Fact-checked by Grok 2 weeks ago

Go variants

Go variants encompass a wide array of modifications to the ancient East Asian board game known as Go (Weiqi in China, Baduk in Korea, Igo in Japan), which fundamentally alter elements like board size, initial stone placement, player count, capture rules, or victory conditions while retaining the essence of strategic encirclement and territorial control.^[1] These adaptations range from ancient regional traditions to contemporary inventions, often played as side events in tournaments or for educational purposes to introduce newcomers to Go's mechanics.^[2] Historical variants emerged primarily in East Asia, reflecting cultural and regional evolutions of the game dating back over 2,500 years. In Korea, Sunjang Baduk, which developed during the 7th century, begins with 16 pre-placed stones in a central box formation on a standard 19x19 board, and stones inside claimed territory are removed before scoring to simplify boundary determination, emphasizing aggressive play over long-term enclosure.^[1] Similarly, Tibetan Go uses a smaller 17x17 grid with six stones per player pre-positioned on the third line from each edge, allows White to move first, enforces a strict ko rule prohibiting immediate recapture, and awards bonus points—40 for controlling all four corners and 10 for the center—to incentivize bold territorial grabs.^[1] These early forms highlight how Go's rules were not rigidly standardized until the modern era, allowing for localized interpretations that influenced strategy and teaching methods across China, Korea, and Tibet.^[3] Among the most popular modern variants are those designed for accessibility, teamwork, or altered objectives. Atari Go, also called First Capture Go or Capture Go, simplifies the game by declaring the first player to capture an opponent's stone(s) the winner, eliminating territory scoring and pass rules to focus on tactical encirclement and atari threats, making it an ideal introductory variant for beginners.^[4] Rengo extends Go to team play with two players per side (Black and White teams), where teammates alternate moves in sequence without communication, often incurring a small penalty (like three captured stones) for out-of-order plays, fostering collaborative strategy on a standard board.^[5] Pair Go, a variant of Rengo, pairs one male and one female per team, promoting mixed-gender participation and is frequently featured in international tournaments.^[1] Multi-player Go accommodates three or more participants using additional stone colors on a single board, adapting capture and territory rules to prevent alliances and handle complex interactions.^[1] Other notable innovations include board topology changes and scoring tweaks. Toroidal Go (or Daoqi) plays on a board where opposite edges connect, creating a wraparound effect that eliminates edge advantages and alters approach strategies, often on smaller grids for computational study.^[6] One Color Go challenges players to use identical-colored stones for both sides, relying on position and memory to distinguish groups, which heightens the importance of connectivity.^[1] Environmental Go, developed by mathematician Elwyn Berlekamp, integrates a deck of cards valued from 1/2 to 20 points that players can claim instead of placing stones, blending probabilistic elements with traditional placement to explore game theory.^[1] These variants, alongside niche forms like Blind Go (memory-based without board visibility) and Paper-and-Pencil Go (marking captures instead of removing stones), demonstrate Go's adaptability, supporting educational tools, casual play, and research into combinatorial game theory.^[1]

Regional variants

Tibetan Go

Tibetan Go, known locally as mig mangs (མིག་མངས་, meaning "many eyes"), is a traditional board game adapted from the ancient Chinese game of Go, with roots possibly tracing back to pre-Buddhist Bon traditions in Tibet or an introduction from China around the 7th century CE.^[7] Archaeological evidence, including a stone board discovered in Lhasa, supports its long-standing presence in Tibetan culture, though exact origins remain debated between parallel indigenous development and external influence.^[8] Unlike the standard 19×19 grid used in Chinese, Japanese, and Korean Go, Tibetan Go is typically played on a 17×17 board with 289 intersections, though some variants use 15×15 boards; these were often made from cloth for portability or carved from wood and stone for durability in high-altitude regions.^[7]^[9] The rules closely resemble standard Go in terms of stone placement, territory enclosure, and capturing by surrounding opponent stones, but feature distinct modifications suited to local play. A relaxed ko prohibition allows recapture only after waiting one full move, preventing immediate snapbacks while simplifying enforcement compared to stricter positional superko rules elsewhere; this rule states that a player may not immediately play on an intersection from which the opponent has just removed a stone.^[8]^[7] Scoring incorporates area or territory methods with bonuses, such as 5 points for controlling the central intersection (tengen) and a 20-point penalty for losing both starting corner positions if the opponent retains theirs, emphasizing strategic control of key points over exhaustive territory accumulation.^[8] Games begin with a fixed opening where White moves first, often starting near the corners, and play proceeds with moves required to be within one intersection of the previous one in some traditions, fostering a more contained and tactical style.^[7] Culturally, Tibetan Go held significance among Tibetan aristocracy as part of the Bon tradition's "Nine Activities," serving as a tool for intellectual and strategic training, though it was generally prohibited for Buddhist monks due to gambling associations; nevertheless, some monastic communities adopted it discreetly for meditative and divinatory purposes, where black and white stones (rdel) doubled as astrological tools to interpret fortunes or resolve disputes.^[7]^[9] Its ritualistic elements tied it to broader Tibetan cosmology, including dualistic concepts in the "Great Wheel of Time," making it more than a mere pastime but a vehicle for philosophical reflection.^[7] In the modern era, Tibetan Go has seen limited revival efforts since the 2000s, driven by international Go organizations amid the decline of traditional players following the Cultural Revolution and Tibet's political upheavals.^[7] Scholarly research, including a 2003 seminar at Oxford University organized by Go historians, and documentation in databases like GoGoD have helped preserve and clarify rules, such as the Tibetan ko variant, fostering renewed interest among global Go enthusiasts.^[8] Despite challenges like restricted access to Tibetan regions, these initiatives have introduced the game to wider audiences through translations and online resources.^[7]

Sunjang Baduk

Sunjang Baduk, also known as an ancient form of Korean Go or Baduk, originated in Korea during the 16th century and remained the predominant style until the mid-20th century, when it largely faded due to the adoption of Japanese-influenced rules.^[10] This variant emphasizes haengma, or the strategic flow and influence of stones, over the pure enclosure of territory seen in modern Go, as the preset opening forces players to develop positional strength from established bases rather than an empty board.^[11] It was played extensively during the Joseon Dynasty (1392–1897), with surviving game records dating primarily from the 1880s, including an early documented match in 1937 between players No Sa-ch'o and Ch'ae Keuk-mun.^[10] The game is played on a standard 19x19 board, but begins with a unique preset position consisting of 16 stones—eight for each player—arranged in a symmetrical, box-like formation along the fourth line, simulating initial weak groups that demand careful nurturing.^[12] Black makes the first free move, traditionally at the tengen (center point), after which play proceeds similarly to standard Go, including rules for ko and seki, though traditional games lacked komi, with later variants introducing 4.5 or 6 points for White in even games.^[12] This setup leads to an aggressive middlegame focused on connecting and expanding influence from these bases, promoting positional judgment as each move interconnects with the existing structure to maximize overall board control.^[11] Scoring in Sunjang Baduk employs a variant of territory rules, where neutral points (dame) are filled, dead or internal stones are removed to clarify boundaries, and only the surrounded empty intersections count as territory for each player, with prisoners ignored and no points awarded in seki positions.^[13] Boundary stones must not be in atari at the end, ensuring stable influence, while plays inside one's own territory incur no penalty, akin to aspects of Chinese scoring but retaining the precision of territory evaluation.^[13] This method rewards surrounding influence by valuing the effective control of space through connected groups, rather than solely enclosed areas, and results in games that highlight strategic depth in influence development over territorial accumulation.^[14]

