The Way to 9x9 is a book by Monteo da Westi, author of The Art of 9x9 Go: How to Beat Dan Players at GoQuest. His study on 9x9 has been conducted since 1995 by playing with AIs and Go players at Yahoo Games and GoQuest, totally around 10,000 games. It took around 10 years part time before breaking through Shodan in 2016. So, to reduce kyu players’ retention time, this book describes 9x9 from basic concepts and principles to advanced strategies and tactics.
|Table of contents|
Books that deal with both basic and advanced knowledge of 9x9 do not exist. This book is an effort to fill this gap.
The origin of 9x9 has long been unknown for sure. Nobody knows exactly when the game was invented and by whom. The oldest 9x9 game records have been said to be of the two matches by Go Seigen and Miyamoto Naoki in 1968. By then, the most common board sizes are 8x8, 13x13, 17x17, apart from 19x19. That is probably why Karl Baker uses the 8x8 board size as illustrations in his classic 1986 book, The Way to Go. In Great Britain, Francis Roads noted in 1971 that they have experienced playing on the 13x13 and 10x10 boards. These facts confirm that the 9x9 is not common prior to 1970.
Popularity of 9x9 in Japan begins with the MiniGo matches broadcasted every Sunday morning from 1987 to 2002 by Yomiuri TV.. In the global scale, 9x9 has gained in popularity since the late 1990s, after David Fortland published Igowin for Windows—the strongest free 9x9 Go program for the time being. In the early 2000s, 9x9 even more popular when players can play 9x9 with stronger players at Yahoo Game, where there is a 9x9 Go Room for both beginners and advanced players. In smart phone age, since 2010, a more polular 9x9 Go server has become GoQuest by Dr. Yasushi Tanase. More than 30,000 names are active 9x9 players there, with around 25,000 kyu players (Elo<1700). Many of them have played more than 10,000 matches and still cannot break through the kyu boundary, even the minimal rating requirement for Shodan promotion has been relaxed from 1750 to 1700.
Go is a game of war. All equipments and actions in the game have real world equivalents during at war.
- Board, made of paper, wood, stone, or computer screen, represents the earth or terrain where the war between the two sides will be taking place.
- Visible lines, nine vertical lines and nine horizontal lines, which represent the road that soldiers (stones) can move and settle a camp at each intersection (point), to subsequently form a group (2 or more stones connected) and an army (2 or more groups working together).
- Invisible (or border) lines, including
- The imaginary diagonal line slant from the bottom left to the top right of the board, which marks the shared boarder line that divides Black’s and White’s homelands. That is why classy players, like Cho Chikun, when having Black, always open the game within the upper left zone, which is Black’s homeland. Uneducated players, including some AIs, do not follow this tradition and often impolitely play within White’s homeland (the lower right zone). Hence, you can see whether one is an educated player or not by observing his or her first move when having Black.
- The fifth vertical line represents the Milky Way above the sky that links the Earth (Origin or Gen) and the Heaven (Ten). Land on the left side of the Milky Way is the East region, and the right side the West region.
- The fifth horizontal line marks the central region of the two countries. Land above this line is the North region, and below this line the South region. Hence, the majority of Black’s homeland is located in the northwest region, and the majority of White’s homeland the southeast region.
- Edge, the outermost square, replesents the great mountain range that encloses all the battle fields in the war zone of 81 points.
- Marks, symbols in the go board, represents a military map where strategy and tactics are made explicit for discussion and making agreement among the generals and consultants.
- Within board
- Stones include - and -
- Star points are the five solid dots on the board at 3-3, 3-6, 6-3, 6-5, and 5-5, representing the key stars above the clear terrain, which help the soldiers and the commanders know easily where they are or where to go. With the knowledge of these stars, it is easier to perform a fast manoeuvre.
- Triangle includes
- Square includes represents potential territory and stones in danger.
- Cross includes
- Circle includes
- Text includes a - z and A to Z
- Outside board Integer and text are used to indicate the position for any move or any point in the board, representing geographic coordinates of the earth surface or the names of towns or cities in the war zone. There are three coordinate systems:
- Japanese: using only integer from 1 to 9, running from left to right and top to bottom along the edge.
- SGF: Using only small letter from a to i running from top to bottom and from left to right. So, the positions in the board are referred to as aa, ab, ..., and ii.
- Within board
- Stones represent soldiers or troops of the two countries:
- Black stones represent the soldiers that occupies the land first. Black has a chance to develop the land before White. So, a weaker player is allowed to have Black and to play first.
- White stones represent the soldiers that occupies the land after Black one time step.
