Gato Go: “This photo (left) is from the Facebook page of one of the Ecuadorian players I met in Quito,” writes Bob Gilman.
In a surprise victory, Murakawa Daisuke 8P defeated Iyama Yuta 9P on December 16 to win the 62nd Oza. The Oza is Murakawa’s first major title. This is the first time a player from the Kansai Kiin has won the Oza since Hashimoto Shoji 9P did so 33 years ago in 1981. The final game was played in Toba, Mie Prefecture, Japan. The upset attracted a lot of attention in Japan, because former Oza Iyama Yuta currently dominates the domestic Japanese go scene.
- excerpted from Go Game Guru; click here for the full report, including game records.
A new go school in Singapore welcomes students from the United States. The Go Academy offers a wide range of classes, from a 1-day introduction to kyu and dan intensive camps, reports Director Daniel Chan. More info available on Facebook.
“It’s interesting to read about the work of University of Edinburgh to use machine learning to improve the level of playing in computers, (Scottish Neural Network Takes Computer Go to Next Level 12/16/2014 EJ)” writes Nin Lei, Distinguished Engineer and CTO, Analytics and Big Data, STG IBM Systems and Technology Group. “However, the title in their article creates an impression that their research is creating a program that can beat the best human players. If their probability of guessing their next move is only 44%, then their chance of guessing it wrong is 56%. In a sequence of 10 moves, the chance of getting the complete sequence correctly is 0.44 ** 10, which is a very small number.” Noting that checkers “has been solved via machine learning,” Lei says that “it appears it is promising for go as well.” But because machine learning predicates that there is a pattern in the underlying data set, Lei warns that “it could be so complex that machine learning can only attain a certain level of accuracy. It seems to me a program needs to have very high level of accuracy before it can play a good game at strong human level.” Lei also says that “Since machine learning is based on pattern recognition, I wonder if a professional can trick the program by using moves that may not be optimally locally but will create patterns that the program has not seen before. I applaud the work they are doing,” Lei concludes. “It is innovative by using a different approach than the existing strong computer programs. It will be interesting to find out if someday they can come up with an algorithm that can improve the accuracy significantly.”
12/22: the chance of getting the complete sequence correctly has been corrected to 0.44 ** 10 (from 0.56).
Late in the afternoon of December 17, go players at the 2014 SportAccord World Mind Games attended their second awards ceremony in the Beijing International Convention Center. This time the awards for pair go were given out. China’s Yu Zhiying and Mi Yuting collected their second gold medals, Korea’s Choi Jeong and Na Hyun collected their first and second silver medals, respectively, and China’s Cathy Chang (more formally, Chang Kai-Hsin) and Lin Li-Hsiang received their first bronze medals.
Medals were also awarded for individual contract bridge, Basque system chess, super-blitz draughts, checkers, and xiangqi. Four countries picked up their first gold medals here: Monaco, for whom Geir Helgemo came through in open bridge; Cameroon, whose draughts star Jean Marc Ndjofang won the men’s superblitz; Vietnam, whose Ngyuen Hoang-Yen shone in women’s xiangqi; and Italy, which proved to have the world’s top two checkers players.
All told there were 24 separate events in this year’s world mind games, and mind athletes from 20 different countries and territories won medals. Between them, China and Russia took the gold medals in half the events, winning six each. Russia outpointed China in silver medals, but China outpointed Russia in bronze, won medals in a greater number of disciplines (all but draughts), and won the greater total number of medals – by a wide margin if China’s medals in the team and pair events in go and bridge are counted as multiple medals. All of Russia’s medals came in individual competition in chess and draughts. Full results and further details can be found on the SportAccord schedule and results page and news page.
At the closing ceremony at the V-Continent Beijing Parkview Wuzhou hotel, Mr Hai Zhenwen, deputy secretary general of the organizing committee, praised the successful conclusion of the four-year series of SportAccord World Mind Games in Beijing. China’s former ace weightlifter Ma Wenguang, representing SportAccord Asia/Pacific, thanked the city of Beijing and expressed a hope that world mind games would continue elsewhere in the Asia-Pacific area in the future. Mr Vlad Marinescu, Director General of SportAccord, said that he had been humbled at finding himself in the midst of so many geniuses but inspired by the enthusiasm of Beijing’s children, and noted that on the publicity front, this year the games had achieved a 50% growth over last year on all media platforms. Mr Chen Jie, deputy director of the Beijing Municipal Bureau of Sports, thanked the organizers for their hard work, after which he and Mr Marinescu exchanged gifts, and then everyone settled down to a good dinner.