Batoo

Batoo is a Korean variant of the board game Go, combining elements of traditional play with innovative mechanics designed to appeal to younger players through online formats. Emerging in November 2008 as an attempt to adapt Go into a more dynamic videogame experience, it derives its name from "baduk" (the Korean term for Go) and "juntoo" (meaning "battle"), emphasizing strategic confrontation.^[15] The game was exclusively played online on dedicated servers, which have since become defunct, though efforts to revive it on new platforms continue.^[16] This digital focus allowed for portable play on computers and mobile devices, adding strategic depth through hidden elements and competitive bidding in fast-paced settings.^[17] The game is typically played on an 11x11 board, though 13x13 variants exist, featuring special plus-point and minus-point locations that award or deduct five points when occupied.^[15] Before the main game begins, each player places three base stones simultaneously on designated spots, each worth five points; overlaps result in penalties to encourage thoughtful positioning.^[15] A key unique mechanic is the bidding system for the first move, where players wager points to determine turn order, introducing an element of risk and psychological strategy akin to komi adjustments in standard Go.^[17] Additionally, each player has access to one hidden stone per game, placed secretly and invisible to the opponent until revealed by a "scan" (costing two points) or by the opponent playing on its location, which adds uncertainty and reconnaissance tactics to the gameplay.^[15] Moves are constrained by a 25-second time limit per turn, with penalties of two points for timeouts and a maximum of three such infractions allowed, promoting quick decision-making suitable for mobile sessions.^[15] Capture and placement follow core Go rules, but the hidden stone mechanic necessitates online implementation or a referee for over-the-board play to verify secrecy.^[17] Scoring integrates traditional territory control with modern twists: one point per captured stone, five points for base, plus, or minus spots, and the total enclosed territory at the end, combining positional control with the value of hidden reserves and bids.^[15] Optional features include customizable characters with special abilities (three free, one purchasable via in-game currency) and themed boards or colored stones, enhancing replayability.^[15] Rankings range from level 5 (beginner) to 35 (expert), fostering competitive progression.^[15] Despite its innovative approach, Batoo saw limited adoption and server shutdowns, but community interest persists through discussions on revival and implementation on platforms like Online-Go Server.^[16] This online-centric preservation highlights its adaptation for contemporary, mobile-friendly strategic play, distinguishing it from purely physical regional variants.^[17]

Variants modifying core gameplay rules

Stoical Go

Stoical Go is a variant of the board game Go that replaces the standard ko rule with a stricter prohibition on captures to eliminate all known infinite cycles and promote strategic patience.^[18] Designed by abstract game designer Luis Bolaños Mures in September 2012, the variant aims to simplify repetition prevention while preserving much of Go's core mechanics.^[19] The key rule modification is that a player may not make any capture on their turn if the opponent captured exactly one stone on the previous turn; all other rules, including board setup on a standard 19×19 grid, stone placement, group capture by liberty deprivation, and area scoring (territory plus prisoners), remain unchanged from standard Go.^[19] This effectively extends ko resolution by forcing players to respond elsewhere on the board before attempting a recapture, thereby discouraging hasty responses and encouraging broader positional play.^[19] By generalizing the ko ban to any single-stone capture, Stoical Go renders traditional ko fights and related cycles impossible without needing complex superko rules, though it disallows certain snapback captures that are legal in standard Go. The variant is available for play through digital implementations like Zillions of Games software, making it accessible for experimentation.^[20]

First capture

First capture is a variant of Go in which the objective is changed such that the first player to capture any stone of the opponent immediately wins the game. This simplifies the rules by eliminating territory scoring and endgame considerations, focusing solely on tactical capture mechanics. The game is played on a standard 19x19 Go board, with players alternating turns to place one stone of their color on empty intersections, following core Go placement rules.^[21] Capture occurs when a group of connected opponent stones is surrounded and left without liberties, as in standard Go, but the game terminates upon the first such event, often resulting in short, intense matches with hyper-aggressive openings aimed at immediate threats.^[22] Invented by Japanese professional Go player Yasutoshi Yasuda in the early 2000s, first capture—also known as Atari Go or the capture game—was developed as part of his innovative teaching methodology to introduce beginners to Go's fundamental concepts.^[21] Yasuda detailed the variant in his 2002 book Go as Communication, emphasizing its role in making the game accessible without overwhelming new players with complex scoring systems.^[23] The variant gained traction in online Go communities during the 2000s, where it became favored for quick, casual games due to its brevity and emphasis on direct confrontation rather than long-term strategy.^[21] In educational contexts, first capture serves as an effective tool to teach basic surrounding and capture tactics, allowing novices to grasp liberty concepts and group connectivity early without needing full game rules.^[24] Yasuda's approach, often called the Yasuda method, has been adopted by Go organizations worldwide, including tours by professionals to schools and clubs, where it helps build intuitive understanding of threats and responses.^[25] A related variant, sometimes referred to as reverse first capture, alters the win condition so that the player who avoids having their stone captured longest prevails, promoting defensive play over aggression, though it remains less common.^[22]

Tenuki Go

Tenuki Go is a modern variant of the board game Go that introduces a restriction on player responses to promote strategic tenuki plays, or moves elsewhere on the board away from the opponent's most recent placement. The name derives from the Japanese term "tenuki," which refers to ignoring an opponent's local development to pursue opportunities in another area. In this variant, the core rules of Go remain intact, including stone placement, capture mechanics, and standard territory-based scoring, but with an added constraint that fundamentally alters move sequencing and tactical decision-making.^[26] The variant was popularized, and possibly invented, by the online player known as ChessWhiz in the early 2000s, with an early recorded sample game dating to September 4, 2004, between ChessWhiz and Xopods on the KGS Go server. This example demonstrates the variant's dynamics through a full game file in Smart Game Format (SGF), highlighting how the restriction reshapes opening and middlegame strategies. Unlike traditional Go, where players freely choose any empty intersection, Tenuki Go forbids the next player from placing a stone adjacent—either orthogonally (side-by-side) or diagonally—to the opponent's immediately preceding move, effectively compelling a tenuki response unless no other legal options exist.^[26]^[27] This rule creates a game that diverges significantly from standard Go, emphasizing global strategy over localized skirmishes and transforming routine exchanges into ko-like threats. Players must identify and execute "bigger" moves elsewhere on the board before returning to contested areas, as each placement elsewhere sets up a potential recapture or extension opportunity upon return. The restriction profoundly impacts life-and-death analysis in groups, as standard response patterns become illegal, leading to more complex tactical fights where timing and threat assessment are heightened. Passing is handled under standard Go rules, though the enforced tenuki often extends play by delaying resolutions in key sectors. Overall, Tenuki Go shifts focus toward sente (initiative) management and board-wide aji (potential), making it a challenging exercise in restraint and prioritization for experienced players.^[26]