- Bowls represent the two countries from which the soldiers (stones) are sent to the war zone (board) by the central governments (Go players) after observing the whole situation and made decisions to develop the war.
- Clock represents a war season or duration of the war affair.
- Game recorder represents a history book from which we can learn from mistakes and successes, and improve our strategy and tactics so that we will be well prepared for the next war.
- Matching: in theory, anybody can challenge anybody. The weaker take Black. The much weaker are allowed to place 2-5 handicap stones at the star points before the first move. In my opinion, giving proper Komi to Black, without handicap stones will enable a more realistic fight. In practice, at GoQuest, you cannot challenge anyone except your friend in a non-rating game. The server will select an appropriate opponent for you. You can only choose to play with human players only or to play with AIs (sometimes or often). The concept of matching is simple. If you want to know how strong you are, play a few matches with a comparable player. If you want to know whether you are the best player in the world, challenge the world number one player. It is quite regretable that no one knows for sure who is the best human 9x9 player, although my analysis shows that he is Yamada Kimio.
- Side selection
- The first move
- Legal moves
- Game objective: To be the winner, or to beat the opponent (ideally all opponents), by placing stones on the board to gain more points than the opponent AFTER the last move. To do so, stronger players often do deeper and more accurate calculations of potential points ahead (for human players) or %wining (for AI players), which involves mental (or computer) simulation of best-move projections and evaluation of the resulting potential scores for both players.
- Scoring: Go players need to evaluate the score for themselves throughout the game. The evaluaion can be done by many methods, from simple counting to sophisticated %wining calculation, all of which are based on the same information observed from the game (living stones, dead stones, prisoners, potential territory) and the initial conditions (total space and komi). Many computer go programs have been developed to calculate %wining, by comparing Black's potential territory with White's potential territory. If %wining = 50, then Black and White will gain equal points (komi included). Human players don’t calculate %wining because it is a time consuming process. They quickly do a rough count of the territory and spend more time thinking about the next good moves and choose the one they believe the best or they prefer.
- Talking: Top go players don’t talk during the game. This is a tradition of classy go players. They talk by hands—the way they play tell something that can often be better understood after the match. That’s why GoQuest doesn’t provide a chat box service to Go players.
- Stone collection
- Joseki results in balance potential territory, which results from symmetrical moves of both players at the opening.
- Fuseki. There are 200 most common 9x9 positions and the move that has best win rate with 95% confidence, according to Henry Hemming. Among them are the 40 well known Fuseki, listed from the highest to the lower ratings (with the best player of each pattern in parenthesis), to give you some ideas on the best opening patterns and the best openners, as follows:
- VS Tengen (Linbee R2806). Linbee has never been defeated by the opponents’ Tengen opening. He uses Komoku (4-3) to react Tengen.
- Tengen (Linbee R2795)
- Hoshi (Linbee R2793)
- Takamoku (Saturn777 R2764)
- New Orthodox (Saturn R2704)
- VS Hoshi (Linbee R2678)
- VS Takamoku (Linbee R2672)
- Soccer Juggling (HippolytaX R2664)
- Windmill (HippolytaX R2659)
- Almost Equilateral1 (bluedragon R2655)
- Slider (Bambi17 R2631)
- Bean Throwing (HippolytaX R2606)
- Sword (HippolytaX R2591)
- VS Moku-hazushi (HippolytaX R2585)
- VS Komoku (HippolytaX R2582)
- Jump Attachment! (Bambi17 R2572)
- Komoku (naotin15 R2561)
- White Slice (HippolytaX R2559)
- Black Boomerang (HippolytaX R2557)
- Moku-hazushi (bluedragon R2555)
- Hand Fan (bluedragon R2545)
- Secret Agent033 (HippolytaX R2540)
- Almost Equilateral2 (HippolytaX R2526)
- Orthodox (doctory R2529)
- Boots (Bambi17 R2521)
- Pendulum (hebogo15 R2506)
- Curveball (HippolytaX R2504)
- Flower (HippolytaX R2468)
- Almost Equilateral3 (HippolytaX R2460)
- VS Sansan (tata01 R2447)
- Andromeda (HippolytaX R2402)
- Sansan (HippolytaX R2401)
- Lunar Eclipse (doctory R2395)
- Side Attachment! (Miyumiyumichan R2334)
- Sea Fairy (Krabicka6D R2243)
- Flower Fairy (diem R2243)
- Zazen (Wing1973 R2220)
- Cross Line (spaoh12 R2152)
- Kodachi (sho1sho2sho3 R2150)
- Head Butt (mysr R2150)
- Counting: Like AlphaGo, you need to find the best move every moment during the game. First, you observe all the stones on the board and have a set of good moves in your mind (policy choices). Second, you do some game variations (optimal plays for both players) some moves ahead according to your ability (mental simulation). Then, you see the territorial and initiative impacts of your choice at the end of each variation (policy evaluation). Finally, you pick the move that best fits your strategy (policy recommendation). Black’s situation can be calculated by human and not by AIs as follows:
- which notes the number of points Black leads White for the time being. Minus situation means White leads Black and is thus bad for Black.
