One of the things that has given me most joy over the years is listening to a real expert who can talk to an audience of laymen and make his subject sound to them as if it was totally obvious, or something that they felt they had known all along. The reality, of course, is that what they have just learnt was neither obvious nor known to them (but that won't stop some of them from believing they know almost as much as the expert all along!).
We find the same phenomenon in go, but with the rider that we don't often get to hear real experts with that ability to talk "down" to us. Pros are usually too busy earning their shekels to take time out for us.
We are lucky to have one who does make time for us. Shibano Toramiru is of course a real expert. Just turned 20, he has won two major titles (Meijin and Oza) and, as I write (March 2020), has started successfully on is way to winning a third (the Judan). Despite that stressful and frenetic schedule, over that past year and a bit he has written a still on-going weekly column for the Go Weekly newspaper. You can tell easily that he wrote it himself. It is entitled Fuseki Revolution. It is about the changes wrought by AI. But there's nary a win rate or decimal point in sight, and he makes no reference to specific bots. Unless you follow the entire series, you would hardly realise he was talking about AI at all. The latest instalment, No. 55, for example, mentions the word AI once, and that's it.
That latest instalment, about the high one-point shimari, is a perfect example of his lucidity - his ability to say things that are unobviously obvious.
What prompted me to post here was not that particular instalment, since the whole series is of the same superbly high standard. It was simply that I watched a film last night called Nightcrawler. It interested me because of its journalistic theme - a stringer who is so new to the game that he doesn't even know what a stringer is (a freelance) who captures saleable video footage by being first on the scene at accidents and shootings and so on. He goes to a tv studio to sell his stuff. In one scene, where he has especially gory film to sell, there is the news director, her full-time professional assistant, and the stringer. The assistant is spluttering with moral indignation at broadcasting such images, whereas the news director is already two steps ahead ("no way we are not using this - we can pixellate out the faces", etc). I enjoyed it on the basis of "been there, done that", and given my background I found it brilliantly written. In go terms, the news director was a 9-dan. Her wimpy assistant (male, of course - it's 2020) was a 1-dan. Both pros, but worlds apart. There was also the amateur stringer in the background, soaking all this up and stirring up the pot by asking dumb questions. Which enables the news director to give some non-obvious but obvious tuition - just like Shibano (and he's only 20! - though, again, it's 2020, of course).
Shibano is talking at present about the four main shimaris, and the way he does it is to talk about the strengths and weaknesses of each of them, but in such a way as to lead up to showing why AI has shown a liking for the high two-space shimari.
Here I will mention just the high one-space shimari.
Shibano begins: "If I were to express the strong point of the high one-space shimari [ikkenjimari] in a phrase, it would be that "it is superior as regards potential for developing a moyo."
Obvious, right? Yes. But look at, for instance, Sensei's Library. This obvious fact is so obvious SL doesn't even think it worth mentioning, apparently, even though Shibano leads with it. Instead, SL leads with influence. Shibano doesn't mention influence (or thickness).
Shibano then mentions another strong point, that the high one-spacer is hard to reduce (which harks back to his treatment of the low one-spacer which is easy to reduce). Obvious? Again so obvious SL doesn't mention it. SL does talk about reduction of a moyo built on a high two-spacer, but that is muddying the waters by bringing in an extra concept - the antithesis of lucidity. Shibano talks about reduction (not an ideal word in English; keshi) of the shimari. A moment's reflection will alert us to the fact that it was reduction of shimaris was one of the first startling things that came out of the AI revolution.
In connection with such reduction, he mentions (as he does over and over throughout the series) the theme of 'over-concentration'. This has become a real buzz-word among Japanese pros. Strangely, the SL article doesn't mention it.
The SL article goes on to waffle on about the tactics (nor exactly the big idea here?) of the high two-spacer, demonstrating its weak points. Shibano says the high two-space "actually doesn't have any weak points as such."
With the true expert's eye for devastating simplicity, he gives a superbly simple tewari example.
Black has just played the square-marked stone. White had previously just played his triangle stone in the top right, against Black's high one-spacer. Many amateurs would have thought, "Great! I can defend at A and get a secure corner." Other amateurs might have thought, "Oooo, that's a bit nasty. Not sure what to do. Maybe pincer? Or leave it for now?" Whatever they do, there will be a worry in the back of their minds. But Shibano's reaction (as Black) is "Thank you. I can get on with my development." And by way of explanation he simply remarks that if the two triangled stones had not been there, would White have now played at his triangle? And would Black have answered with over-concentration at A? Of course Black would play at his triangled point. White has made him stronger, not weaker (the development potential outweighs the open skirt, Shibano says).
Proceeding, Shibano ponders how White should answer the high-two spacer. He mentions the "much-played" attachment in the following diagram, i.e. one that he sees a lot in pro internet play, where he spends most of his time. He lets others do the hard graft of data mining and doesn't use a bot himself. He just looks at the results as they appear in actual games. "Much-played," but SL doesn't mention it. Hmmm.
Shibano says A and B are the first thoughts but in either case Black is being somewhat forced. C goes against the idea of the development-based shimari. It ends up over-concentratedly defending the corner. That leaves D. (Note again the utterly pared-down simplicity of thought.)
It is interesting how he then talks about D.
Shibano says the sequence up to 13 is an example of a "recent joseki." That's it. He doesn't waffle on about whether it's "even", or about the (flawed) ponnuki shape. He doesn't mention thickness or territory, or who has sente. He doesn't even talk about his own main theme - development potential.
What he does say is that White's attachment (an AI play, remember) is about either forcing Black or overconcentrating him, and that D (the hane on top) is about Black thwarting White's intent. That's a word that comes readily to Oriental go players because of its associations with their martial arts, but I have found that western martial artists often find it hard to grasp properly - at least in that mode of expression. Perhaps in go too? (No SL entry, anyway.)
Now, Shibano does not use the word "even" about the above joseki, but I think we can safely assume that that idea is embedded in his thoughts somewhere - but not about the local position. It will be even in terms of the whole position (note, incidentally that White has used two fewer stones locally, and has sente).
Just as in haiku, what is left out is usually more revealing than what is put in. If this position really is even, i.e. simply satisfactory for both sides, Black has not really made proper use of his first-move advantage. So, digging deeper, he needs to try a wee bit harder. Think that way, and you can begin to see where the high two-space shimari comes from, and also why the low two-spacer might be preferable to the low one-spacer, and why both - because of being susceptible to overconcentration - might be inferior to their high cousins. At least in a bot's eyes.
Incidentally, while I may appear here to have been slagging off SL in particular, that was not really my intention. I find previous books on shimaris, even by pros, have similar defects - the main defect being that they are not by Shibano!. I chose SL simply because it will be the first port of call by many here, since books are so unpopular.
My own intentions were two-fold. One was a bit of "effortful practice" on my part to help ensure I understand Shibano, by re-writing some of what he says in my own words. Any infelicities are mine alone, of course.
Another intention was to show that the AI revolution can be, and is being, talked about in terms other than numbers and win rates. In human terms, in fact. But it seems we humans may need to re-appraise our previous thinking and terminology. I recommend Shibano as our guide.
Which, since I mentioned haiku and lucidity , reminds me of Basho's famous "May the moon be your guide" poem:
月ぞしるべ こなたへ入らせ 旅の宿
Last edited by John Fairbairn on Mon Mar 09, 2020 5:34 am, edited 1 time in total.
|