| Life In 19x19 http://www.lifein19x19.com/ |
|
| Endgame practice http://www.lifein19x19.com/viewtopic.php?f=15&t=17832 |
Page 1 of 1 |
| Author: | lichigo [ Sun Oct 25, 2020 2:49 pm ] |
| Post subject: | Endgame practice |
Attachment:
File comment: What is the value of 1 with the new endgame theory? Studying endgame and I saw this shape today.20201025_223801.jpg [ 1.96 MiB | Viewed 20127 times ] If white play first black makes 0. If black play he makes 1point and 1 1/4point left right? So should I do 1 and 1 1/4 divided by 2= so 1 1/8? Expected territory is 1 1/8 and a move 7/8? |
|
| Author: | Bill Spight [ Sun Oct 25, 2020 4:38 pm ] |
| Post subject: | Re: Endgame practice |
The original position is, assuming all stones are immortal, worth ½ point of territory for Black, on average. Black to play can gain ½ point on average for 1 point of territory. White to play can hold Black to 0 territory. Just like the following position.  |
|
| Author: | lichigo [ Sun Oct 25, 2020 10:54 pm ] |
| Post subject: | Re: Endgame practice |
Thank you very much for your answer ^^ so with the robert jasiek's way I can say C=1 M=0.5 right? Sorry if it is not good 
|
|
| Author: | Bill Spight [ Sun Oct 25, 2020 11:28 pm ] |
| Post subject: | Re: Endgame practice |
The original territory count is ½ point for Black. White can play to L-18 or L-02 to reduce the count to 0, thus gaining ½ point. Black can play to L-18 or L-02 and temporarily gain more than ½ point, but White replies and keeps on replying, thus allowing Black to make only 1 point, for a gain of ½ point in the end. |
|
| Author: | lichigo [ Mon Oct 26, 2020 12:08 am ] |
| Post subject: | Re: Endgame practice |
So C=0.5 and M=1? Just starting the robert jasiek's books about endgame so I am not fully aware of all concepts of the new theory. |
|
| Author: | RobertJasiek [ Mon Oct 26, 2020 12:13 am ] | ||
| Post subject: | Re: Endgame practice | ||
The position is mentioned at https://senseis.xmp.net/?MiaiValuesList%2F000To099#toc5 where only the initial move value and count are stated. Bill gives another hint. However, we do not determine values by guessing. Instead, we must calculate systematically and, if necessary, verify that the calculated values are correct. There are different methods of doing so: - combinatorial game theory a la Mathematical Go Endgames using infinitesimals and cooling tax - graphical thermography a la Spight et al - thermography as linear algebra a la Spight - iterative calculation of counts, gains and move values (modern endgame theory described by various people including Spight and Jasiek) combined with verification by the method of making a hypothesis a la Jasiek - sometimes applicable, possibly much faster, sophisticated methods When long sequences are involved, I find the method of making a hypothesis the easiest, although an explicit verification is needed. Since the position is a standard shape, I have already solved it much earlier in the attachment. Needless to say, calculation of values is iterative and backwards: we determine the values of follow-up positions before the values of the initial position. Initially, we do not know whether the values of the initial position are derived from the follow-up position created by move 1, move 2 or move 3 of Black's alternating sequence. Longer sequences override shorter ones so we first verify that it is correct to calculate the initial values from Black's long alternating sequence. We find that these tentative values are confirmed because they are consistent with the (tentative) gains of the moves of Black's alternating 3-move sequence and White's 1-move sequence: the move value is at most each gain of the moves of both alternating sequences starting from the initial position. The values of "Black's long gote" endgame position are: count C = 1/2, move value M = 1/2, gain of Black 1 = Gb1 = 3/4, gain of White 1 = Gw1 = 1/2. EDIT: stating the values.
