It is currently Mon Nov 20, 2017 12:33 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 15 posts ] 
Author Message
Offline
 Post subject: Extension problems
Post #1 Posted: Mon Apr 10, 2017 2:28 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
One of the most popular books I have done, at least in terms of sales, was The Go Consultants. This is a commentary on a famous "Old versus New" pair-go game between Suzuki Tamejiro/Segoe Kensaku and Go Seigen/Kitani Minrou in 1934-35. The commentary was by the players themselves, designed to show how they were thinking. Readers seem to like the novelty of this.

But this style of commentary was actually the norm in those days. Almost every game in Kido in the 1920s and 1930s featured detailed comments by both players. This was also the period of the life cycle of New Fuseki - ideas that were incredibly similar to those of modern AI. It was also an era when time limits of 8 to 18 hours each offered scope for deep and accurate thought. The insights offered are extreme distillations, of course. But that makes themt malt whiskies rather than the tepid lagers of modern commentaries.

The difference seems most noticeable to me in the fuseki - natural enough, perhaps, given the innovation in the air. I thought it might be of interest to present fuseki problems from these games where an extension was possible. Your task is not so much to guess the extension played (quite often there are valid alternatives) but to get clear in your mind the factors that need to be considered in making extensions. Then see what the pros offer as a distillation of their in-game thoughts. Over time, this series of problems (if popular enough) might lead to kyus especially developing a rather more refined policy network.

I also hope to shed some light on what lies behind some of the words we use.



This no-komi game featured Kato Minaichi versus Takenaka Kotaro in 1936. Kato had just played the triangled stone but confessed himself unsure whether to play there or A. Now, though, Black will naturally play on the upper side. Where and WHY?



Takenaka chose the triangled point. He first mentioned the approach move at A but then White could pincer at B or C. On the face of it, he seems to be saying nothing about why he rejected it, but there's an interesting translation subtlety. English students are (or were, in my day) repeatedly advised to eschew the passive tense in writing. It is therefore common for a translator to render the Japanese here (literally "I would be pincered") as "White would pincer." This changes the subject and the emphasis - massively, I think. The passive is also usually considered "adversative" in Japanese. The upshot, anyway, is that Black is implying "I don't want to, or must not, be pincered."

Next, Takenaka mentions B - "one space wider." He concedes this is plausible, but then adds: "It would lead to an ikken-tobi game," with Black jumping to D.

I confess that I am not entirely sure why he seems to be implying that an ikken-tobi game is not to his taste, but there is, I think, an important clue in the word "wide." The Japanese word (hiroi) has almost a technical status and is very common. It could be translated often as, say, "open" but "wide" extensions are an important aspect of open play. In those no-komi days, it could be taken for granted that everyone knew that Black tended to shy away from open/wide play so as to preserve his advantage of first move. I'm certain that applies here, but would be interested to hear opinions on how this relates to ikken tobis.

But Takenaka goes on to explain that he decided his move was superior to B because White will eventually play the shoulder hit at E, and the thickness he will accrue there will make invasion at the top troublesome for Black. I think the interesting aspects are that he is thinking of the reducing move as thickness and that he is expecting the shoulder hit - in other words he is not thinking of this area as Black territory or with any expectation of getting the ikken tobi to build up his framework. On the other hand, he appears not to be thinking of any follow-up moves - he is treating this as an upper-right quadrant only problem (and the course of the game seems to validate that - perhaps we can see this as Black laying down a spider's web, luring White in).

Note also that, while he evidently expected the shoulder hit on the right side, he strongly implied that, had he extended wider on the upper side, he would get the ikken tobi there, i.e. White would not shoulder hit there. Interesting to consider why - the direction of the thickness seems to hold the answer.


White's reply to the hidden move chosen above also involved an interesting extension, with reasoning that I find very rare in amateur circles:



I would expect many amateurs to play the triangled stone instead of A, but for reasons such as "Keep away from thickness" or "Else Black could invade." But Katos' reasoning was that A would give Black perfect momentum to jump to B. Momentum (choshi) is a major strategic concept mostly absent from amateur play. Incidentally, neither White's move nor A is strictly an extension (hiraki), but a tsume, often translated as "checking extension." So this term alone tells us a lot about what was going on in White's mind (i.e. he was blocking Black's development, this being the main purpose of a tsume, so going to A and encouraging Black to develop at B must be a no-no).