Non-contact Go

Non-contact Go is a variant of the board game Go that modifies the placement rule to forbid stones from being placed orthogonally adjacent to an opponent's stones, creating an isolation mechanic that emphasizes territorial expansion over direct confrontation or combat.^[28] Played on a standard square grid, such as the 19x19 board, players alternate turns with Black starting; on each turn, a player places one stone of their color on an empty intersection that maintains at least one empty point buffer from all opponent stones, or passes their turn. The game concludes after two consecutive passes by both players, at which point scoring occurs exclusively by territory: a player's score consists of the number of their stones on the board plus all empty points orthogonally adjacent solely to their own stones. The player with the higher score wins, though ties are common due to the symmetric nature of play and the inability to disrupt opponent territory; optional komi can be added to White's score to balance Black's first-move advantage.^[28] Captures are impossible in this variant because the no-adjacency rule prevents surrounding opponent groups, shifting the focus entirely to strategic growth and space division without tactical fights. A simplified sub-variant eliminates passing altogether, with the player unable to make a legal move declared the loser, further streamlining play for introductory purposes.^[28] The primary challenge lies in balancing aggressive expansion to claim territory while avoiding self-imposed restrictions, as the buffer requirement can lead to blocked areas if placements are not carefully planned to preserve flexibility for future moves.^[28]

Multi-player variants

Pair Go

Pair Go is a team-based variant of Go designed for mixed-gender pairs, where two players per team—one male and one female—collaborate to play a single color on a standard 19x19 board, alternating moves without direct communication during the game. Invented in 1990 by Hisao Taki in Japan to promote social play and increase female participation in Go, the format emphasizes partnership and strategic harmony between teammates. In Europe, Pair Go gained traction in the mid-1990s, with the European Go Federation (EGF) actively promoting it for inclusivity and to foster a welcoming environment for players of all genders and skill levels.^[29]^[30] The core rules retain standard Go mechanics, including Japanese scoring and the prohibition of suicide moves, but adapt the turn structure for teams. Partners alternate making moves for their color in a fixed sequence: the female player of the black team starts, followed by the female of the white team, then the males, repeating thereafter. No consultation or signaling is permitted during play to prevent hints, with violations resulting in penalties or forfeiture; however, teams may discuss strategy beforehand. If a player passes on their turn, the game continues with the next player in sequence, allowing the partner to play subsequent moves for the team, and the game ends only after two consecutive passes by opposing teams. This setup highlights the dynamics of trust and intuition in partnerships, as players must interpret their teammate's intentions through moves alone.^[31]^[32] Pair Go has fostered international competitions, with the inaugural International Amateur Pair Go Championship held annually in Japan since 1990, drawing representatives from over 20 countries and underscoring its global appeal. In Europe, the first European Pair Go Championship occurred in Amsterdam in 1995, becoming a regular event from 1997 onward, often integrated into the European Go Congress to encourage broad participation. The championship continues annually, with the 34th held in Osaka, Japan, in August 2025. These tournaments emphasize collaborative play and have contributed to Pair Go's role in building community, particularly through its focus on mixed pairs that blend competitive strategy with relational elements.^[33]^[30]^[34]

Rengo

Rengo is a collaborative team variant of Go that originated in Japan during the Edo period (1603–1868), where multiple players per side work together on a single game to emphasize social interaction and shared strategy. Traditionally played in Go salons, it allowed enthusiasts to engage collectively, contributing to the game's cultural role as a communal activity beyond individual competition. In this format, teams typically consist of two players per side, who alternate moves in sequence without communication or consultation, distinguishing it from solo Go by requiring coordinated strategy through observed play alone. The game adheres to standard Go rules on a 19×19 board, with Black moving first and conventional scoring. Informal variants, such as "consulting rengo," permit discussion among teammates or external advice, often used for teaching or casual play. Time limits are enforced at the team level, such as shared clocks or byoyomi periods, to maintain pace in tournaments or club settings.^[35]^[36] Rengo remains popular in contemporary Go clubs and events worldwide, often as a social evening activity that accommodates players of varying strengths by averaging team ratings for handicaps.^[35] Variants like consulting rengo are commonly used for teaching or casual play among novices.^[37]

Zengo

Zengo is a three-player variant of Go designed to accommodate odd-numbered groups, where participants cycle through playing moves for black and white stones without fixed color ownership.^[38] Invented in the early 2000s as an educational extension of standard two-color alternation from Go, it allows players to experience both offensive and defensive perspectives in a shared board environment.^[38] The game proceeds with players taking turns in sequence, with the first player placing a black stone, the second a white stone, the third another black stone, and so on, effectively making each move "neutral" in the sense that it benefits neither player exclusively since all contribute to both colors over time.^[38] Standard Go capture rules apply uniformly across all stones, regardless of the placing player, preventing any player-specific advantages in removals and maintaining fairness in group liberties.^[38] To avoid alliances, the rules enforce strict turn rotation without consultation, ensuring that no two players consistently support one color.^[38] Scoring, when used, can vary and may involve tallying captured stones or evaluating overall board balance, dividing points based on contributions to black and white territories proportionally among players, but the variant often emphasizes learning over competitive winning.^[38] Since around 2010, Zengo has seen increasing casual play, particularly on online platforms like the Dragon Go Server, where it supports up to 15 players and has been available since 2012, fostering its use in teaching and recreational settings.^[39]^[38]

Micaël Go

Micaël Go is a multiplayer variant of Go originally designed in 2023 by a member of the André-Grasset College go club in Montreal, for the purpose of enabling teams of unequal size to play balanced matches. Designed for 3 or more players divided into teams, it features equal distribution of stones among team members on a standard 19×19 board, with refills to maintain balance. The game proceeds with black and white stones alternating as in standard Go, but players use their allocated stones in any order until exhausted, then receive a refill (e.g., initial 12 stones per player for teams of 3-4). Non-team members do not communicate except for expressing dismay over blunders, fostering a collaborative yet structured environment that highlights team strategy. At the end, territory and captures are scored conventionally for the teams. This variant draws from team elements in rengo but uniquely addresses unequal team sizes to develop strategic balance and group dynamics. It has spread to the U.S., Europe, and Korea, featured in tournaments like the European Go Congress 2024 in Toulouse, promoting both Go proficiency and inclusive multiplayer play in educational and casual settings.