|
|||
| Author: | lichigo [ Mon Oct 26, 2020 12:22 am ] |
| Post subject: | Re: Endgame practice |
Thank you so much for your answer, the pdf is perfectly answering my question. Endgame problems 2 is coming ^^ |
|
| Author: | Bill Spight [ Mon Oct 26, 2020 1:38 am ] |
| Post subject: | Re: Endgame practice |
½ + ½ = 1. |
|
| Author: | lichigo [ Mon Oct 26, 2020 1:47 am ] |
| Post subject: | Re: Endgame practice |
Thank you very much. This endgame was harder to understand that I expected but it seems clear now ^^ I think endgame problems 2 will be interesting if there are many problems like today. |
|
| Author: | RobertJasiek [ Mon Oct 26, 2020 3:05 am ] |
| Post subject: | Re: Endgame practice |
lichigo wrote: I think endgame problems 2 will be interesting if there are many problems like today. Answer: https://www.lifein19x19.com/viewtopic.p ... 75#p260975 |
|
| Author: | Bill Spight [ Mon Oct 26, 2020 4:21 am ] |
| Post subject: | Re: Endgame practice |
lichigo wrote: Thank you very much. This endgame was harder to understand that I expected but it seems clear now ^^ I think endgame problems 2 will be interesting if there are many problems like today. Yes, it is not so easy to understand. We are used to the hane-and-connect, which is a 3-move sequnce, but the 3 move sequence of this position is not so obvious.
|
|
| Author: | lichigo [ Mon Oct 26, 2020 12:34 pm ] |
| Post subject: | Re: Endgame practice |
I have some trouble to understand the pdf, my friend and I we think C= 5/8 and the M=5/8 . Why is it wrong because your answer is both 1/2. |
|
| Author: | RobertJasiek [ Mon Oct 26, 2020 2:40 pm ] |
| Post subject: | Re: Endgame practice |
Please tell us how you calculate your values, then we might identify your calculation mistakes! The PDF is a bit dense but you might state where you think you lose track in its description. |
|
| Author: | lichigo [ Mon Oct 26, 2020 2:53 pm ] |
| Post subject: | Re: Endgame practice |
We thought when Black plays 1 that we need to divide the next shape by 2. 1 1/4 divided by 2 but it seems wrong. And in the pdf Gb3 doesn't seem clear. And I am not familiar yet with the gain. I am gonna check Endgame 2-values Gain chapter tommorow ^^ |
|
| Author: | RobertJasiek [ Mon Oct 26, 2020 5:10 pm ] |
| Post subject: | Re: Endgame practice |
lichigo wrote: We thought when Black plays 1 that we need to divide the next shape by 2. 1 1/4 divided by 2 but it seems wrong. You find the answer here: https://www.lifein19x19.com/viewtopic.p ... 99#p260999 Quote: And in the pdf Gb3 doesn't seem clear. And I am not familiar yet with the gain. I am gonna check Endgame 2-values Gain chapter tommorow ^^ Understanding of how Black's or White's gains are calculated is mandatory in general and presumed in the PDF. Gb3 is the gain of a black move, namely move 3 of Black's alternating sequence. Its calculation Gb3 = B - C5 is the count B of the position after the move minus the count C5 (where the index 5 refers to the diagram subnumber 5 of Dia. 73.5) of the position before the move. |
|
| Author: | lichigo [ Tue Oct 27, 2020 12:54 am ] |
| Post subject: | Re: Endgame practice |
Thank you for the informations. |
|
| Author: | Bill Spight [ Tue Oct 27, 2020 3:55 am ] |
| Post subject: | Re: Endgame practice |
See here ( https://www.lifein19x19.com/viewtopic.p ... 03#p261003 ) for the thermographs of this position and positions in its game tree. Thermography is a method for finding the count of a position, how much each play gains, and possibly other information about it. |
|
| Author: | Bill Spight [ Sun Nov 01, 2020 5:48 am ] |
| Post subject: | Re: Endgame practice |
An easy one, but one that most people may not think about. How much does  gain, on average?
|
|
| Author: | lichigo [ Sun Nov 01, 2020 7:47 am ] |
| Post subject: | Re: Endgame practice |
I will say 2/3 of a point. |
|
| Author: | Bill Spight [ Tue Nov 03, 2020 8:07 am ] |
| Post subject: | Re: Endgame practice |
By filling at B-19 and then at A-18 Black reaches a local score of 0 in the corner in 2 plays. White to play takes the ko with sente and after Black fills at B-19 to save two stones White can fill the ko for a local score of -1 with 1 net play. As with a simple ko the difference between winning it and losing it is 1 point in 3 net plays. The average gain per play is thus 1/3 point.
|
|
| Page 1 of 1 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|
