| Life In 19x19 http://www.lifein19x19.com/ |
|
| 4k/5k (KGS) game analyzed with crazy sensei http://www.lifein19x19.com/viewtopic.php?f=4&t=13994 |
Page 1 of 1 |
| Author: | jeromie [ Sun Jan 29, 2017 2:05 pm ] |
| Post subject: | 4k/5k (KGS) game analyzed with crazy sensei |
I thought it would be interesting to see what nuggets of information I could glean from crazy-sensei's analysis of one of my recent games. It's an interesting way to review a game, because the engine can tell you which moves were a mistake but it can't tell you why. It's a bit overwhelming to see how many of my moves are considered mistakes, though. I have to pick and choose which ones to look at more closely. Here's the game: Here's a link to the game with analysis on crazy-sensei.com. I was playing black. I've dipped back into 5k on KGS and my opponent was 4k, so there was no handicap. I tried 1 in this game based on feedback from the one other game I analyzed with crazy-sensei. My normal play is at the space marked with a square, but the analysis prefers the approach. Overall I think this is a matter of style/preference, but I figured it would be worth trying this approach. I know it's popular these days. This is the first big error in my play. Crazy Stone would have played at the space marked with a square. The bot actually recommended another move in the upper left before this, but the error bars for my other plays were relatively small. After white plays the marked stone, however, another move is urgent. This shows one of my repeated weaknesses: I often fail to discern when I need to add another stone to a group once my opponent has played near it. When my opponent jumped toward the center with the marked stone I played at the circled point, which is the right idea but in the wrong location. It's possible for white to completely surround the black group at that point and make it a liability rather than a strength. My next few error were place where I continued to miss a chance to make sure this group did not get hemmed in. Here, I should have played at the marked point to weaken the white group on the side. Once I let that group become stable, my winning probability dropped significantly. I missed another opportunity to play in the same place a couple of moves later, and the game turned in favor for white. It is my instinct to split groups of my opponents' stones, but sometimes I can apply that in the wrong situation. In this move, both of the white groups are stable; splitting doesn't do much. On the other hand, white has a chance to solidify a large territory in the lower left if I don't make a play there soon. I should have played at the marked location. I'm surprised at the size of the error bar on this move. I made crazy-sensei's second choice, but the marked point was much better. I can't kill white with proper play, but I can reduce the size of the territory and seal him in while retaining sente. This move was only off by one point, but it was enough to make it bad. I didn't read a way to get the black stone back out, but my feeling was that I was far enough behind that the normal move along the sector lines would (edit: not - big difference in meaning!) be enough. All this did was help white solidify the territory, though, and I was way behind at this point. I ended up winning the game when my opponent made a big mistake and lost the upper left. There are still lots of error bars after that, though, because the game was still relatively close and we both played kyu level endgame. It's eye opening to see how many opportunities to put the game away I missed. I definitely liked using the crazy-sensei tool to analyze my game. It still requires me to think about why a move was good or bad, but at least I have some direction regarding which moves to examine and which alternate lines to consider. |
|
| Author: | daal [ Sun Jan 29, 2017 3:05 pm ] |
| Post subject: | Re: 4k/5k (KGS) game analyzed with crazy sensei |
jeromie wrote: This move was only off by one point, but it was enough to make it bad. I didn't read a way to get the black stone back out, but my feeling was that I was far enough behind that the normal move along the sector lines would be enough. All this did was help white solidify the territory, though, and I was way behind at this point. I think the sector line is between E10 and K8 (not H7) - this would mean you are pretty far in as opposed to the CS move which is just a bit over the edge... |
|
| Author: | Bill Spight [ Sun Jan 29, 2017 4:41 pm ] |
| Post subject: | Re: 4k/5k (KGS) game analyzed with crazy sensei |
One thing that Crazy Sensei indicated about your endgame was not taking your sente. I noticed one point where it missed the right play. Here is the sequence in the game, starting with move 197. Crazy Sensei's recommended sequence: The sente sequence,  -  , is fine. But then  is only a 1 point move. White should tenuki and play at  .  is wrong locally. -  is tesuji, gaining 2 points. 
|
|
| Author: | jeromie [ Sun Jan 29, 2017 5:34 pm ] |
| Post subject: | Re: 4k/5k (KGS) game analyzed with crazy sensei |
Daal, I edited my original post: I meant to say that the move along sector lines would NOT be enough. I knew I'd was pretty far in but thought I might be able to make something work, and figured my position was hopeless if I could not. |
|
| Author: | Gomoto [ Thu Dec 07, 2017 10:14 am ] |
| Post subject: | Re: This move ? |
jeromie wrote: This move was only off by one point, but it was enough to make it bad. I didn't read a way to get the black stone back out, but my feeling was that I was far enough behind that the normal move along the sector lines would (edit: not - big difference in meaning!) be enough. All this did was help white solidify the territory, though, and I was way behind at this point. <Admin> Moved this post here as I thought it was a good reply to a spam reposter (they are getting cleverer, including image!) about to be deleted </Admin> |
|
| Page 1 of 1 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|