Cat's Face or Tiger Shape  

Crimson: Empty Triangles are bad for many reasons. The list starts with a wasted stone which connects two stones that can't be separated anyway, and would be better placed to form the CatsFace, which is lighter, connected better and has more eye potential.

Oryx: Many thanks Crimson! It's now three months after I wrote this jotting, and I think I can comfortably say that I do understand why the empty triangle is bad. In most cases: see GoodEmptyTriangle for the exception to the rule.

Empty Triangle And Liberties  

Crimson: Also, empty triangles are heavy, and lack liberties in close fighting. Compare liberties with three stones in a row.

Here, the two shapes have the same number of liberties, even though the lower one (the Farmer's Hat) contains one more stone. This is because of the two empty triangles.

Oryx: As is more fully explained in A Static Treatise on Shape, an extension of a bad shape is more bad shape.

More Empty Triangles & No Added Liberties  

Oryx: This can get to absurd lengths; in general, a blob of stones clumped together is either a waste of perfectly good moves that could have been out establishing moyo, or else it's the beginning of some Go parody (see the terrifying The Power of the B2 Bomber, or the famous treatise, "Get Strong At Empty Triangle").

Actually, Empty Triangles are GOOD! When you see your opponent making them, you can sit back, and thank him for the free sente he just gave you, and go invade some of his loosely-held territory.

After a 3-3 Invasion  

If I understood Ms. Guo correctly, then at this point B should extend his strong wall towards the Star Point at .
This would be even more important if B had territory in the upper right corner.

License to Kill  

Lynx pointed out that my seemingly live white group along the side here is really dead -- because while B invades, I can only defend one side at a time.
After , (a) and (b) are miai, and white dies. Painfully. So it goes.

Lynx: I'm not sure if it is dead, but it is clearly very weak. Perhaps a Dan could look at this and sort it all out?

Alex Weldon: The diagram doesn't really show enough to say whether it lives or dies. Depends on what's going on above. is weak, so if W has friends nearby to connect to, he can likely live or escape with c. Note that c gives B the option to double-hane which can lead to ko, or W giving up his group for a ponnuki on the outside, among other possible results.

Oryx: As I recall, I had nothing close to connect to (top line in diagram added after Alex Weldon's comment). But you raise a very important point, that we high-kyu players too often forget: things always go better with some friends around.

double-hane  

tderz: In this diagram, White would be dead without ko, = A, = E.

double-hane  

tderz: if Black answers a ko threat with and White recaptures with = F,
then this isn't anymore a flower-ko for Black as also separation white G is threatening.

Because of these complications I estimate this a very game-depending problem.
W can be forced to live by crawling.

 

We are at the very start of the game, four innocent moves, nothing special... then Black decides to look hungrily at White's lower right corner.

 

White doesn't feel terribly threatened, and tries to isolate with ... only to see invade at the 3-3 point at the very start of the game!

czarny: White is the most natural response for B's keima kakari to hoshi - the purpose of this move is to attack the single stone in order to :
1) avoid b's from developing along the side
2) build thickness in case of invasion

White takes the bait!  

They exchange a move each, then White grins with avaricious glee: and hanes at ! AH HAH, he says to himself, Black's a goner here!

He doesn't read, however, the threat implied by ...

Continuations after the descent  

^[1]Tderz: White is hoping in hamete-manner for the submissive Bd. White must be prepared after Wa for Bc or Bb, which are both bad for White. Bc usually leads to the greater damage for White. White can also consider simply playing c (eliminating the weakness) instead of a (has been played already in a professional game, but there will have been a white stone around or left of e, I'm sure). Black will then play a (then White g?)

Variation: hane instead of descend  

The black cut c now leads to similar outcomes as with the descent variation.

Variation (cont'd)  

(= b above [1]) is a subtle response, which shows that the hamete should lead to a bad result for White:

Variation (cont'd)  

White has committed aji-keshi with this exchange, lost the attacking aji at f and one ko threat. The price of playing hamete (might look small but is decisive in even games). I only played the descent variation in handicap games.

 

Black extends at , which draws to block; Black cuts with , White atari... ending with Black atari on , and suddenly, because of the overriding presence of , the White group of -- is short of liberties. Black takes the corner, and White's in tatters.

And all this, because White's unfortunate hane.

[phantom] Interesting example... It seems a little specific to actually remember and apply to games, but I think it does a good job of making the point that other stones can really change 3-3 invasions.


ThaddeusOlczyk Kogo describes the hane as a serious error for just this reason. It seems the hane is played out of a desire to kill the invading stone. One should always keep in mind that unless the proper steps were taken, 3-3 invasions should succeed. That is unless I am the one invading, in which case the invasion will fail.

Making the best of a bad job  

Rich: The best result I can see after this is for white to play in the original diagram as shown; there's still some aji there for later, though. White will have to take care in responding to a, black gets a big corner and could play at b for a nice thick shape along the lower side. White is left with a small group and bad aji.

Charles But see go bloopers. I wouldn't believe Kogo about anything tricky. The only attested pro example I know is Otake's hane-tsugi and cut.

Countering the descent  

Bill: Through White is quite thick. (Compare this diagram with the ones where White responds to the 3-3 with the outside block.) Isn't this good for White? Suzuki and Kitani agree (Small Joseki Dictionary).

Move 118 -- WHY?  

At this point B is ahead by maybe 50 points or more. Ok, so he cuts at Q9. But can't W just kill this single stone off? What is TheCap thinking?

Rich: W can't capture this stone.

Oryx: shhhh... don't give away the ending! :) Some of us don't see it yet, not till we read through!

