move to get a solid base, and the thickness is a little useless because of white's strong-ish group on the lower side.)
Let's compare this to if white had an incomplete shape in the upper left:
Now black has no weak groups. (Well, the lower left is weak, but black has sente and can fix it immediately, then go on to attack the weak white group.)
Or here (or 2 at triangle), black's group is stronger than before because white's upper left corner is weaker - either white defends and black gets a free jump to the centre, or white doesn't defend and black can settle while taking away white's base by sliding. This is a messy fight for both sides because everything is weak, but at least it's a fight on roughly equal terms.
So, in practice, this invasion might not happen any more - white has nasty aji left behind and doesn't want black to be able to poke at it. And if white goes back to defend, black can defend the large extension, and black made a lot of points in one gote move (and one sente move), which is better than the more solid joseki, which only makes a few points in gote.
I'm unsure about all of this, and probably shouldn't be talking about go when I can't sleep, but these are my thoughts. I'm open to correction.
-
EdLee
- Honinbo
- Posts: 8859
- Joined: Sat Apr 24, 2010 6:49 pm
- GD Posts: 312
- Location: Santa Barbara, CA
- Has thanked: 349 times
- Been thanked: 2070 times
jts wrote:#1 The black extension on the left side is worth way more than the black extension on the right side.
It's not clear to me. On the right, B has a nice follow-up of (a).billywoods wrote:Absolutely - because the extension on the left is a 2-space extension, and on the right it's a 1-space extension. This is completely clear.
billywoods wrote:I'm unsure about all of this...
For cherryhill, just this (again).SoDesuNe wrote:Don't worry too much about this.
-
EdLee
- Honinbo
- Posts: 8859
- Joined: Sat Apr 24, 2010 6:49 pm
- GD Posts: 312
- Location: Santa Barbara, CA
- Has thanked: 349 times
- Been thanked: 2070 times
Many people (perhaps most amateurs) are mistaken about a great many things(*) like these in Go.peppernut wrote:...finally understanding the difference between 3rd and 4th line was really eye-opening.
For me personally, it was the most important thing for reaching 10kyu.
They like to think, Oh, I finally understand [insert your favorite: attack, invasion, sabaki, moyo, thickness, weakness, etc...]
It's not like that, at all.
It's more like this:
At ~10k, there is the ~10k understanding of 3rd line and 4th line.
At ~5k, there is the ~5k understanding of 3rd line and 4th line.
At ~1d, there is the ~1d understanding of 3rd line and 4th line.
At ~3d, there is the ~3d understanding of 3rd line and 4th line.
At ~6d, there is the ~6d understanding of 3rd line and 4th line.
At near pro level, there is the near-pro-level understanding of 3rd line and 4th line.
At pro level, there is the pro-level understanding of 3rd line and 4th line.
At top pro level, there is the top-pro-level understanding of 3rd line and 4th line.
There is no limit to this.
There is no such thing as "the" difference between 3rd line and 4th line.
There is also no such thing as "the final understanding" of the 3rd line and 4th line.
We hear people say "3rd line: territory; 4th line: influence" -- that is not "the" difference;
that is merely a general idea, a guideline (for beginners, perhaps)
which is not even true in many cases.
Most likely, what you experienced was a better understanding of 3rd vs. 4th line,
and this, together with many other improvements, helped you reach ~10k level,
which is nice.
(*) Stolen from Episode VI.
-
tchan001
- Gosei
- Posts: 1582
- Joined: Wed Apr 21, 2010 6:44 pm
- GD Posts: 1292
- Location: Hong Kong
- Has thanked: 54 times
- Been thanked: 534 times
- Contact:
Re: Opening Theory Made Easy
So what's EdLee's current understanding of the 3rd and 4th line?
http://tchan001.wordpress.com
A blog on Asian go books, go sightings, and interesting tidbits
Go is such a beautiful game.
A blog on Asian go books, go sightings, and interesting tidbits
Go is such a beautiful game.
-
EdLee
- Honinbo
- Posts: 8859
- Joined: Sat Apr 24, 2010 6:49 pm
- GD Posts: 312
- Location: Santa Barbara, CA
- Has thanked: 349 times
- Been thanked: 2070 times
-
peppernut
- Dies in gote
- Posts: 63
- Joined: Tue Oct 09, 2012 4:51 am
- Rank: 5k
- GD Posts: 0
- KGS: peppernut
- Has thanked: 12 times
- Been thanked: 21 times
Re:
EdLee wrote:They like to think, Oh, I finally understand [insert your favorite: attack, invasion, sabaki, moyo, thickness, weakness, etc...]
It's not like that, at all.
[snip]
Most likely, what you experienced was a better understanding of 3rd vs. 4th line,
and this, together with many other improvements, helped you reach ~10k level,
which is nice.
Definitely, I agree with you. It's a very personal thing too. What helped me isn't necessarily the same concept that's going to help someone else. Nor is the specific thing that I picked up on the final thing to understand. Nor does everyone learn things in the same order.
I know it's really personal, which is why I wanted to chime in with my own thoughts on the matter. I remember when I was starting to learn Go, I was very confused by the height of extensions. I had literally zero understanding about why one would extend on the 3rd line versus the 4th. For me, all the talk about 4th line being for development, or favoring the center, these ideas were just too abstract to apply. Why not just put everything on the 4th line, I thought. The 4th line looks bigger. Bigger frameworks mean more territory. And why would you ever try to make a smaller territory when you could make a bigger one?