- James Davies
This is the game in which China’s Yu Zhiying and Mi Yuting won the gold meal in pair go by defeating Korea’s Kim Choi Jeong and Na Hyun. Click here for the sgf game file.
Black 37 (A in diagram 1, played by Choi) defends against a double peep at B, but lets white jump to C. Black is playing too safe. She should have capped at 1 in diagram 2. If white peeps, pushes, and cuts as shown in the diagram, black can give up two stones, after which white will be unable to reach the center. Then black can develop on a truly large scale by pressing white down on the lower side as shown (moves from 11 to 15 in diagram 2).
Black 43 (played by Na) was also too conservative. Black should have occupied white 52, the junction point of his framework on the right side and the white framework on the lower side. When white got to play 52, the Chinese pair had a territorial lead.
White 64 may have been an overextension. Black 65 immediately started to threaten white’s thin position. With 66, white began dancing around to protect the weak white stones in this area and the weak white group at the top. But white danced successfully. By the time white played 92, white’s weak stones and group were out of danger.
Black 111 (Na) may have been the decisive mistake. Before playing here, black could have made sente moves at 112 and 150. Instead, white was able to play 112 and 148-150 in sente, gaining approximately seven points.
Black’s last chance was to start the ko at 157, but black lacked the necessary ko threats. Black 159 was inadequate. White simply ended the ko with 160, gaining as much at the top as was lost in the bottom left corner. When black renewed the ko challenge at 183, white accepted by cutting at 184. When Yu took the ko with white 194, Na decided that black’s prospects were too poor to continue and offered to resign, and his partner agreed.
- Ranka, based on commentary by Michael Redmond 9p.
This is the game in which Chinese Taipei’s Chang Kai-Hsin (Cathy Chang) and Lin Li-Hsiang (White) won the bronze gold meal in pair go by defeating Japan’s Fujisawa Rina and Ida Atsushi (Black). Click here for the sgf game file.
The opening pattern up to white 8 has become very popular quite recently and appeared in several SportAccord games.
Up through black 71 (the marked stone in diagram 1) the position is about even. White appears to have more territory, but the white position on the lower side is thin and can easily be invaded, so black is not behind. In fact, if black had played 71 at 1 in diagram 2, black might well have been ahead. White cannot cut black apart. Given the continuation through black 9, black is solidly linked up and the white group cannot make two eyes on the right side, so black will be able to attack it in the center.
The move that Ida Atsushi chose for black 71 was readily answered by white 72, and black 73 let white break up black’s right side territory in sente with 74 and 76. Aside from losing territory, black was placed on the defensive, and would be hampered in going deeper into the white territory at the bottom.
From this point on, the fighting in the center became quite confused. White’s top group seemed to be in trouble, but it could never be killed unconditionally, except at the sacrifice of an even larger black group. When white 168 connected the ko in the top left, white had a clear territorial lead.
By white 186 black’s position had become untenable and the Japanese pair resigned.
- Ranka, based on commentary by Michael Redmond 9p.
Another series of classic go books has just been released by BadukTV, reports Shawn Ray (Clossius). The first set of four books was “The Profound and Mysterious,” a life-and-death exercise book written between 1347 and 1349, during the Yuan dynasty. The second and just-released series is “The Art of Closing,” a 6-book set filled with ancient problems put together by previous masters and translated by Cho Hye-yeon. “It is a level below that of ‘The Profound and Mysterious’ so players 5-kyu and stronger should be able to benefit greatly from it,” Ray tells the E-Journal. “Though I think anyone can take a lesson or two from it.” Click here to buy both as a bundle of all 10 of the books. For more info e-mail Ray at Clossius.ShawnRay@gmail.com
This is the game in which the Russian brother-sister pair of Ilya Shikshin and Svetlana Shikshina, both former European champions, defeated Russia’s Natalia Kovaleva and France’s Fan Hui to take fifth place in the mixed pairs competition. Click here for the sgf game file.
White 8 to 16 are one variation of a popular joseki. This variation and others appeared several times during the week of World Mind Games.
White 38 (the marked stone in diagram 1) was a mistake. The critical issue here is the relative strength of the groups in the center. If the white pair had played as shown in diagram 2, their own center group would have had the upper hand, and they would then have been free to deal with the loose black framework on the right side.
If black had captured white 98, the white group in the center would have been in serious trouble. Apparently Ilya decided that invading the bottom right corner first would be a safer way to win, but this is not necessarily true. If white had played 123 in sente before black did so, then given the same continuation on the rest of the board, the final margin would have been only half a point. As it was, black won by 1-3/4 stones, or 2-1/2 points.
- Ranka, based on commentary by Michael Redmond 9p