Information-limited variants

Blind Go

Blind Go is a memory-intensive variant of Go where players do not see the board and must maintain the entire position in their minds throughout the game. Players alternate announcing their moves verbally using standard coordinate notation (e.g., "3-4" or "Q16"), and a referee records these on a hidden board to verify legality, such as ensuring no suicides, self-captures, or ko repetitions occur. If a player announces an illegal move due to a memory lapse, they are prompted to select another intersection, maintaining the standard rules of placement, capture, and passing otherwise.^[22]^[40] In the modern form, each player typically has their own hidden board or relies solely on mental visualization, with moves called aloud to the referee who updates both positions if needed. The game proceeds on a standard 19×19 board, though smaller 9×9 boards are common for practice due to the increased difficulty of memorizing larger configurations. At the endgame, players must fully recall the position to propose territory counts, though the referee's record serves as the authoritative source for final scoring under Chinese or Japanese rules, including komi adjustments. This requirement tests comprehensive board recall, often leading to errors in complex endgames with scattered groups.^[22]^[40] Blind Go enhances players' visualization and spatial reasoning skills, as strong performers can track up to 150 moves or more before lapses occur, a capability more demanding than in blindfold chess due to Go's larger board and subtler positional nuances. Professional-level play is rare but demonstrated by individuals like Bao Yun, who has completed full 19×19 games from memory. The variant originated from blindfolded board game traditions dating to the 7th century, adapted for Go to challenge cognitive limits.^[40] Tournaments featuring Blind Go emerged in the late 20th century, with notable events including human-computer matches at the IEEE SSCI in Paris in 2011, where the program MoGoTW defeated professionals in three out of four games by exploiting memory burdens on humans. These competitions underscore the variant's role in AI research and player training. Earlier informal exhibitions date to the 1970s in Go clubs, promoting it as a skill-building exercise.^[40] A related variant, silent blind Go, eliminates verbal announcements entirely, forcing players to internalize moves without auditory reinforcement from the opponent or referee, thus heightening the memory demands and often limiting games to smaller boards.^[22]

One Color Go

One Color Go is a variant of the traditional board game Go in which both players place stones of identical color on the board, eliminating visual cues for ownership and compelling participants to rely on memory or external aids to track their respective positions. Unlike standard Go, where black and white stones clearly delineate player control, this setup tests visualization and recall by forcing players to mentally associate stones with the player who placed them. The game is played on a standard 19×19 grid, with alternating turns and the same core objective of controlling territory through encirclement. Captures follow the principles of standard Go, where a group of stones loses all adjacent empty intersections (liberties) and is removed from the board, but determining which player's group is affected requires remembering the sequence of moves or the last player to contribute to the surrounded formation. Scoring is conducted by assessing territory and captured stones at the game's end, often involving a post-game reconstruction where players verbally or visually clarify ownership to tally points accurately. In practice, disputes over ownership or captures are resolved by replaying the game from the beginning to verify positions, ensuring fairness through collective recollection or reference to move logs. This variant builds essential skills in memory retention and error prevention, as players must carefully read ahead to avoid blunders in ambiguous configurations, making it a valuable training tool particularly for intermediate learners. It gained notable play during events like the European Go Congress, where participants have demonstrated proficiency by managing multiple simultaneous games under these constraints. Although informal and unranked in most settings, its use of notes or secondary boards for tracking is prohibited in strict tournament play to maintain the memory challenge. Online adaptations of One Color Go, available on servers such as KGS and OGS, simulate the uniform stone appearance while internally recording ownership via timestamps and move histories, allowing captures and scoring to proceed as in standard play without visual distinction. These digital implementations facilitate the variant's accessibility, often resulting in games that end in resignation due to the heightened cognitive demands.

Phantom Go

Phantom Go is a high-uncertainty variant of Go designed to simulate imperfect information, where players cannot see their opponent's stones or moves until specific events like captures force revelations. Phantom Go was first played at the 12th Computer Olympiad in Amsterdam in 2007. The game traditionally involves players using separate boards, with a referee overseeing the full board state to verify the legality of each proposed move and only revealing details when an illegal placement occurs—such as attempting to occupy an intersection already taken by the opponent—or when a capture happens, at which point captured stones are removed and announced by count to allow board synchronization.^[41]^[42] If a player proposes an illegal move due to guessing an opponent's position incorrectly, the referee instructs a retry or, in stricter variants, a pass, adding tension and forcing reliance on inference rather than direct observation. This setup contrasts with other information-limited variants like blind Go, which may involve verbal move announcements but no physical hiding. The core emphasis is on probabilistic strategy, where players model possible opponent configurations based on sparse feedback, such as atari warnings or capture numbers, to anticipate threats and opportunities without full visibility.^[43]^[41] Scoring occurs after both players pass, with the referee revealing the complete board for standard territory and capture evaluation, often using Japanese or Chinese counting rules adapted to the final state. This final reveal underscores the variant's focus on uncertainty management over tactical precision. Extensions include phantom rengo, a team-based adaptation where multiple players per side collaborate under the same hidden mechanics, promoting discussion and shared probabilistic modeling during turns.^[42]

Non-standard board variants

Toroidal Go

Toroidal Go is a variant of the board game Go played on a toroidal surface, where the opposite edges of the standard grid are connected, creating a seamless, edgeless topology equivalent to the surface of a doughnut. This setup eliminates the traditional advantages and disadvantages associated with corners and edges in standard Go, resulting in a more uniform playing field across all intersections. The board is typically 19×19, though smaller sizes like 11×11 are common in digital implementations, with each point having exactly four adjacent neighbors due to the wraparound connections.^[44] The rules largely follow those of standard Go, with liberties and captures adapting to the topology: a stone's connections wrap around the board, allowing groups to extend influence across what would be edges in a conventional setup. For instance, a chain of stones can connect from one side to the opposite, potentially encircling opponents in novel ways. Due to the perfect symmetry of the toroidal board, no komi is theoretically required, as neither player benefits from first-move territorial bias near boundaries; however, some implementations use a small komi (e.g., 3.5 or 4.5 points for White) to fine-tune balance on finite grids.^[44]^[45] This variant emerged as a speculative idea among mathematician players of Go, attracting interest for its exploration of non-Euclidean geometries in gameplay. The edgeless design shifts strategic focus toward the center, promoting more centralized formations and reducing edge-oriented tactics, as confirmed by observations in early discussions where "the center is the game." Software support for Toroidal Go appeared in the early 2000s, with platforms like Little Golem offering online play on 11×11 boards since at least 2004, enabling tournaments and analysis that highlight these central tendencies.^[44]^[46] Mathematically, Toroidal Go holds appeal for its graph-theoretic properties, such as being vertex-transitive—meaning the board is homogeneous, with every intersection equivalent under symmetries like rotations, reflections, and translations. This structure facilitates studies approximating infinite boards through periodic boundary conditions, useful in theoretical analyses of Go on unbounded grids.^[47]

