| Life In 19x19 http://www.lifein19x19.com/ |
|
| Optimal number of groups http://www.lifein19x19.com/viewtopic.php?f=18&t=8280 |
Page 1 of 1 |
| Author: | Raspingbard [ Sun Apr 21, 2013 5:44 pm ] |
| Post subject: | Optimal number of groups |
I have a small project idea for any go programmers out there who want to tackle it. I've always wondered whether there is an optimal number of groups to aim for in a game. Or rather, whether there are optimal ratios of groups that might have a higher winning percentage, e.g. four groups to three groups wins 58% of the time. If anyone has access to a large database of pro games, it would make for an interesting project to compute the win statistics for all the games based on their group ratios. |
|
| Author: | EdLee [ Sun Apr 21, 2013 5:47 pm ] |
| Post subject: | |
That's an interesting idea. Would you count seki as 2 groups ? |
|
| Author: | Toge [ Mon Apr 22, 2013 2:05 am ] |
| Post subject: | Re: Optimal number of groups |
Fewer groups are better than large number of them, since each individual group has to have walls around them, resulting in inefficiency. Fundamental tactics like cutting and connecting are about separating opponent into two groups and keeping own groups together respectively. Resulting game position at the end depends on fighting ability. Knowing the optimal number of groups, if it exists, wouldn't be good measure to govern decision making. Sometimes you have to invade (=create separate group) if situation calls for it. Endgame may present opportunity to connect two independently living groups for no profit. |
|
| Author: | John Fairbairn [ Mon Apr 22, 2013 3:00 am ] |
| Post subject: | Re: Optimal number of groups |
There is already a proverb that summarises pro experience - if you make six groups one ought to die. However, confirmation of that from a database would be useful, and it would be interesting to know not just the optimal number but also the optimal ratio (own groups:opponent' groups). If anyone does tackle this, do exclude ancient Chinese games as group tax applied, and so emphasis was put on making as few groups as possible. Also, since groups can be connected up somewhat artificially in the last stages of a game, I'd suggest choosing a mid-point such as move 100 to count at. |
|
| Author: | RobertJasiek [ Mon Apr 22, 2013 3:43 am ] |
| Post subject: | Re: Optimal number of groups |
There is no single optimal number in practice. It all depends on strategy and positional context. E.g., often when I have 6 groups on the board, I reflect whether 5 would have been better. Usually, the answer is "no". |
|
| Author: | Uberdude [ Mon Apr 22, 2013 3:55 am ] |
| Post subject: | Re: Optimal number of groups |
I like to win games with 7 groups. Breaking proverbs is fun. Not sure if I've won with 8. |
|
| Author: | billywoods [ Mon Apr 22, 2013 8:18 am ] |
| Post subject: | Re: Optimal number of groups |
Toge wrote: Fewer groups are better than large number of them, since each individual group has to have walls around them, resulting in inefficiency. But there are two players on the board. If you have a half-board moyo, your opponent probably does too; if you have 6 groups (and haven't resigned yet!), your opponent probably does too. "Inefficient" in your post seems to mean "doesn't get me much solid territory", but it doesn't seem fair to call a move by black "inefficient" without taking into account the move it forces white to make, for example. |
|
| Author: | Uberdude [ Mon Apr 22, 2013 8:28 am ] |
| Post subject: | Re: Optimal number of groups |
It's inefficient for you if your opponent creates an alive group inside your (former) territory! |
|
| Author: | Chew Terr [ Mon Apr 22, 2013 8:44 am ] |
| Post subject: | Re: Optimal number of groups |
I always liked my friend's 'pseudopod' strategy when we were learning. He would make one small, living group, and just try to use that to reduce the whole rest of the board, just to be sure that I couldn't kill his stones. =D |
|
| Author: | Uberdude [ Mon Apr 22, 2013 9:43 am ] |
| Post subject: | Re: Optimal number of groups |
Chew Terr wrote: I always liked my friend's 'pseudopod' strategy when we were learning. He would make one small, living group, and just try to use that to reduce the whole rest of the board, just to be sure that I couldn't kill his stones. =D The Go equivalent of Red Alert's silo base attack 
|
|
| Page 1 of 1 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|