Life In 19x19 http://www.lifein19x19.com/ |
|
#215: The Intrepid vs. drishti: Deep in the Endgame Weeds http://www.lifein19x19.com/viewtopic.php?f=37&t=8305 |
Page 1 of 5 |
Author: | The Intrepid [ Sat Apr 27, 2013 2:36 am ] |
Post subject: | #215: The Intrepid vs. drishti: Deep in the Endgame Weeds |
Deep in the Endgame Weeds The diagram below is the adjourned position from a 9-stone handicap game between drishti and me which we started playing a week ago (Black, drishti, to move). The hour was getting late, so we got kicked out of the go club before we could finish the game. Since drishti could not come to the club this week, we decided it would be great to play the rest of the game online ... and where better to do this than on lifein19x19.com? ![]() As can be seen from the position, the game is at the late-endgame stage. Since beginning players, double-digit kyus, etc. don't usually have many opportunities to get deep into the weeds of an endgame, I strongly encourage such players to follow this game and to do so not passively but actively.
For beginning players: 1). Who is leading in the game? 2). How large is the lead qualitatively? 3). Can you calculate, or at least estimate, the score? 4). It's Black to move, so how should he play next? For players of any strength: 5). What's the difference in score between White playing at "a" and between Black playing at "b": a). without the knowledge of what the rest of the board looks like and b). with the knowledge of the rest of the board?
|
Author: | bgrieco [ Sat Apr 27, 2013 7:32 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
To the Intrepid : 1) White 2) IMHO : w has 2 whole corners + half ul corner; almost all upper side; the lr corner advances to the center taking almost 1/4 of the board + 9 prisioners; b has ur corner + r side; l side with half ul corner; 2 points on lower side; 2 points on upper center; 3) roughly 25 points ahead 4) b could try something at the ul corner: A18 or he could start closing it's borders: P19 or P9 or B4 5) w playing a takes only 1 point; b playing b takes at least 2 points from w; a) and b) I can't see no global effects of this play, unless b manages to take the upper side, which I don't see how. |
Author: | The Intrepid [ Sat Apr 27, 2013 7:42 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
@bgrieco Actually, the situation is the other way around. I (White) am hopelessly lost. I'll write about it in a second in my next post. |
Author: | tj86430 [ Sat Apr 27, 2013 8:08 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
A question about prisoners: does ![]() |
Author: | Shaddy [ Sat Apr 27, 2013 8:51 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
For beginning players: 1). Who is leading in the game? 2). How large is the lead qualitatively? 3). Can you calculate, or at least estimate, the score? 4). It's Black to move, so how should he play next? For players of any strength: 5). What's the difference in score between White playing at "a" and between Black playing at "b": a). without the knowledge of what the rest of the board looks like and b). with the knowledge of the rest of the board?
1. Black, I think 2. Not large, less than 10 points 3. Please no 4. T5 sente, then p19 5. I don't know, but if the rest of the board is tiny, Black actually plays n18 |
Author: | The Intrepid [ Sat Apr 27, 2013 9:25 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
Prisoners = ![]() ![]() |
Author: | tj86430 [ Sat Apr 27, 2013 9:30 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
The Intrepid wrote: Prisoners = ![]() ![]() That's how I interpreted it. That being the case: Even after looking other posts, I count white leading with around 5 points. I'm very interested in learning why white is hopelessly lost. |
Author: | illluck [ Sat Apr 27, 2013 9:54 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
I started responding but then lost my post T.T Anyways, my responses weren't too interesting except for: 4. T5 and then N18. 5. About 4 points, but as above I think N18 is better for both a) and b). |
Author: | drmwc [ Sat Apr 27, 2013 4:18 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
Not for players. I'm drunk, so treat this analysis with caution!
B 1 looks powerful. If W 2, B connect up for a big gain. So W may resist:
However, B 3 threatens to connect again, and W can't resist without collapsing. Another line for W:
Again, it's a disaster for W. Sober edt: It's not a disaster, but B gets quite a lot.
So W needs to let B connect up, and it's very big. So the line below seems a little duff for B.
|
Author: | The Intrepid [ Sun Apr 28, 2013 5:38 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
Warrning! A wall of text incoming! ![]() First, please note that the title of the thread contains a play on words, or, more accurately, play on idioms. Whereas "getting into the weeds" means getting into the detail of something (often implying getting into too much detail), the expression "in the weeds" usually refers to someone (especially a waiter at a restaurant) being hopelessly behind. Thus, although the title is meant to convey that the game will contain detailed endgame analysis, the title is also meant to signify that I’ll need to perform a miracle to win this game ... or, more precisely, I’ll need to induce drishti into making a serious mistake if I am to have a chance. My evaluation of the position is that Black is ahead by 13 points (including prisoners). (There is some trickiness as to whether Black gets to capture the ko in the middle or not, but I think he does; otherwise, the lead would be 12 points.) This is a lot in itself in any endgame, but here the real problem is that almost wherever Black chooses to play (and it’s his move), he can expect to keep sente, while just about all of my potential plays end in gote. If the position where within 5-6 points, I would adopt the strategy of making optimal moves, hoping to capitalize on small mistakes from my opponent--a reasonable strategy given the playing strength disparity and given that drishti has not had much endgame practice or practice counting in the past. However, with 13 points difference and with the given sente/gote-plays distribution, such a strategy is doomed to failure. Thus, I need to resort to making at least one overplay somewhere (a move that leads to a loss of points, objectively speaking, but one that may induce a mistake from the opponent). As a parenthetical remark, I should say that it’s very enjoyable for me to play lost positions. It’s very liberating in many ways. When the position is balanced or one player has only a small lead, a player needs to be weary of pressing his attack too far, of overextending, of making mistakes big and small. However, when the game is hopeless, such as here, it’s possible to try playing anything, even the most ridiculous invasions and overplays, because whether one loses by 10 points or by 100, it’s still only 1:0. Beginning players: take note from this attitude! Rather than hanging your head and resigning after a huge blunder, look at the position again and ask yourself: “what’s the craziest invasion I can do in the position”? Or, stated differently, “what would Magicwand do”? ![]() Answers to previous questions: 1). Black 2). Black is significantly ahead 3). 13 points in Black’s favor (more on that in another post) 4). I’ll come back to what Black should play in my next post 5). To calculate the score difference between White playing in the upper right corner and between Black playing there, we ought to consider the following diagrams:
![]() Thus, we end up with this diagram:
If Black plays in the area first, this leads to the following sequence:
Beinners beware!!!