Moves 119 to 128  

W suspects something with this stone, or perhaps she figures it's small, and she needs to deal with larger matters first. The game progresses... it's still there, the easy kill.

macisajt POUNCES!  

OK, so TheCap loses one measly stone. But why did he throw it away like that? Or did he?

A Little Black Magic  

... and suddenly a very nice 8-point or more profit garnered. Cute trap, if I do say so my [untutored!] self!

Rich: ....so there is no trap :)

Oryx: well, that's easy to say (and see) now, but beforehand, during the game, it was pretty amazin' to me.

TheCap eventually went on to win the game, but not before macisajt showed her 7d strength, winning a cliffhanger of a semeai in the lower middle, winning back nearly 30 points for a 7.5 point final margin. One day hopefully I'll present that battle, and we'll see if, courtesy of 20-20 hindsight, TheCap could have pulled it off.

 

ThaddeusOlczyk: Actually there is nothing surprising in this combination. Many kyu level players should see it.[101]

, and a and b are miai. White plays one Black takes the other and the cutting stone escapes.

This would be plain to either TheCaptain or macisajt, and macisajt did not play these these moves to capture the cutting stones, but rather to extend her right group into Black territory in sente. ( Which and a do. )

The stones were captured way earlier in the game when Black played 8 and 10 in your sequence. Why White did not fight harder, I do not know, but understand that White did not feel they were worth saving.

 

If one replaces one of the weak white stones with a black stone, Black would respond to 1 with 2. This is a difference of roughly 7 points for White. One says that the threat of reviving the dead white stones is aji, which White uses to gain 7 points.

Move 106: The invasion starts  

At this point I'd just blown a group on the middle left, but I had some nice area on the right and a bit on the lower middle left... so I thought I'd try invading.

C0nfuseki: I personally think this invasion can be countered by Black hitting underneath at Square. But more importantly, I think Black can both save the corner and threaten the White group by playing a nice move like Circle instead. This allows for Black play at a to kill the White group.

symplicity: I would agree that the circle or thereabouts looks pretty large right now, though I'm not sure how a kills white afterward. Anyways I agree, this invasion should not live. In fact, I'm not even convinced white can live if black ignores it, since black is so ridiculously thick all around.

Move 106: The invasion is dead in gote?  

symplicity: Say black ignores it. If white tries to escape at , is enough to seal it in. If white tries to make a base at , the long knight's move at removes essentially all of the eyespace. Since Black has no major cuts or weaknesses to worry about, so I don't see anything white can do.

Moves 107 to 116  

nakedjazz sees the threat to his center and blocks with , while I continue with the invasion. I have a lovely spread of white stones that has every promise of life, though at the cost of little to no influence towards that center.

I'm fairly satisfied with this result, but I wonder if I could have invaded differently, and retained influence to the middle... Comments? Ideas?

symplicity: I know you played as white, but if you put yourself in black's place, this position could actually serve as a good case study in killing much-too-deep invasions, if you're interested =). Note black's weakness-free massive walls on both sides. That's a good sign that black should be able to kill this invasion. So (assuming you aren't confident enough to play tenuki), how should black do it? There's no need for black to start playing crazy contact moves like . These only complicate things and help white establish a base. Instead, the easiest way is often is a peaceful, seemingly lazy attack.

Focus on 3 things: protect your weaknesses, stop the escape, and remove the base/eyes.

Black has no weaknesses or cutting points, so we can check that one off for now.

It would take several moves for white to escape all the way out, on pretty much all of the circled points. So we don't need to worry about escape much yet.

So let's remove any hope for white getting a base.

Peaceful death  

We can do that with a move like . It may even be too passive, but it is simple, gives white no chance to complicate. If white blocks at a, black has no weaknesses still, so he can simply hop in from the other side at b or c and again white has no base. If white plays c or d, black plays the simplest possible move to deny any base, a slide at e.

Peaceful death 2  

Following the game sequence, is still fine for black. But now, there is no need to play a reckless contact move ( at a) as in the game. White definitely can't escape now, so just deny a base. For instance, is fine, or if you aren't confident enough that you can connect up with the right, b or c are also perfectly fine. Even blocking at d works.

Peaceful death 3  

Even a few moves later, black has an easy, simple way to kill. It's clear white has no way to escape, so rather than playing on top, as in the game, black can simply undermine white's base at . So, the rule is: just remove the base, stop the escape, and play solid and simple. -symplicity

Coconuts: In my eyes this invasion is far, far too deep. Black can rightly consider almost all of the space above tengen as territory. Suppose was at ...there are almost no options left for white. Jumping out seems impossible, the cap of is all but unnecessary. The result is amazingly good for White...but I don't think we can consider this anything but a swindle. =\

Around a is the edge of this very territorial moyo, and the only real safe place to play keshi. Unfortunately for white, black's gains in this game seem too great to overcome if black plays solidly.

As a more general response to your thoughts about the keima, is a move that seeks a base, ie tries to find some eyespace and more area to live in. You're right, the keima from the fourth to the third line, or the third to the second, is a good way to look around for the base, especially when it is a slide underneath an enemy stone. You will see this type of move frequently in many situations, not just deep invasions. For a simple example, take a look at perhaps the simplest joseki. Both black and white make keima extensions in this joseki, each looking for a base more than territory.

The keima is a good starting point when your stone starts feeling a little cramped, in need of some settled space to live. Don't consider it an iron rule, though (or anything else in this game! =) ), as sometimes you are left with an ugly angle play that does more harm than good.

Oh, if you have it around, can you show some of the moves that led to the position in the upper left? I ask because W at b looks interesting, if only white were connected or had one or two more liberties. Black is tantalizingly weak up there! Cheers =)