- Human players
- Go Seigen opens exclusively on the 4-4 point when having Black, to set the corner more quickly. He loves sente at the end of sequence, by playing contact moves with the aim of keeping the initiative. So, he gets around the board more quickly and could reach the big points first. He has an exceptional ability at converting thickness into profit; a keen positional judgement backed by accurate reading and a special intuition; and a penchant for rarely losing a ko fight that he initiated. He plays much more quickly than his opponents.
- Miyamoto Naoki encourages reading a couple of moves and evaluating the global position, instead of relying on proverbs. He puts a lot of emphasis on thickness and seeks a simple joseki that fits the global position, rather than complications. He strongly discourages entering at the 3-4 point when White has opened at the 5-4 point but to invade at the 3-3 point instead as soon as White has made two extensions.
- Cho Chikun tries to match strength with strength, lightness with lightness. His ideal is to play in such a way that no one can tell who the player was. He has no special preference in the fuseki and if possible he would like to master every style.
- Ma Xiaochun plays light, airy, elegant and dodging style, which goes against a fierce, combative style. His airy style of play would continue to give attacking players a difficult time.
- Takemiya Masaki is famous for his 'cosmic style' of play - focusing on the construction of large moyos and taking territory in the centre. He calls his style 'natural style', and considers that the centre-oriented style for him is the natural way for his stones to move and focuses on'playing with your heart' and playing moves which 'please'. He likes Dosaku’s style, and playes modern opening strategies. He plays very slowly.
- Cho Hunhyun plays complex moves that requires natural brilliance but quick thinking, which is contrast to a simple style with magnificent calculation and deep reading.
- Yuki Satoshi has a dynamic and flexible style, which is very effective in 9x9 because the game situation changes drastically after each move.
- Yamada Kimio is a territorial player who is adept at invading and living within opponents’sphere of influence.
- AI players
Modified meta-analysis was conducted using data from MiniGo matches and tabulated as Synthetic Tournament, which gives an idea on the best 9x9 human player in history. It turns out that the Champion is Yamada Kimio; the Runner-up is Cho Chikun. Yuki Satoshi wins the third place and Cho Hunhyun the fourth place.
|A||Yamada Kimio 8p (B+3.5)|
Cho Hunhyun 9p
|Yamada Kimio 8p (W+6.5)||Yamada Kimio 5p (B+R)|
|B||Takemiya Masaki 9p (B+2.5)|
Yuki Satoshi 9p
|Takemiya Masaki 9p|
|C||Cho Chikun 9p (B+R)|
Yata Naoki 6p
|Cho Chikun 9p (B+T)||Cho Chikun 9p|
|D||Yamada Shiho 9p|
Kurahashi Masayuki 6p (W+2.5)
|Kurahashi Masayuki 6p|
|Group||Play off||3rd Place|
|E||Yuki Satoshi 8p (B+R)|
Yamada Shiho 9p
|Yuki Satoshi 8p (B+T)|
|F||Cho Hunhyun 9p (B+T)|
Yata Naoki 6p
|Cho Hunhyun 9p|
 The game records can be downloaded at gobase.org
 Noted in 1969 by Scott A. Boorman in The Protracted Game: A Wei-Chi Interpretation of Maoist Revolutionary Strategy.
 For more information, see British Go Journal , April 1971 issue, Page 11.
 The game records can be downloaded from Professor Andries E. Brouwer’s MiniGo page
 The openings in boldface will be explored in details to demonstrate how to beat dan players with them.
 The 10 game records will be analyzed in The 9x9 Grand Slam: Game Analysis with Deep Learning Technology. This technology has been validated by analyzing the fourth game between AlphaGo Vs Lee Sedol; the board situations (%wining) are well consistent with those of AlphaGo’s analysis: AlphaGo’s %winning drops at move 79, and get worse and worse until after Move 105, and get better and better from Move 131-141. Hence, it is very reliable for use in 9x9 game analysis. For example, bad moves can be identified in the Final, which leads to Yamada’s victory.