Thus, the final diagram is this:
The value of (Triangle) is equal to 0, 0.5, or 1 depending on how much we know about the rest of the board. Given the knowledge of only the local position, it is unknown who will play first in the area, so the best we can do is say that each player has 50% to do so. If White is the first one to play, neither player will have any points in the area (i.e., Value(White first) = 0). If Black is the first one to play, he will have 1 point, White none (i.e., Value(Black first) = 1). Therefore, without knowing what the rest of the board looks like, (Triangle) = (50% * 1) + (50% * 0) = 0.5 => Score 2 = 4 - 0.5 = 3.5. (Incidentally, from a mathematical point of view, it is more convenient and less error-prone to assign negative values to points of one of the sides, say B points are negative, White positive. Also, it would perhaps be more accurate to write 3.5 +/- 0.5 instead of just 3.5, but I’m getting technical.) Hence, we conclude that playing at the upper-right corner equals Score(limited knowledge)= Score 2 - Score 1 = 7 - 3.5 = 3.5 If we do know what the rest of the board looks like, then we might be able to determine whether (Triangle will or will not yield Black an extra point. After all, at the end of the game, there are no half-point values anywhere. In the present situation, I believe that Black will be able to make the point there, leading to the conclusion that the Score(full knowledge) = 3. To summarize, the answer to question 5). is: a) 3.5 b) 3 (perhaps I’ll return to this later). For beginners: 6). Now that you know my evaluation of the position, try again to count the board and to play out the endgame to see if you can match my result. In the very least, make sure that you can get that Black, and not White, is leading by a significant number of points. Being able to do such evaluations accurately is of paramount importance! Knowing who is ahead and by how much dictates the strategy which one ought to adopt at a given point. 7). What’s the difference in score between White playing in the lower-left corner (at "a) and Black playing there (at "b")? For each player, is the play sente or gote?
|
Author: | The Intrepid [ Sun Apr 28, 2013 5:55 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
drishti sent me a message that the following is the move he wishes to play:
White has terrible shape in the middle: cutting points everywhere, the dead 9 stones still have aji, and so on. However, is ![]() ![]() For beginners: 8). What is Black theatening to play as followup to ![]() For players of any strength: 9) What is the point difference between ![]()
10). What should White play in response to ![]() |
Author: | illluck [ Sun Apr 28, 2013 8:05 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
The Intrepid wrote: ...If Black plays in the area first, this leads to the following sequence:
Not exactly. Black has this move here:
Therefore, drwmc and I (implicitly in my estimate of the difference between the two moves) visualizes the below sequence:
|
Author: | The Intrepid [ Sun Apr 28, 2013 2:42 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
As explained previously, I cannot simply play moves that are optimal from the objective point of view. I have to make overplays. There are two areas where I could make up 10 point or more. The right side is one of them, seeking to take advantage of the aji of the marked White stones and of the potential shortage of liberties of the marked Black stones.
The second area is in the middle left, where I have a remote chance to cut off the big Black group in the middle while poking out it's potential second eye.
The second option does not seem promising at all, but it is definitely possible to go wrong for Black on the right side. As far as the questions I posed earlier, I'll return to these later. As a teaser, I'll just briefly answer question 7): The difference between White playing at "a" and Black playing at "b" is 6.5 points. Details to be posted later.
|
Author: | illluck [ Sun Apr 28, 2013 5:18 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
The Intrepid wrote: As far as the questions I posed earlier, I'll return to these later. As a teaser, I'll just briefly answer question 7): The difference between White playing at "a" and Black playing at "b" is 6.5 points. Details to be posted later.
Just a side comment, if you get a 0.5 yose value difference when it's a symmetrical position, double check your calculations. Also, one idea is to extend down instead of hane to keep sente while removing the possibility of the opponent playing the hane himself. |
Author: | drishti [ Sun Apr 28, 2013 5:40 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
Trying to make sure that a) my groups don't get split up and b) The Intrepid's groups stay split. |
Author: | The Intrepid [ Sun Apr 28, 2013 5:53 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
The overplay did not yield the desired result. ![]() ![]()
Oh well, let's see how optimal his moves will be from now on. |
Author: | The Intrepid [ Sun Apr 28, 2013 6:34 pm ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
I need to lay low for a little bit. My hope is that Black becomes careless about the liberties of his group on the right, which could offer me a chance to make something happen there. Specifically, I am aiming to play ![]() ![]()
Incidentally, ![]() ![]() |
Author: | drishti [ Mon Apr 29, 2013 5:27 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
|
Author: | Bill Spight [ Mon Apr 29, 2013 8:25 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
The endgame is fighting strength. -- Takagawa Edit: Unhidden. |
Author: | illluck [ Mon Apr 29, 2013 8:41 am ] |
Post subject: | Re: #215: The Intrepid vs. drishti: Deep in the Endgame Weed |
Black's last move is very strong! I was thinking of K7, but this cleanly captures the two stones. |
Page 1 of 5 | All times are UTC - 8 hours [ DST ] |
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |