Life In 19x19
http://www.lifein19x19.com/

Two games against GNUGo - another reliable opponent :)
http://www.lifein19x19.com/viewtopic.php?f=4&t=11991		 Page 1 of 1
Author:  tekesta [ Thu Jul 02, 2015 8:26 pm ]
Post subject:  Two games against GNUGo - another reliable opponent :)
Hello. Today I played a total of 8 games, all blitz because of the tiny time limit, against GNUGo level 1 (10k) on OGS. Of these 8 I won 2 and lost 6. On here I post the last 2 games I played against this engine. For comparison, I have included one of my won games.







Much obliged in advance for your comments.
Author:  Schachus [ Fri Jul 03, 2015 4:30 am ]
Post subject:  Re: Two games against GNUGo - another reliable opponent :)
Hi,

I think that you should attack opponents weak groups a lot harder. For example game 1 move :w35: tenukied from his baseless group in the lower right, that doesnt even have a lot of space to run to and ist far too heavy to be sacrificed. You should definitely add another move there threatening to completely surround it from the center(I'm not strong enough to tell which one, maybe o10 trying to build on the right after the process(the push at :b36: first is also good, maybe even better, but then you should come back an attack lower right with :b38: )). Also after white finally jumped out at :w47: his group is still weak, you should let him connect to top right so easily as in the process imo.

Another thing I noticed is that the hane from outside at :b18: (game 1) and :b10: (game 2) seems to be just bad(and both times you did end up with bad results there). After the white answer, it is like white extended first from a crosscut(which is a good thing for him), in an area, where you "should be stronger", because you played first in that corner. Just hane from the inside instead, which seems to be the usual joseki.

Greeings,

Schachus
Author:  Charles Matthews [ Fri Jul 03, 2015 11:39 am ]
Post subject:  Re: Two games against GNUGo - another reliable opponent :)
tekesta wrote:



Game 2. Nothing very dramatic occurs in this game. But White catches up steadily. It is instructive to see how this happens, in detailed sequences.

:b10: is a mistake, as been pointed out. You add quite a lot to the loss, though, by adding further stones. This may have lost 20% of the handicap.

:b26: It is usually not possible to "play on both sides", as you try here, and get a good result. Here it makes sense just to answer at Q6: corners are important. Up to :w40: your pairs of stones on the M line and O line suffer from bad shape. :w49: and :w51: are worth more than :b50: and :b52:. Ignoring atari when the capture is small is good play. and you are left with the small change.

:b60: would obviously be bigger at J11. The cut is covered by a net.

:w79: tries something: I don't believe White deserves the result obtained.

Black 110: you really had to try M17. White cuts, you play M18. You might learn something about liberties, at least.
Author:  tekesta [ Fri Jul 03, 2015 4:55 pm ]
Post subject:  3 games against GNUGo - another reliable opponent :)
Here is my latest game against the GNUGo engine on OGS, which I played after spending hours working outdoors in the intense heat & humidity of a Florida summer. I thought I was behind as White got two living corner groups on the bottom.
Author:  tekesta [ Fri Jul 03, 2015 5:13 pm ]
Post subject:  Re: Two games against GNUGo - another reliable opponent :)
Charles Matthews wrote:
:b10: is a mistake, as been pointed out. You add quite a lot to the loss, though, by adding further stones. This may have lost 20% of the handicap.
Now that I realize it, it was an open invitation for White to cut Black apart. I played this way in an effort to emphasize outward influence. Of course, now I know sound reinforcement comes first.

Quote:
:b26: It is usually not possible to "play on both sides", as you try here, and get a good result. Here it makes sense just to answer at Q6: corners are important.
I played this way to undermine White's base. Looks like the timing was off on this one. :b28: should've been at Q6, even if :w29: is at M4.

Quote:
Up to :w40: your pairs of stones on the M line and O line suffer from bad shape. :w49: and :w51: are worth more than :b50: and :b52:. Ignoring atari when the capture is small is good play. and you are left with the small change.
The capture was greedy on my part; I did not want White to connect, but then this would probably be better left for later on in the game as it is a small-scale move. Letting White consolidate the moyo (large-scale move) worked against me. Shows I far I have to go as far as whole-board thinking is concerned. I often let this happen and this leads to me losing games.

Quote:
:b60: would obviously be bigger at J11. The cut is covered by a net.
No contest on that one. I went for a solid shape. A play at J11 seemed too loose at the time. Now that I think about it, J11 would've been a better place to play as it would reduce White's moyo a little.

Quote:
:w79: tries something: I don't believe White deserves the result obtained.
Bad reading on my part. :b82: at S18 is to make Black live as White already had him surrounded.

Quote:
Black 110: you really had to try M17. White cuts, you play M18. You might learn something about liberties, at least.
Another one that I missed! Then again, if I had longer time limit I could've spotted the move after a couple minutes of reading. This is why Go experts recommend playing long games instead of shorty ones. As for :b110: I was hoping to use a tendril to split White in two.

Well, Charles, thanks for the review m(_ _)m
Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/