| Life In 19x19 http://www.lifein19x19.com/ |
|
| Why is white dead? II http://www.lifein19x19.com/viewtopic.php?f=11&t=16767 |
Page 1 of 1 |
| Author: | Jika [ Mon Jul 15, 2019 12:40 am ] |
| Post subject: | Why is white dead? II |
Another question on a 321go exercise: I've played out a solution why white is dead, but it is long and involves 2 ko threats until white is captured. I think that A2 is a false eye, but I'm finding it hard to really destroy it. |
|
| Author: | jlt [ Mon Jul 15, 2019 1:15 am ] |
| Post subject: | Re: Why is white dead? II |
If you absolutely want to remove White's stones from the board: (  plays elsewhere). But in a real game you usually don't need to do that, for instance withyou get some territory (triangled intersections), so your outside group is alive and you don't need to remove the white stones from the board. |
|
| Author: | Jika [ Mon Jul 15, 2019 2:59 am ] |
| Post subject: | Re: Why is white dead? II |
diagram 1) I think I played 1,2,3,(4) but then overlooked 5?! (I can't remember now because your solution looks so obvious. I should play out these things in an SGF editor and save the file(s) and post them here...) diagram 2) I see my opponents doing stuff like that, so it seems to work against me, but I myself can't wrap my mind around why a jump that far is enough to give me territory. OK, if I was white, I could not work out how to get the triangled intersections; but why are the un-triangled intersections closer to G7 not, well, triangled? What would be ways to learn this? Playing and watching what the others do? (This was the first time I realized it is possible to download SGF from OGS games. Maybe I should do that and replay them for me, so I can see what the opponent does without time pressure?) Thanks! |
|
| Author: | jlt [ Mon Jul 15, 2019 3:20 am ] |
| Post subject: | Re: Why is white dead? II |
Actually I was a bit too conservative, Black can get a little more than that. I assumed wrongly that F8 doesn't get a triangle because of this: but actually after  , Black can play the other hane to capture the stone:so a triangle was missing at F8, as well as other intersections probably. P.S. How far you can extend in general: https://senseis.xmp.net/?ExtensionFromAWall |
|
| Author: | Joaz Banbeck [ Mon Jul 15, 2019 8:22 am ] |
| Post subject: | Re: Why is white dead? II |
jlt wrote: Jika wrote: ...I...can't wrap my mind around why a jump that far is enough to give me territory... When a more experienced player looks at jlt's diagram, he sees this: or even this: Those two stones are connected. White cannot separate them. See https://senseis.xmp.net/?TwoSpaceExtens ... annotBeCut This may seem complicated to you now, but soon it will happen without even trying. You will look at a position and be able to say "Those stones are connected", or "That territory is mine" without even stopping to consciously think about it. |
|
| Author: | Joaz Banbeck [ Mon Jul 15, 2019 8:31 am ] |
| Post subject: | Re: Why is white dead? II |
Soon after that, you will be able to look at the position like this, and say "That stone is too close. It is inefficient". ...and you will play this: ...because you will see that it probably takes all of this territory: Be patient. It will happen. |
|
| Author: | Mike Novack [ Mon Jul 15, 2019 9:22 am ] |
| Post subject: | Re: Why is white dead? II |
Jika, I am going to try to help you with this referring to an earlier question "why dead" question you had. You are looking at it backwards. You should be asking: a) Is there any possible way those surrounded white stones could get out? b) Is there any possible way for them to form a live group? (two separate eyes) Then c) Are the surrounding stones alive? Note that they cannot be out apart (the cutting stone could be captured. They have lots of liberties, more than enough to capture the surrounded stones if they had to. d) If the surrounding stones are not alive AS THEY STAND (if surrounded, attacked) could they definitely be made so? Note that the moves being suggested to you not only gain territory but advance making the outside stones unconditionally alive. So you don't have to capture those surrounded stones NOW << the ko would be irrelevant at the end of the game when there are no possible ko threats >> What you have to watch for are things that might affect "c" and "d". For example, what if white made a play adjacent to that cutting point? Do you have to respond. The point is, immediately capturing that group WASTES MOVES. Having checked "a" and "b", you address "c" and "d" problems, trying to do that in a way that also gains territory. Remember, if the surrounding stones are connected/alive at the end of the game you don;t have to actually capture the surrounded stones because they CANNOT form a live group. MORE ADVANCED LATER (this whole thing will get revisited when you have reached the stage you need to learn about seki and double ko) |
|
| Author: | Bill Spight [ Mon Jul 15, 2019 9:32 am ] |
| Post subject: | Re: Why is white dead? II |
Jika wrote: diagram 1) To copy a diagram, click on "Click Here To Show Diagram Code" and then copy and paste the code. Like so: I have added  and  because I suspect that the exercise is about how to take off a group of stones when the opponent resists.Quote: I think I played 1,2,3,(4) but then overlooked 5?! (I can't remember now because your solution looks so obvious. You are learning, Grasshopper.  But because you had some trouble with this exercise, may I suggest playing the Capture Game? (The first player to capture one or more stones wins. No passes, or each player could just make uncapturable groups, which would lead to a boring game.  ) You can play it by agreement even if the server is not set up for it.  Quote: I should play out these things in an SGF editor and save the file(s) and post them here...) Good idea! Quote: diagram 2) I see my opponents doing stuff like that, so it seems to work against me, but I myself can't wrap my mind around why a jump that far is enough to give me territory. Usually, making territory by extending is not good enough. Real territory is made in skirmishes with the opponent, as a rule. Quote: OK, if I was white, I could not work out how to get the triangled intersections; but why are the un-triangled intersections closer to G7 not, well, triangled? On a real board my first thought would be to play  or "a". And then the question would be what points on this board are not Black territory? White might well be able to live, but who knows? Let him try.Quote: What would be ways to learn this? Playing and watching what the others do? Kibitz SDK and dan games in real time. Or download the SGF files and play them over, yourself.