Last edited by John Fairbairn on Mon Apr 10, 2017 9:11 am, edited 1 time in total.

This post by John Fairbairn was liked by 5 people: Bill Spight, BlindGroup, gowan, joellercoaster, Waylon
Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #2 Posted: Mon Apr 10, 2017 7:37 am 
Judan

Posts: 6482
Liked others: 1518
Was liked: 2414
FWIW, for Black I prefer C, precisely because of its whole board effects. That said, I have had success with plays like the one in the game, an idea that I picked up from Segoe's Sakusen Jiten. :)

As for White's extension in the second diagram, it is fairly normal from the large knight's enclosure. If White extends one more space, Black's jump becomes a dual purpose play, enlarging the Black framework while aiming at White's thinness.

_________________
Don't cry for me, Sergeant Tina.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #3 Posted: Mon Apr 10, 2017 8:41 am 
Lives with ko

Posts: 169
Liked others: 108
Was liked: 24
KGS: 5k
IGS: 8k
Universal go server handle: BlindGroup
John Fairbairn wrote:
I thought it might be of interest to present fuseki problems from these games where an extension was possible. Your task is not so much to guess the extension played (quite often there are valid alternatives) but to get clear in your mind the factors that need to be considered in making extensions.


Great idea! Thank you for doing this.

John Fairbairn wrote:
I confess that I am not entirely sure why he seems to be implying that an ikken-tobi game is not to his taste, but there is, I think, an important clue in the word "wide."


Could someone explain what an ikken-tobi game is? I under stand the term as the one-space jump, but don't know what it means in terms of a type of game.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #4 Posted: Mon Apr 10, 2017 11:26 am 
Tengen

Posts: 4235
Location: Cambridge, UK
Liked others: 159
Was liked: 2083
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Funnily enough I played a restrained extension like the second one here 2 days ago at the British Open against Andrew Kay (though it was in middlegame and the side group much thicker so even more logical), but as I was overcome with fighting spirit and stupidity in overtime and chose to fight a ko rather than connect a dragon of about 30 stones in the endgame the subtleties of such extensions were rendered rather moot!

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #5 Posted: Tue Apr 11, 2017 3:20 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
Here is the next example (these positions are cropping up as I'm transcribing 1930s games). This one is between top players Kubomatsu Katsukiyo and Miyasaka Shinji. Miyasaka is the chap expected - at least by himself - to succeed Shusai as Honinbo.



Obviously Black to play somewhere on the lower side. But where and what factors come into it?



You will note that Black has adopted a slightly unusual but still standard line of the joseki on the left which emphasises solidity. So that is one factor. Kubomatsu said he chose the triangled point because he intended to continue on the left side as shown in the game - see below. Also, he rejected Black A because that gives White too good a move at B.

But maybe the most interesting thing is the word he used to described his play. It was not an "extension." It was a yose move. Yes the yose (if translated properly as "boundary play") can begin as early as move 15 and in centre field.

This usage is a little unusual, though certainly not rare. Kakoi (surrounding could be used instead). But either way the term gives us an insight into Kubomatsu's thinking in this game - he was going for early territory, hence an early boundary play.

The position after the next few moves will illustrate further what he planned:



Black's play on the left side could also be called yose rather than thickness building. As I've said before, thickness is really thickness only if it can function as thickness, and in this game there was no call for that. In fact this game could be called one long yose sequence - see the final position.


It is possible to argue that Kubomatsu was being wise after the event in using the term yose, but given the plethora of articles (and a famous book: Tengen and New Fuseki) by him, and many other games, in which he emphasised the centre and discussed the problems associated with sealing it off, I think we can safely assume the "Great Amateur" was an AlphaGo trailblazer of the 1930s. He won by 6.