My first steps, before being able to understand anything about frameworks, were to understand that if you play this:
then you have to anticipate White may attack like so:
White's stone is placed a knight's move away, one line lower, and is ready to take away your territory.
if you make a third line extension then you prevent this:
however if you extend further to the 4th line, it's bad shape
This is all I was trying to get at. I realize it's not the final word on high vs low moves, but for me, it was the start.
-
billywoods
- Lives in gote
- Posts: 460
- Joined: Sun Jul 29, 2012 1:12 pm
- Rank: 3 kyu
- GD Posts: 0
- Universal go server handle: billywoods
- Has thanked: 149 times
- Been thanked: 101 times
Re:
EdLee wrote:It's not clear to me. On the right, B has a nice follow-up of (a).
Thanks - but he has the same follow-up on the left too, both gote for black, and white ends up with much the same position. It looks more interesting than it is.
EdLee wrote:For cherryhill, just this (again).
Nonetheless, I'm sure everyone would quite like a stronger player to come along and explain some of what the author had in mind. I've done the job to the best of my ability, and would be grateful if someone could come along and patch my answer together (or refute it entirely). I think we'd all learn from it.
-
jts
- Oza
- Posts: 2662
- Joined: Sat Sep 18, 2010 4:17 pm
- Rank: kgs 6k
- GD Posts: 0
- Has thanked: 310 times
- Been thanked: 632 times
Re: Opening Theory Made Easy
@ EdLee - would you agree that breaking appreciation of the 3rd and 4th line down rank-by-rank is a little bit too much? Different people excel in different areas of the game.
@ Billywoods - I think we're approaching Cherryhill's question from different directions. The clever point made in OTME is that you can use the "add a second extension" principle in conjunction with the "finish high positions with low extensions" principle to figure out when a high, gote extension from the corner is appropriate, and when a low, sente extension is appropriate. I guess I have always taken it as a matter of faith that 6-4 and 6-3 are both legitimate extensions, and so it made perfect sense to me that these two principles provided a way to identify the one or the other as more appropriate. You seem to take the central question to be, "Given that 6-4 plus a side extension looks so much bigger on the board, how can 6-3 plus sente ever be equally efficient?" (I agree that this is a tough question to answer, but I think you could just as reasonably ask, "Given that sente is so important in the opening, how can 6-4 plus a side extension ever be justified?") Is that a fair way of stating the question you're trying to answer?
PS, to both - It's clear that (a) in the low position is no better than the corresponding move in the high position (B15), but is it even generally viable for B at all? It looks context-dependent to me. B needs a ladder, no? And even if the ladder works, W can still connect his stones and be happy with having undermined B's side position. B15, on the other hand, is just pure endgame profit.
@ Billywoods - I think we're approaching Cherryhill's question from different directions. The clever point made in OTME is that you can use the "add a second extension" principle in conjunction with the "finish high positions with low extensions" principle to figure out when a high, gote extension from the corner is appropriate, and when a low, sente extension is appropriate. I guess I have always taken it as a matter of faith that 6-4 and 6-3 are both legitimate extensions, and so it made perfect sense to me that these two principles provided a way to identify the one or the other as more appropriate. You seem to take the central question to be, "Given that 6-4 plus a side extension looks so much bigger on the board, how can 6-3 plus sente ever be equally efficient?" (I agree that this is a tough question to answer, but I think you could just as reasonably ask, "Given that sente is so important in the opening, how can 6-4 plus a side extension ever be justified?") Is that a fair way of stating the question you're trying to answer?
PS, to both - It's clear that (a) in the low position is no better than the corresponding move in the high position (B15), but is it even generally viable for B at all? It looks context-dependent to me. B needs a ladder, no? And even if the ladder works, W can still connect his stones and be happy with having undermined B's side position. B15, on the other hand, is just pure endgame profit.
-
billywoods
- Lives in gote
- Posts: 460
- Joined: Sun Jul 29, 2012 1:12 pm
- Rank: 3 kyu
- GD Posts: 0
- Universal go server handle: billywoods
- Has thanked: 149 times
- Been thanked: 101 times
Re: Opening Theory Made Easy
jts wrote:You seem to take the central question to be, "Given that 6-4 plus a side extension looks so much bigger on the board, how can 6-3 plus sente ever be equally efficient?" (I agree that this is a tough question to answer, but I think you could just as reasonably ask, "Given that sente is so important in the opening, how can 6-4 plus a side extension ever be justified?") Is that a fair way of stating the question you're trying to answer?
Yes, I think so.
On the one hand, white shouldn't consider the left area big because there's a low white stone there, so black shouldn't either, and it's good to take sente if there's something else big on the board. On the other hand, if white is playing here, white thinks the corner is the biggest area on the board; the 6-4 plus side extension does give black some extra points and an opportunity to expand into the centre later, which white doesn't really have, and will hopefully equal or outstrip whatever profit white made on the lower side; and it's fine for black to take gote if there's nothing else this big. The only reason the author can have had for saying this was outright a bad move (not depending on sente/gote), as far as I can see, is because of what I said earlier - an immediate fight would be good for white. But I'm very unsure.