|
|
| Author: | Jika [ Mon Jul 15, 2019 9:58 am ] |
| Post subject: | Re: Why is white dead? II |
Wow, guys, you are great! Thanks a lot. @ Mike: Maybe my question could be reduced to "I think white's lower eye is false, but I've found only a very long-winded way to prove that." So, the question "Can white make two eyes" is in there, the gut feeling is "no", but I failed in seeing why exactly. @ Bill: The exercise is in the life-and-death chapters, so sometimes one has to make a group live, sometimes to kill, and the website contains a few variants of correct and wrong moves. They do not go on until the stones are actually taken off the board, so sometimes I get a "solved!" but don't fully understand why. But I do have an app for capture go, and I'm quite bad at it, so I'll take this up again. How can one ask a human on a server to play this? (technically) |
|
| Author: | Bill Spight [ Mon Jul 15, 2019 10:06 am ] |
| Post subject: | Re: Why is white dead? II |
Jika wrote: The exercise is in the life-and-death chapters, so sometimes one has to make a group live, sometimes to kill, and the website contains a few variants of correct and wrong moves. They do not go on until the stones are actually taken off the board, so sometimes I get a "solved!" but don't fully understand why. They are assuming a level of experience you do not yet have. For now, play the exercises out to the bitter end. Quote: But I do have an app for capture go, and I'm quite bad at it, so I'll take this up again. Capturing is the basic skill at go. With the app at its lowest setting (if it has more than one) you should practice until you win every game as Black. Quote: How can one ask a human on a server to play this? (technically) Depends on the server, I guess. Chat? |
|
| Author: | Mike Novack [ Mon Jul 15, 2019 11:16 am ] |
| Post subject: | Re: Why is white dead? II |
Jika wrote: Wow, guys, you are great! Thanks a lot. @ Mike: Maybe my question could be reduced to "I think white's lower eye is false, but I've found only a very long-winded way to prove that." So, the question "Can white make two eyes" is in there, the gut feeling is "no", but I failed in seeing why exactly. Maybe an approach would be to try to learn what groups ARE (unconditionally) live groups. Not at the stage where obvious (fully connected, two spaces in separate interiors of the groups) but before tht stage. In other words, when still open spaces. a) First try fully connected groups. one or two spaces inside -- clearly dead (can never get to two separated spaces_ three space, either in a row or a triangle--- whose move is it? Four spaces, in a row or an L shape or a zig-zag shape --- does it matter whose move it is? in a T shape -- whose move is it five spaces --- in a Q shape -- whose move is it? Otherwise )any other 5's where it matter whose move? Six spaces --- in a rectangle? Other groups with 6 spaces? b) Notice in your example the group was NOT fully connected (so a part could be captured off from it. Are there groups composed of separate connected pieces where the pieces are not connected BUT the groupis still alive? What has to be true? << think about when each of the pieces touches two different spaces. Once you understand when/why groups are alive you will be in a better position to learn when they can be killed. In other words, prevented from reaching on of the live states. |
|
| Author: | Pio2001 [ Mon Jul 15, 2019 12:15 pm ] |
| Post subject: | Re: Why is white dead? II |
Jika wrote: Maybe my question could be reduced to "I think white's lower eye is false, but I've found only a very long-winded way to prove that." Hi Jika, There are only three kind of false eyes. The false eye by the large diagonal : The false eye by the one point jump : And the false eye by the second line : The eye on your diagram is the false eye by the second line. Also, when dealing with a ko, ask yourself what happens if Black wins the ko, and what happens if White wins the ko. A key feature of your position is that if White wins the ko, she dies, and if she looses the ko, she also dies. Jika wrote: How can one ask a human on a server to play this? (technically) If you are playing in Japanese rules, it is horribly difficult (confirmation of life and death, forbidden to recapture a ko without passing for this ko, ko bans lifted, black to play and white to play at the same time, have to restore the initial position after the confirmation...) ...Technically, you have to call an admin as a referee for your game. Under chinese or AGA rules, it is easy : just wait until the end of the game, and when you have finished with your territory, just capture the stones while your opponent passes. If you do that in japanese rules, you loose points. In Chinese or AGA rules, you don't. |