3D Go

3D Go extends the traditional game of Go into three dimensions, introducing volumetric boards that stack multiple layers to enable intricate spatial strategies beyond the flat plane. The variant has roots in the 1980s, with an early physical implementation featuring a 9x9x9 board built by Knut Roll-Lund and tested at the Kristiansand Go Center in Norway during spring 1983.^[48] This marked one of the first documented efforts to play Go in a fully three-dimensional space, adapting core mechanics like placement and capture to a cubic grid structure. Core rules in 3D Go maintain the essence of surrounding and capturing opponent groups, but liberties are defined in three dimensions, typically using a cubic lattice where each position connects to up to 6 face-adjacent neighbors. A group is captured when all its adjacent empty positions—its liberties—are occupied by the opponent, requiring complete enclosure in x, y, and z directions. Scoring shifts from two-dimensional area to three-dimensional volume, evaluating the enclosed space controlled by each player at game's end. Some implementations, like Freed Go software released in 2004, support configurable neighbor counts of 3, 5, or 6 to model various 3D topologies while enforcing standard Go features such as ko rules and territory counting.^[49] Physical realizations often involve stacking components to simulate depth, as seen in Margo, a 3D Go variant invented by Cameron Browne in 2006 and published by nestorgames in 2012. Played on a 7×7 grid, Margo uses colored balls stacked in holes or on platforms, where groups survive only if they maintain freedoms at the board's surface level; captured supporting balls become persistent "zombies" that remain in play. A smaller 4×4 variant called Spargo emphasizes tactical depth in this stacked format. These designs address 3D playability by limiting perpetual expansion, but they introduce challenges like reduced visibility into stacked interiors and the risk of illegal self-atari placements in deeper layers, where a move might immediately expose a group to capture without escape routes. Software versions like Freed Go facilitate digital play on diverse 3D graphs, mitigating physical visibility issues while preserving the variant's strategic complexity.^[50]^[49]

Round Go

Round Go is a variant of Go played on a circular board designed to eliminate the corners of traditional square grids, allowing for more balanced territorial expansion and fluid gameplay. Invented by Harald Schwarz in 1996, it was proposed for aesthetic reasons to create a more symmetrical and visually appealing playing surface, diverging from the rectangular boards common since the game's ancient origins.^[51]^[52] The variant was first introduced publicly at the European Go Congress in Abano Terme, Italy, during the summer of 1996, where it gained initial interest among players for its novel board shape.^[52] The board consists of 81 intersection points arranged in a circular pattern approximating a 9×9 grid, with lines forming a donut-like structure that curves at the edges. Stones are placed on these intersections following grid-based placement rules from standard Go, but the curved edges ensure every point maintains exactly four liberties, even along the perimeter, preventing the "dead corners" that can influence strategy in square boards.^[51]^[52] This design promotes the formation of more organic and flowing shapes, as players cannot rely on corner advantages for early territory control. Otherwise, the rules mirror those of standard Go, including alternation of turns, capturing by surrounding opponent stones, and scoring by territory and prisoners at the game's end. Komi adjustments and handicaps are applied similarly, tailored to player skill differences—for instance, 3 komi for even games or additional stones for larger disparities—to maintain fairness on the symmetric board.^[52] The variant's aesthetic appeal has led to printable board templates and occasional play in European Go circles, emphasizing conceptual balance over the directional biases of rectangular layouts.^[51]

Hexagonal Go variants

Hexagonal Go variants adapt the core mechanics of Go to hexagonal lattices, where each intersection or cell typically connects to six neighbors rather than four, fundamentally altering strategic connectivity and liberty dynamics. These variants emerged in the 2000s, drawing inspiration from connection-based games like Hex to explore non-square grids while retaining Go's territorial objectives and capture rules. Unlike standard Go on a square grid, where liberties are orthogonal, hexagonal setups increase the number of adjacent positions, leading to higher initial branching factors in gameplay due to expanded move options, but result in shorter games through accelerated enclosure and capture opportunities from enhanced group interlinking.^[53] In these variants, stones are placed on intersections or cells of a hexagonal board, with captures occurring when an opponent's group is fully surrounded and deprived of all liberties, mirroring standard Go but adapted for the grid's geometry. Boards are commonly hexagonal in shape, such as an 11x11 configuration yielding approximately 91 cells or 157 intersections depending on placement convention, though smaller sizes like 8x8 (169 intersections) are used in specific implementations to maintain playability. Scoring remains area-based, tallying controlled empty spaces (territory) plus captured stones, with adjustments like komi for balance; for example, in one analyzed game on a hexagonal grid, scores were calculated as stones on board plus territory, resulting in a close 67–66.5 finish including 6.5 komi. This setup emphasizes fluid invasions and escapes, as the six-directional connectivity reduces the efficacy of simple walls compared to square grids, where liberties are more linearly constrained.^[54]^[55] A prominent subvariant is TriGo, which modifies placement to accelerate pacing on a hexagonal board with an 8-intersection-per-side triangular lattice (169 points total). Players alternate turns, with Black placing one stone first, followed by both players placing two stones per turn thereafter—either adjacent to their own group or anywhere unoccupied, subject to no-suicide rules; passing forfeits both placements. Captures follow the liberty model, with groups needing three eyes for secure life due to the grid's density, and a simple ko rule prohibits immediate recapture to the prior position. The game ends on consecutive passes, scored by surrounded empty intersections plus captured stones, fostering aggressive multi-stone responses that heighten tactical complexity over standard Go. TriGo is supported online through abstract strategy platforms, enabling accessible play and analysis.^[55]^[56] Another influential example is Keil, invented by Luis Bolaños Mures in November 2019, designed explicitly to preserve Go's tactical depth like crosscuts and ko fights on a hexagonal grid of triangles' intersections (equivalent to hex-cell spaces). Placement is single-stone per turn after Black's first move, with mandatory liberty for the placing group and positional superko to prevent cycles; captures remove liberty-less enemy groups post-placement. This variant mitigates the grid's reduced natural liberties (often three per point in vertex placements) by introducing "linked cells" for conditional connectivity, allowing nuanced ko resolutions and maintaining game length comparable to 9x9 Go despite the altered topology. Keil's innovations, detailed in abstract strategy analyses, highlight how hexagonal variants can balance increased strategic variance with familiar Go principles, promoting deeper study of enclosure in multi-directional environments.^[54]^[57]

Piece and placement variants

Omino Go

Omino Go, also known as Tetromino Go, is a variant of the board game Go invented by R. Wayne Schmittberger in the early 1990s.^[58] In this game, players place polyomino-shaped pieces called tetrominoes—each consisting of four connected squares—rather than individual stones, transforming the traditional placement mechanic into a puzzle-like fitting challenge.^[58] The game is played on a standard 19x19 Go board, maintaining the core strategic elements of territory control while emphasizing spatial arrangement and connectivity.^[58] Tetrominoes connect edge-to-edge, forming the seven distinct shapes familiar from Tetris: the straight I, square O, T, L, J, S, and Z pieces, with rotations allowed but no reflections.^[58] On each turn, a player places one complete tetromino (a "platoon" of four connected units) anywhere on the board, provided it does not overlap existing stones; Black's opening move is restricted to a domino (two units) to balance the starting position.^[58] Groups of connected tetrominoes from the same player are treated as single entities for capturing purposes, following standard Go rules where an opponent's group is captured if it has no liberties—empty adjacent spaces.^[58] This mechanic encourages players to strategically enclose enemy shapes, adding a layer of geometric puzzle-solving to the territorial gameplay.^[58] The full rules and variations of Omino Go were first detailed in Schmittberger's 1992 book New Rules for Classic Games, which explores innovative twists on traditional games.^[59] The variant highlights the puzzle-like nature of fitting irregular shapes onto the board, promoting creative placement strategies over simple point-by-point expansion.^[58]

