Alex Weldon: Here's a position I'm using as an example in my book-in-progress (Breaking Bad Habits). It's for the chapter on creating two weak groups (why not to do it, that is). The players were an 8k and a 9k on KGS. White has invaded and lived in the upper left corner, and Black played the marked stone to seal him in. White proceeded to try to cut through immediately (starting with z), which I would judge to be a mistake, because the marked White group is still somewhat weak, while Black is fairly strong on both sides. White will just end up with weak cutting stones, no Black stones to attack, and probably be on the receiving end of a severe splitting attack.
Hopefully, everyone agrees with me that z is too early, and that White should tenuki, leaving aji in Black's shape to aim at later. However, after each example of a bad habit, I give a suggestion for better play. Although there are dozens of moves that are better than cutting immediately, I'd like to choose one that's as close to "best" as I can manage. Usually, I'm pretty confident in my suggestions, but this one has had me second-guessing myself for long enough that I thought I'd ask for help.
Being the territorial player that I am, my first instinct was to suggest a low extension on the left side, to threaten taking all four corners, while leaving Black an open skirt to reduce his territory and weaken his left side somewhere, in case of a later cut. Although I'd probably favour b, a and c are also thinkable.
While I was trying to make up my mind about those three, it struck me that they might be the wrong direction entirely, especially since even a isn't severe enough to get an answer. Something like d is a good idea, since the bottom side is fairly big and it does something to reinforce the weak group.
Another possibility is to combine direct reinforcement of the weak group with aiming at Black's thin shape and reducing the moyo with a move like e. Seems a bit slow, however.
Because all of these moves are large and multi-purpose, it's hard for me to pick the best of them. If it's really as hard to choose between them as it seems to me, I'll just select one to use, but maybe one of them (or something entirely different) will seem much better to the other dans here.
Andy: obviously this is not dan-contributed advice, but if white is really interested in doing something about black's moyo at the top, how about f as a probe? This is bigger in scope than e, and looks more directly at the aji around z. I see no urgent moves for either player on the bottom. White's marked group has sufficiently good shape that I don't think it requires immediate direct reinforcement. It would be important for white to answer flexibly (or not at all) to a black response to f, should black respond directly.
dnerra: I like White g best. A move on the fourth line is better for limiting the prospects of Black's moyo. I think this is most obvious when compared with a, where after the exchange Black h, White c, Black could play around tengen; this move would both expand black's moyo, and increase the pressure on white's group in the center.
As an aside to your question, I think the most obvious argument against a "cut" at z is that you are cutting two living groups.
Alex Weldon: Well, Black's shape on the left is not good, so the cut does have some aji. But yes, if White cuts immediately, Black can easily defend while attacking.
jantiff: as White I would feel quite uncomfortable at the thought of Black x. I think this aji makes a play at d prematurate. f seems a bit vague, I'm with dnerra on g.
Alex Weldon: I don't like f because it also creates two weak groups, which is exactly what I'm advising against. I'm undecided on the high vs. low issue in the lower left. I'd like to hear what Charles, Bill, unkx80, think on the issue, and on my question in general.
Anonymous Coward: How about white "i"? If we presume that white is not going be able to invade the top very deeply, white might as well reduce and strengthen his group.
zinger: another idea: how about white j, to enclose black on the right and make the bottom into a pseudo-moyo?
Alex Weldon: I prefer k to j and e to i. All good ideas, though... keep them coming. As I suspected, this position makes it quite hard to pick the right move... so far, there haven't even been two of us agreeing on a single move. :-)
(Hicham) : what about t? It strenthens the white center group and limits black moyo potential. If black tries to cut it seems to me that white has several good options. This move seems to solve two problems at once. But I am only 3k EGF, so I am interseted in what high dan players think if this position. Otherwise I like c or g, maybe more c cause I have a rather territorial style.
Alex Weldon: I like t for the same reason I suggested e. I don't know which is better - my choice of e was mainly because I have a fondness for large knight's moves.
Chris Hayashida: Silly question, but is the lower right white group settled? If it isn't, I think it would be in more trouble than the marked white group. Otherwise, I think a 3-3 invasion and then splitting the two weak white groups might be difficult for White. But then again, I don't think I could win this game as White. :)
Bill: I second dnerra's play at g. But like Chris, I have little hope for White in an even match. It's not just that White is significantly behind, but what can White threaten effectively? Is f a shobute?