This post by John Fairbairn was liked by 2 people: Bill Spight, BlindGroup
Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #6 Posted: Tue Apr 11, 2017 8:45 am 
Judan

Posts: 6482
Liked others: 1518
Was liked: 2414
Kubomatsu's (Black's) play is kind of obvious, especially with no komi. But I wonder: With 7-ish komi, would Master play at B? :)

_________________
Don't cry for me, Sergeant Tina.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #7 Posted: Tue Apr 11, 2017 8:57 am 
Judan

Posts: 6482
Liked others: 1518
Was liked: 2414
John Fairbairn wrote:
Final position.


I think there is a typo somewhere. Can't White live (or Black kill) on the right side?

_________________
Don't cry for me, Sergeant Tina.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #8 Posted: Wed Apr 12, 2017 2:17 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
Bill: You are right - white stone missing at S14.

Today's example is rather simple but a profound truth lies behind it. Iwamoto to play on the upper side, against Maeda Nobuaki.



In passing, how does White capture the three black stones in the lower right?



Iwamoto played at the triangled point and commented that this sort of situation of having to decide between this move and a jump to A arises often, but it is never clear which is better. I think it is important to remember that fuseki is very hard even for pros, even in cases that allegedly come up often.

However, there are not too many database examples of this particular upper-board shape. In most cases, Black does indeed play on the middle of the upper side, though not always at the precise triangled point chosen by Iwamoto. But I can find no example of the jump. The same situation obtains even if we change Black's shimari to a small one.

That is not to say, of course, that Black does not give the jump very serious consideration. Assuming a genuine decision is being made, therefore, as in Iwamoto's case, what sort of factors go into the mix? In Iwamoto's case he did not explain directly (nor, unfortunately, did he comment on high or low), but he did make a comment about his previous play which probably indicates one thing that swayed him: he said that the result in the lower right was a little overconcentrated for Black, whereas White was rather thick.

Maeda commented that had Black played the jump to A, he would have responded with White at the triangled point. But Iwamoto said to that that he would just play B, which seems to hint at a little regret on his part. He won anyway, without counting.

I confess to missing the best capture in the lower right - or, more precisely, I was too lazy even to look: White plays S1, not the P1 I on auto-pilot would have done.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #9 Posted: Wed Apr 12, 2017 3:04 am 
Judan

Posts: 6482
Liked others: 1518
Was liked: 2414
Didin't Karigane play that jump against Go Seigen?

_________________
Don't cry for me, Sergeant Tina.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #10 Posted: Wed Apr 12, 2017 3:43 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
Quote:
Didin't Karigane play that jump against Go Seigen?


Yes, but not when he had a choice between that an extension (Game 1 of their match I assume you mean).

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #11 Posted: Wed Apr 12, 2017 7:26 am 
Judan

Posts: 6482
Liked others: 1518
Was liked: 2414
John Fairbairn wrote:
Quote:
Didin't Karigane play that jump against Go Seigen?


Yes, but not when he had a choice between that an extension (Game 1 of their match I assume you mean).


Right. I looked it up after getting up just now. He had the option of extending on the other side. :)

_________________
Don't cry for me, Sergeant Tina.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #12 Posted: Wed Apr 12, 2017 9:24 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
So that others know what we are talking about, here is the position:



In fact Go said Karigane's triangled move was dubious. He assumed Karigane played it because it married well with the X stone and/or because he was worried about White playing there as an erasure. He shot down the first reason by arguing that the space was in fact too wide, and the second by claiming that Black would answer the cap at A, and that was the ideal relationship with the X stone.

But he didn't say quite that the right alternative was extension on the right side. Rather he suggested Black should first exchange B for C before plonking down at D. This may be a stylistic thing. It seems equally valid to me to play D first and then aim for B. The rights and wrongs of this order-of-play issue is something that I don't thin I've ever seen discussed in fuseki books.