Coin Go

Coin Go, also known as Pillar Go, is a variant of the board game Go that introduces neutral, uncapturable elements in the form of coins or pillars placed on specific board intersections. These coins serve as visual markers for forbidden placement spots, effectively creating inner walls or barriers that influence strategy without altering the fundamental rules of stone placement, capture, and scoring. The variant was proposed and discussed in online Go communities around 2022, aiming to add tactical complexity through fixed neutral points that affect liberties and territory control.^[60]^[61] Under standard Coin Go rules, players cannot place stones on intersections occupied by coins, treating them as impassable. Adjacent stones connected to a coin lack a liberty in that direction, potentially facilitating captures, but the coins themselves cannot be removed or captured regardless of surrounding groups. Scoring follows conventional territory rules, with coins not contributing to any player's score. Coin placement varies by setup: they may be positioned randomly or by mutual agreement at the game's outset, or players can alternate placing them before normal play begins. To mitigate potential advantages, the pie rule is commonly applied, allowing the second player to swap colors after the first move.^[60] In one popular implementation, each player receives an even number of coins—such as three—and can spend a turn to place one on an empty intersection or, optionally, remove an opponent's coin, introducing dynamic decision-making about resource use. This mechanic encourages strategic timing, as coins act as semi-permanent obstacles that can block approaches or deny liberties without the risk of loss. The board remains the standard 19x19 grid, preserving Go's core emphasis on enclosure and influence while the coins provide a neutral layer of interference.^[61]^[62] Coin Go balances traditional skill with subtle positional constraints, making it suitable for experimental or casual play. Random initial placement adds a minor element of chance, though the focus remains on deterministic strategy once the board is set. It has been featured in discussions on Go variant forums, highlighting its potential for creative adaptations like combining with other rulesets.^[63]

Block Go

Block Go emerged in the early 2020s within maker and DIY communities as a physical variant of Go proposed by user "illia" on BoardGameGeek in June 2020. It introduces stackable wooden blocks to create vertical dimensions on a standard 2D board. Players place stacks of blocks of their color on intersections, with the height of each stack determined by the number of liberties it possesses immediately after placement, extending the game into a vertical dimension where height represents available liberties. This mechanic transforms traditional placement into a construction challenge, where stability and positioning are key to maintaining structures.^[64] Liberties in Block Go are calculated both horizontally across the board and vertically within stacks, extending the classic liberty system into three dimensions. A stack or connected group is captured if all its liberties are filled, causing the entire structure to topple and be removed from play, simulating a collapse. This adds tactical depth, as players must balance horizontal encirclement with vertical threats to destabilize opponents' builds.^[64] At the game's end, scoring follows standard Go area control principles, counting enclosed territory and captures, with stacks treated as groups for liberty and territory purposes. Exclusively designed for physical play, Block Go relies on tangible wooden components, fostering hands-on interaction without digital aids. In contrast to standard Go rules, which explicitly prohibit stacking stones to ensure a flat playing surface, Block Go leverages this forbidden element for innovative gameplay.^[65]^[64] DIY kits for Block Go, often shared through maker forums and craft sites, enable enthusiasts to fabricate custom sets using accessible materials like balsa wood or 3D-printed prototypes, underscoring the variant's emphasis on engineering creativity and accessibility.^[64]

Pixel Go

Pixel Go is a digital-native variant of the ancient board game Go, emerging in the late 2000s and popularized through online servers in the 2010s and 2020s. Invented by player Mikkgo during the Experience Go in China 2009 program in Beijing, it draws inspiration from informal play between lines on a standard Go board, adapting the core mechanics to multi-stone placements for a faster-paced experience. Unlike traditional Go, where players place single stones, Pixel Go requires each move to consist of a 2×2 block of stones, referred to as a "pixel" or "blob," treated as a single unit. This innovation was first implemented digitally on platforms like the Variant Go Server, launched around 2020, enabling easy experimentation and tournaments.^[66] The game is played on a standard 19×19 grid, but the 2×2 placement size effectively scales the board to resemble a 9.5×9.5 layout in terms of strategic coverage, accelerating the filling of space and reducing game length. Players position their blob anywhere unoccupied or partially overlapping their own existing stones, with any overlapping positions simply ignored to avoid duplication. Connection between blobs follows traditional Go rules, where groups share liberties via adjacent edges (not corners), forming larger chains that can be defended or attacked collectively. Captures occur when an opponent's entire group of connected blobs is surrounded and deprived of liberties, removing all stones in the captured blobs at once—emphasizing enclosure of multi-cell units rather than individual stones. Openings often mirror smaller-board Go strategies, such as approaching the 5-4 point, but with added complexity from partial overlaps and edge placements that may extend near board boundaries.^[66]^[63] Pixel Go's blob-based moves enable faster territorial coverage and development, typically shortening games to suit casual or training sessions compared to standard Go's single-stone pace. It supports both area scoring (counting enclosed empty points plus captured stones) and territory scoring (enclosed points only), though Chinese-style area scoring—with neutral points valued—is common in competitive play, alongside positional superko to prevent repeats. Subvariants extend the concept by varying blob sizes, from 1×1 (equivalent to traditional Go) to 3×3 blocks, allowing customization for different speeds or board scalings, while maintaining core connection and capture principles. The variant has been used in online tournaments, such as the first Pixel Go event in 2020, and serves as a training tool on servers to practice grouping and enclosure tactics in a condensed format.^[66]^[67]^[68]