PurpleHaze: Make it 4 votes for g. It seems the most multi-purpose move.
Alex Weldon: Alright, g it is. I agree that the game looks bad for White and, at dan level, is probably almost hopeless. At the mid-single-digit kyu level this game was played at, however, and especially at the mid-double-digit kyu level I'm writing for, anything is possible. If White g and Black answers f, then White e and Black answers passively to defend his top territory (perhaps typical answers for a kyu player trying to defend his lead), White can aim at getting something on the lower side. At least, that's how I felt when I picked out this example.
Now I've done a bit more analysis and it seems to me that even if Black is very passive, all he needs to do is get any sort of life on the lower side, and he'll win by 10 or so. White would have to try something unreasonable to catch up, which is certainly not the sort of message I want to give in my book.
It's a pity not to use this example, since it's such an extreme case of the mistake I'm trying to illustrate. However, I don't want to say "If White plays this way, the game is hopeless. Instead, he should play this move, which is much better, although the game is still hopeless." I guess it's back to KGS to hunt for more kyu players creating two weak groups.
dnerra: Why don't you look at dan games? :P Seriously, I don't think that is the kind of mistake because of which an 8k loses a game. First you have to learn how to effectively do splitting attacks, then you realize that your opponent might be able to do this as well...
Alex Weldon: This should really be on a discussion subpage for my book, but that would require me to make a page for my book, which I'm not yet ready to do. Anyway, the book is about bad habits, which usually develop because the player consistently makes a mistake that opponents his level don't know how to punish - he therefore doesn't realise that what he is doing is wrong, and finds it harder to rid himself of the habit when he does get to the level where his opponents can punish it. I think around the single-digit kyus is where players will start to face opponents who know how to punish mistakes such as creating two weak groups, so I think a cautionary book is in order, to help players rid themselves of their bad habits before they become a serious problem.
As for why I don't look at dan games, it isn't because dans don't have bad habits. Everyone has bad habits. It's because, as a low dan myself, I don't feel qualified to look at a game played by someone of equal or higher rank and say "this is a mistake" with enough certainty that I would put it in a book. Of course, if a high dan amateur or pro wanted to collaborate with me, I'd be happy to let them pick examples and suggest improvements, while I did the explanation and the writing, since it's in those areas that my true talents lie.
zinger: Actually the closer I look, the more I agree with jantiff that black x would be too damaging to allow. Maybe white k would be best, but it does look pretty hopeless.
Calvin: I agree there is a problem finding the best move for white because white is behind, and the exercise of seeking the best overplay to shake up the game is probably not something you want to put in your book. There is a small inconsistency in suggesting g or territorial moves in that area. If you claim that the marked group is weak, then the best move should have some stronger effect which helps defend it. Also, it is not obvious to me that if white cuts at z that his intention is necessarily to make another group another side immediately---if he can cut in sente, he may be able to use the aji of the cutting stones later. Usually in reviews, if I lose a game because I create too many weak groups, the problem isn't the first one or two stones in the weak group---it's trying to save the whole weak group or multiple weak groups rather than making the appropriate sacrifices. Now you saw the whole game, so you know whether white tried too hard to save the stones on the other side of z after the cut. But just seeing z played I would not assume that is the plan. I agree with dnerra that this isn't so much a weak group problem as a questionable cut.
Chris Hayashida: Alex, just a thought, but what about showing what led up to this position as an essay on why too many weak groups are bad? I think the marked stones being separated from the lower right corner is a clear illustration of the point. Instead of worrying about cutting/not cutting, maybe you can just show why it would be really hard for White to win the game... He can't aggressively attack any stones invading at the bottom of the board for fear of hurting his floating group...
Granted, that might be harder to describe, but it might illustrate the point. However, I have to say that many kyu players (myself included) have problems visualizing far enough ahead to realize that a line of play will create two groups, let alone being able to see how weak or strong they are.
Bill: How about just showing the left half of the board? Or not worrying about it? You are showing low kyu mistakes. At that level, White still has chances.
unkx80: Sorry for responding so late. White is so behind, that I do not know what to suggest. Personally I go for e, mainly because the framework looks so disgustingly big, and the marked stones are pretty weak.