|
| Author: | Bill Spight [ Mon Jul 15, 2019 1:34 pm ] |
| Post subject: | Re: Why is white dead? II |
By this, I am pretty sure that Jika means the Capture Game. Jika wrote: But I do have an app for capture go, and I'm quite bad at it, so I'll take this up again. How can one ask a human on a server to play this? (technically) Emphasis mine. |
|
| Author: | EdLee [ Mon Jul 15, 2019 9:25 pm ] |
| Post subject: | |
I like how Pio merged sensei's two separate cases "on the side" and "at the corner" into one "by the 2nd line", but it's still nice to include both a side diagram and a corner diagram ( instead of omitting the latter ). Quote: The false eye by the one point jump : Code: [sgf-problem](;GN[Bottleneck false-eye] But I prefer sensei's naming "bottleneck" over the 1-point jump, because this shape is not a 1-point jump, but rather a "1-point jump, wedged" or a clampy situation. ( Sensei's original naming 'bottleneck' is embedded in post#12's sgf tag. )L: 1-point jump; R: not a 1-point jump |
|
| Author: | Jika [ Mon Jul 15, 2019 10:12 pm ] |
| Post subject: | Re: Why is white dead? II |
I really enjoy your different styles of teaching me: Mike's "think about this" and Pio's/Ed's "look at this"! I've made a note of them all. Thanks! |
|
| Author: | jlt [ Mon Jul 15, 2019 11:42 pm ] |
| Post subject: | Re: Why is white dead? II |
To confuse you even further: things are clear when a group consists of only one string. In the diagram below, "a" and "b" are eyes because they are adjacent to the black string, and only to it. When there are two strings: "a" and "b" are eyes because each of them is adjacent to each black string and only to them. A group can have two eyes that look like false eyes: https://senseis.xmp.net/?TwoHeadedDragon There are two eyes because the black group consists two strings, and "a" and "b" are empty intersections which are adjacent to each of these two strings and only to them. On the other hand, in the original diagram there are three strings but Black cannot play immediately at "a" so we can suppose that "a" is white, so we get two strings The eye "b" is adjacent to only one string, so the group doesn't have two eyes. Another way of seeing this is to fill external liberties with enemy stones and see that the lower white group is in atari. (All the above is pretty obvious for experienced players, however it is not uncommon for SDKs to confuse false eyes with real eyes during a game... including me...) |
|
| Author: | Pio2001 [ Tue Jul 16, 2019 11:13 am ] |
| Post subject: | Re: |
EdLee wrote: I like how Pio merged sensei's two separate cases "on the side" and "at the corner" into one "by the 2nd line", but it's still nice to include both a side diagram and a corner diagram The idea is not mine  It comes from the book Level Up! volume 4 (Lee Jae-Hwan, Yoo Chang-Hyuk). 
|
|
| Author: | Jika [ Tue Jul 16, 2019 12:19 pm ] |
| Post subject: | Re: Why is white dead? II |
@ jlt: String-theory!! Cool. |
|
| Author: | Joaz Banbeck [ Tue Jul 16, 2019 12:54 pm ] |
| Post subject: | Re: Why is white dead? II |
There are two ideas mentioned above which I think are even better if explicitly placed together: Pio2001 wrote: ... There are only three kind of false eyes. The false eye by the large diagonal : The false eye by the one point jump : And the false eye by the second line : ... jlt wrote: ... Another way of seeing this is to fill external liberties with enemy stones and see that the lower white group is in atari... If you use jlt's idea of filling liberties on Pio's diagrams, they look like this: ...and this: ...and this: Once the string of black stones is completely surrounded, black can either save it by filling in his 'eye', or white can take it, in which case if becomes a ko. And it is a ko that - as demonstrated earlier in this thread - black cannot win. |
|
| Page 1 of 1 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|