Scoring and objective variants

Environmental Go

Environmental Go is a variant of the board game Go designed to incorporate an external "environment" in the form of a deck of value coupons, allowing players to opt for point accumulation outside the board during play. Invented by mathematician Elwyn Berlekamp in the late 1990s, it extends standard Go rules to facilitate the study of move valuations, particularly in endgame scenarios, by providing quantifiable alternatives to board actions.^[69]^[70] The core rules follow conventional Go on a 19×19 board, with players alternating turns to place black or white stones, subject to standard liberties, captures, ko restrictions (though modified), and suicide prohibitions. The key innovation is a stack of 40 coupons—cards valued at 0.5, 1, 1.5, up to 19.5 and 20 points in 0.5 increments—placed face down. On their turn, a player may either place a stone on an empty intersection or draw the top coupon, adding its value to their personal score without affecting the board. This option removes traditional ko bans, as coupon-taking serves as a neutral "pass" equivalent, enabling analysis of forced cycles. The game ends when both players consecutively take coupons or pass, after which scoring combines the traditional territory or area count (using Japanese, Chinese, or American rules) with the sum of each player's collected coupons, plus 9.5 komi for White to compensate for Black's first-move advantage.^[70]^[69]^[71] Berlekamp developed Environmental Go as part of his work in combinatorial game theory, aiming to empirically assess professional players' perceptions of move values by observing when they prefer coupons over board plays. For instance, in thick positions with low-gain opportunities, players might take high-value coupons (e.g., 5 points) instead of marginal board moves estimated at 10 points but with positional risks. The variant's structure, with incrementally valued coupons, highlights decision-making thresholds, such as the point at which a board play outweighs the top available coupon. This has proven useful in academic contexts, including analyses in the collection More Games of No Chance (2002), where thick coupon stacks with small increments reveal optimal strategies.^[69]^[71]^[72] The first recorded game occurred on April 21, 1998, in Menlo Park, California, between 9-dan professionals Jiang Zhujiu (Black) and Rui Naiwei (White), lasting 333 moves. Under Chinese (Ing) scoring, White won by 2.5 points; under Japanese/American rules, Black prevailed by 0.5 points, demonstrating how scoring methods interact with coupon totals. A second game followed in July 2000, and additional matches have been played at events like the American Go Congress. These games underscore the variant's viability for high-level play while providing data on endgame economics, such as move 268 in the inaugural match where coupon values influenced ko fights. Environmental Go remains a niche tool for theoretical exploration rather than widespread competition, with game records available for study.^[70]^[71]^[69]

Bid Go

Bid Go is a variant of the board game Go that incorporates an auction mechanism to determine the right to make moves, aiming to address imbalances such as the first-move advantage without relying solely on traditional komi compensation. Developed by Bill Spight in the late 20th or early 21st century, it introduces bidding rounds that add an economic strategy layer, requiring players to strategically allocate resources to prioritize key plays while managing diminishing opportunities as the game progresses.^[73] This variant maintains the standard 19x19 Go board and core placement rules but modifies the turn structure through bids, making it particularly suitable for players seeking fairness in even-strength games. In Bid Go, the game begins with players selecting their colors, typically without komi unless used as a handicap; any handicap stones are placed on the board prior to bidding. Pre-game bidding is not explicitly required, but the initial setup emphasizes mutual agreement on colors to ensure equity. During play, bidding occurs simultaneously at the start of each round, where players submit bids (often written down) representing their willingness to "pay" for the initiative. The average of these bids sets the "temperature," a value that governs the round's intensity and must decrease in subsequent bids to prevent escalation. The highest bidder gains the right to move first in that round, playing a legal Go move unless the temperature drops to zero or below, at which point passing is allowed.^[73] If bids are equal, the player who did not make the last move (or Black if at the start) initiates. Passing awards the passer the current temperature in points but lifts any ko bans, encouraging tactical decisions on when to conserve resources. The gameplay proceeds in rounds separated by bidding phases, with the board following standard Go rules for stone placement, captures, and ko restrictions except where bidding intervenes. A round ends when two consecutive passes occur in the same board position; if the temperature remains positive, a new bidding round resumes play, but if it reaches zero or negative, the game concludes. This structure emphasizes selecting the "largest" available plays, as bids deplete over time, forcing players to weigh immediate gains against long-term board control. Variants like Bill Taylor's Bid Go eliminate distinctions between sente and gote moves to simplify strategy, while ChipUni's Alternate Bid Go allocates a fixed pool of tokens (e.g., around 1000 per player) at the outset, with bids using these tokens to claim moves— the highest bidder pays their bid to the pot, and the game ends if a player exhausts tokens or after mutual passes, using standard scoring thereafter.^[74] In all forms, the winner is determined by conventional territory or point scoring, adjusted implicitly by the bids' influence on move opportunities rather than direct subtraction from the final tally. This auction-based approach fosters an additional layer of economic decision-making, where players must forecast the value of potential moves and bid accordingly to outmaneuver opponents without overcommitting resources early. While no widespread online tools for bidding were documented prior to 2012, the variant's mechanics lend themselves to digital implementation for simultaneous bid submission, enhancing accessibility in modern play. Bid Go thus promotes balanced competition by auctioning privileges dynamically, distinct from static handicaps.^[73]

Multi-player Go

Multi-player Go extends the traditional two-player game to three or more participants by introducing additional colors of stones, allowing simultaneous competition on a shared board. To balance the increased complexity and first-mover advantages, rules are adapted for fairness, often on enlarged boards for four players, providing more space for individual strategies without immediate territorial conflicts.^[75] Players alternate turns in a strict cyclic order—typically sequential, though parallel moves are a variant where conflicting placements result in passes—ensuring the rotation continues even if a player passes. Alliances are forbidden in the core version, preventing cooperative territory claims or shared prisoners, with each participant required to compete independently. Captures function as in standard Go, removing any opponent's stones lacking liberties regardless of which player surrounds them, and prisoners may be divided among captors or returned to the board in specific rule sets.^[75] Scoring divides the board into personal territories, where each player counts enclosed empty points and their own stones using area or territory methods, aiming to maximize individual points without joint victories. The game concludes when all players consecutively pass, promoting a focus on personal expansion over direct confrontations. This setup maintains Go's emphasis on influence and enclosure while scaling to multiple competitors.^[75] Multi-player Go is commonly played informally at Go clubs and tournaments for recreational purposes, adding variety to events with larger groups. Software implementations, leveraging Monte-Carlo Tree Search algorithms like UCT variants (including paranoid, alliance, and confident adaptations), enable computational play at reasonable levels, with extensions supporting up to six players on boards like 9×9 for experimentation.^[76]

Unequal Resources Go

Unequal Resources Go, also known as =/= Go, is a scoring variant of Go where the value assigned to captured stones (prisoners) differs from the value of a point of territory, referred to as a moku. This adjustment allows players to modify the relative worth of captures versus enclosed areas during endgame scoring, while the core rules of play, including stone placement, captures, and ko restrictions, remain unchanged. The variant enables fine-tuned balance in games between players of unequal skill levels by altering resource valuation rather than altering the board setup.^[63] In practice, the game follows standard Go procedures on a conventional board, such as 19x19, with players alternating turns to place stones and form groups. Captures occur as usual when an opponent's group has no liberties, but the scoring phase introduces the inequality: prisoners may be valued higher or lower than territory points to compensate for disparities in player strength. For instance, this can benefit a weaker player by doubling the point value of their captures, effectively rewarding aggressive play without granting extra stones at the outset. Such modifications promote teaching scenarios by emphasizing strategic capture tactics over territorial enclosure alone.^[63] A notable subvariant is Stonethrow, in which each prisoner is worth only half a moku of territory. This reduces the incentive for captures relative to secure territory building, shifting focus toward defensive strategies and long-term board control. Stonethrow exemplifies Unequal Resources Go by halving the typical equivalence between prisoners and territory, often used to create more even contests in casual or instructional settings.^[63] This variant aids in providing handicaps without placing additional stones, making it suitable for club environments where players seek alternatives to traditional komi or positional advantages. Its adoption in teaching contexts helps novices experience balanced games, fostering skill development through adjusted outcomes rather than overwhelming deficits.^[63]