This post by John Fairbairn was liked by: Bill Spight
Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #13 Posted: Wed Apr 12, 2017 9:53 am 
Tengen

Posts: 4235
Location: Cambridge, UK
Liked others: 159
Was liked: 2083
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
If I would play on the top side, I would also play j16. That kind of high and far extension (one further than star point) is one I've tended to play more in recent years. I like that it makes white's checking extension smaller, and being on the 4th line has more effect reducing the influence of white's high shimari (if 3rd line then shoulder hits or caps or knight's moves can develop the centre to left for white later). And s1 on the lower side is easy to see when asked but I too might miss it in a game.

But the first thing I looked at on this board was black's lower left group which I'd assume white attacked at c8 recently. My first feeling would be to see it as a weak group and do something with it: jumping out to f8 is a standard move in handicap games but is rather ponderous (a bit slow but trying to make miai of pincers on left or lower side, but despite the h3 gap white's pretty strong there). So then I would think about g4 to get out and would white answer at h4 (if so pincer at d11 next?) and would that be a good exchange for black (I feel probably, you lose an invasion, but that invasion is unlikely)? j4 is another idea but probably overplay given white's thickness at m4. Black's not going to die if he tenukis, so if black did something like f6 which has no spectacular follow-up I'd worry about white playing a big point on the top side (extend from his shimari and start a double-wing, stop my shimari extending and making double wing), so perhaps the flow of the game is white just played c8 and black cannot afford to make any local answer as it's basically miai with the top: if c8 was the last move and instead white played at k16 (for example) then black would play on the left side (c10 is a local joseki but feels a bit timid, so c13, or something more advanced?).

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #14 Posted: Wed Apr 12, 2017 11:15 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
uberdude: The flow of the game was that Black let White attack his left corner group in a version of the one-weak-group strategy. It lived easily enough in a minor skirmish in this quadrant alone (i.e. no collateral damage), and Black also managed to live in the upper left. That left White no option but to try kamikaze tactics on the side, but he never got beyond move 125. There was np specific commentary by Iwamoto on this skirmish, but Maeda thought he ought to have deferred attacking there to play on the right side. He probably did not play there because he had already become dissatisfied with his cap on the shimari (maybe because Black answered - though with reservations - at O16) and presumably didn't want to throw good money after bad. He decided he ought to have invaded at L16 instead of the cap, though it has to be said that Iwamoto gave that rather a gruff brush-off.

In a sense this isn't a fuseki problem but rather a start-of-the-middle game problem, so the dynamics have changed a bit.

Top
 Profile  
 
Offline
 Post subject: Re: Extension problems
Post #15 Posted: Thu Apr 13, 2017 2:23 am 
Gosei

Posts: 1901
Liked others: 12
Was liked: 2814
A further aspect today - extensions are often not just extensions but may double up as checking extensions (tsume) or, as here, extensions-cum-pincers. That brings in a dose of tactics, because of the joseki implications.



White (Fukuda Masayoshi) has just played at the triangled point. He explained that he has played this sort of move for a long time (so it's obviously stylistic) but he was not sure whether it was better instead to play at A or B, let Black make the shimari at the triangled point, and then shimari himself at C.

There is nothing magical about choosing the next move for Black, but see if you can replicate some of the thought process of Mukai Kazuo.



Mukai said: "There was a strategy behind Black 27 (triangle). If White continues by attaching at the 3-3 point A in the usual way, Black would normally hane inside at B, and then after White C, Black D a furikawari trade would ensue, but in that case Black 27 would be better at E [not explained, but why?]."

He went on: "But the reason I played 27 was that if White did play at the 3-3 point A, then I thought I would try simply playing the hane on the outside at F and making the hanging connection after White B, Black G."

He did not explain why he thought that might be better, but two factors seem to be relevant. One is that Black appeared to have his eye on playing H in the lower left (as happened in the game) rather than the usual I. He said I is bad here because White will attach at H. Also, the open skirt in the upper left will govern Black's assessment of what he can do in the lower left.

White appears to have sussed out Black's thinking and so continued with the attachment at E. This has now become known amongst pros in the last couple of decades, but was a novelty then and remain apparently uncopied until 1996. Fukuda's assessment of it was that it seemed a little heavy. The result of the "joseki" is shown below - more shades of AlphagGo (but White lost here).


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 15 posts ] 

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group