Quote:
I want to play on the border of two moyos and attack the marked stones on a large scale.
- Click Here To Show Diagram Code
[go]$$c
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . j . . . . . O . . . |
$$ | . . . O . . . O . X . i . O . . O . . |
$$ | . . . , . O . . . k . . . . . X . . . |
$$ | . . X . . . . . . . . . . . . . . O . |
$$ | . . . . . . . . . . . . . . . . X O . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . X . . . . . . h . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . , . . . . . , . . . . X , . . . |
$$ | . . . . . . . . . . . . . . . . O . . |
$$ | . . O . . . . . . . . . . W X X X . . |
$$ | . X O . . . . . . . . . g . W . W X . |
$$ | . X O . . . . . . . . . . W . . W X . |
$$ | . . X O . . . . . a c . . . . . X . . |
$$ | . . X O . . . . d f . b . . . X . . . |
$$ | . . X O . . . . e O . . . X . . . . . |
$$ | . X O O . . . . . . . . . . . . . . . |
$$ | . X . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]
Although it would seem to be taking the discussion away from the concept of bases, I think we can widen the discussion in a way that sheds some light on that concept. And we can use known proverbs or heuristics.
First, I think we can probably agree that, in this position, talking about a base at the top is premature anyway. We can get to that conclusion my applying a standard proverb: "Play in the widest area". This seems to be a good proverb to apply because AI seems to follow it whenever strategic considerations override tactical ones. The important thing to note is that (yet again) words can traduce. "Wide" is not limited to gaps on the third and fourth lines, although it does include those. The real meaning is "spacious" and so includes the centre. That is enough (with no tactics to worry about) to tell us that our focus should be around 'a'.
Having grown up in go in the Age of Takagawa, my first instinct would be to look at the cap at 'a'. There is no proverb I can think of that justifies that (as a cap). There is a proverb that speaks against it ("Answer a cap with a knight's move", i.e. 'b' here). In addition, I learnt from bitter experience that a cap may feel like a sente move, but it can instantly turn into a gote liability (as with 'b' in answer to Black 'a' here, stranding the capping stone).
But the known proverb about playing at the pivot point between two moyos (e.g. 'c') also seems questionable here. There is a risk of letting White turn his moyo into territory instantly without being reduced in any significant way. But more pertinent is that Black doesn't really have a moyo yet in the centre of the lower-right quadrant, as it is populated by some white stones. I'll come back to that in a moment.
My sense of what AI (or, more accurately for me, of what AI-inspired pros) would play is the much more assertive (Chinese jin 紧) shoulder hit at 'd'. The thinking behind this "path of the devil" (I would imagine) would be that if White answers at 'e', that would be too submissive and he would end up being reduced far more than by the cap. But if White pushes up at 'f', that would be giving Black momentum (choshi) to move smoothly into the centre. So again, existing human knowledge can be used to justify that (always assuming it turns out to be correct).
More importantly, in my eyes, that would make the centre of the lower-right quadrant bigger (in the sense of 'more at stake').
We can look at even this through human eyes. In the days of the first London Go Centre (so some 50 years ago) there was a big buzz one day when somebody revealed a Russian proverb (I think someone had picked it up at a European Congress). It was along the lines of "if an area is worth 15 points or less, you can tenuki". David Mitchell, who then worked in the LGC, compiled a book or proverbs and included that one - anyone less lazy than me might care to look up the precise form of it. There were the usual "numbers guy" arguments about whether it should be 14 or 16 points, and later arguments about whether it really was the Russians who first discovered it. But, as far as I know, it had not previously surfaced in the Far East, so we can reasonably call it a European proverb, I think.
Is it reliable? It seemed to have dropped out of mainstream use fairly early, but I for one have never forgotten it, and I recall it every time I see a (to me) startling sacrifice in AI-type play. And I usually come to the conclusion that the proverb would be praised by AI if only it could talk. The area in the lower-right centre does seem to be worth around 15 points and so, following this proverb, that area should be left alone. I am therefore guessing that if Black played at 'g' to make the white group there heavier and so attackable, White would ignore it, although I suspect in real play many of us would just grit our teeth and connect.
But if Black starts with the shoulder hit 'd' and White does push him upwards via 'f', any play in the lower-right centre becomes bigger and more urgent. If we then envision any sort of attack on the white group, there would be twin Black forces moving towards the upper centre and converging somewhere around 'h'. In that case, any black stones in that area would be lighthouse stones (zhaoying) for the lone black stone at the top.
In that case, Black would not really have to contemplate making a base with 'i'. And by a similar line of reasoning probes at either 'j' or 'k' in that area are premature.
The best probe to play, therefore, I would suggest, is Black 'd'. My sense of AI play is that this is the sort of whole-board strategic play they adopt almost all the time (though I do think they prioritise moves in the widest areas).
I suspect it also marries with the pro style of play, though in their case they, unlike the bots, are at the limits of their whole-board vision and so are likely to make mistakes. I say this in the light of an experiment in the mid-1970s when Rin Kaiho (I think) and others tried playing on a 21x21 board. They came to the conclusion that it was too big for humans. The game just descended into a series of small tactical skirmishes with no obvious (to the pros) connections between them. They decided 19x19 created the best balance between tactics and strategy that left enough scope for differences in human skill to emerge reliably.
Of course, I could be all at sea here. My sense of AI play is almost 100% derived from looking at games by humans who appear to try to play like the bots. Maybe even that's too big an assumption on may part.