Connection and territory variants

Sygo

Sygo is a two-player abstract strategy game invented by Christian Freeling in November 2010.^[77] It combines elements of traditional Go with the Symple move protocol, emphasizing efficient enclosure of territory while incorporating Othello-style captures to create shorter, more dynamic games compared to standard Go.^[78] The game is played on a standard 19×19 Go board, known as a goban, using bi-colored stones for both players, with White moving first.^[78] In Sygo, players alternate turns with restricted placement options derived from the Symple protocol: a player may either place a single stone to initiate a new group anywhere on the board (provided it does not immediately have no liberties) or add exactly one stone adjacent to an existing group of their own color.^[78] Placement adjacent to an opponent's stones is prohibited, which shifts the focus from direct confrontation to strategic surrounding and growth of groups. Captures occur when an opponent's group loses its last liberty through placement or growth; instead of removal, the captured stones flip to the captor's color, similar to Othello, preventing cycles like ko and promoting fluid board development.^[78] Suicide moves are illegal, and passing is allowed, but the game ends only upon two consecutive passes or resignation.^[78] Scoring follows area rules, where a player's territory consists of their stones on the board plus the empty points they surround; the player controlling the larger territory wins, with draws possible if territories are equal.^[78] This system rewards compact, efficient enclosures over expansive play, as the move restrictions limit total placements and encourage surrounding maneuvers without the prolonged invasions typical of Go. Games are significantly faster than standard Go while retaining the core conceptual depth of territorial control.^[79] As part of Freeling's portfolio of abstract strategy games, Sygo highlights innovations in move protocols to streamline connection and enclosure mechanics.^[77] Online play for Sygo has been available since its creation through the Mind Sports International arena at mindsports.nl, where players can access interactive boards and example games.^[80] The game is also featured on BoardGameGeek, facilitating discussion and rules dissemination among enthusiasts.^[81]

Infinite Go

Infinite Go is a theoretical variant of the game of Go designed to eliminate the constraints of fixed board edges by playing on an infinite grid. The concept originated as a conceptual exercise around the 1980s. In Iain M. Banks' 1985 novel Walking on Glass, a similar idea appears as "Open-Plan Go," where the board expands limitlessly and pieces form infinitely extending lines, highlighting the variant's emphasis on unbounded spatial dynamics.^[82] The rules retain standard Go mechanics for placement and capture: players alternate placing stones on intersections of an imagined infinite lattice, surrounding opponent groups to remove them when they lack liberties. Unlike finite boards, there are no corners or edges to exploit, shifting focus to central confrontations and emergent shapes. Games conclude by mutual agreement once stable fronts—regions of secure territory or uncontested influence—form, often after evaluating captures or positional advantages in a designated central area, such as a virtual 19x19 zone for scoring. This approach underscores pure shape theory, where viability of formations depends solely on internal liberties and connections, free from edge distortions.^[83] Software adaptations simulate infinite boards using expandable grids or periodic boundary conditions as virtual extensions, enabling analysis of long-term strategies and local tactics. These tools prioritize computational study of shape interactions, such as eye formation and cutting patterns, in isolation. Infinite Go has informed AI research in Go by providing a model for evaluating midgame shapes without finite 19x19 limitations, aiding algorithms in pattern recognition and tactical reading on open fields.^[83]

Paper-and-pencil Go

Paper-and-pencil Go is an informal adaptation of the traditional game of Go that enables play without a formal board or physical stones. Players draw freehand circles on any available paper to represent black and white stones, approximating liberties and group connections visually rather than adhering to the gridded precision of standard Go. This approach allows for flexible, accessible gameplay in casual settings, emphasizing strategic intuition over exact measurements. Unlike standard Go, paper-and-pencil versions often dispense with a fixed board size, permitting captures through simple encircling of opponent groups and relying on eyeballed estimates for territory and scoring at the end. The lack of rigid structure makes it particularly suited for travel or impromptu sessions, where a pencil and scrap paper suffice. A popular sub-variant known as "napkin Go" further simplifies setup by using restaurant napkins as the playing surface and everyday items like drinks or coins as temporary markers for stones, fostering quick games during social outings. This variant promotes deeper understanding of Go's core concepts—such as surrounding territory and capturing—by stripping away equipment needs and encouraging creative adaptation, though it sacrifices some of the precision found in gridded play.

Reversi Go

Reversi Go is a hybrid variant of Go that fuses the surrounding mechanics of traditional Go with the stone-flipping capture from Reversi (also known as Othello), where captured opponent groups change to the capturer's color rather than being removed. The variant emerged in the early 2000s among Go enthusiasts.^[84] The game employs a standard Go board of 19×19 intersections (though smaller sizes like 9×9 are used in some implementations), starting empty with Black placing the first stone and players alternating turns. Stones are placed on empty intersections, and a legal move must ensure the player's own group retains at least one liberty after placement; suicide moves are typically prohibited. Capture occurs when an opponent's group loses all liberties, prompting an immediate flip of those stones to the capturer's color, which may trigger chain reactions of additional flips if the newly converted stones surround other groups.^[85]^[84] Games conclude when both players consecutively pass, often after the board fills due to the absence of stone removal. Scoring follows area rules, tallying each player's stones on the board plus surrounded empty territory, with adjustments like komi for White (e.g., 7 points on smaller boards) and a "button" mechanic adding 0.5 points to break ties. Handicap options allow the weaker player up to 9 pre-placed stones. This design balances Go's emphasis on territorial control with Reversi's conversion strategy, making captures dual-purpose by both securing space and augmenting one's forces through flips.^[85] The variant's flip mechanic renders traditional ko prohibitions unnecessary, as color changes and persistent stones prevent positional repetition while promoting aggressive play and multi-step conversions over static enclosures. Digital versions of Reversi Go have been playable on online platforms since the early 2010s, facilitating experimentation with its faster board-filling dynamics compared to pure Go.^[84]

References

[1]
Go - GamesCrafters
Sunjang Baduk: This variant of Go evolved in Korea during the 7th century. Instead of an empty board, Sunjang Baduk starts with 16 stones in a box-like ...
[2]
Play Atari Go Online and Train Your Stone Capture Skills - Go Magic
to be the first to capture one of the opponent's stones. You can also watch a video tutorial on this topic. The bot may takes a long ...
[3]
History of Go: China's Legacy to Global Phenomenon - Go Magic
Jan 31, 2024 · Traditional Go in Korea, dating from the late 16th century and possibly rooted in the game's original form, differed significantly from Chinese ...The Dawn of Go: A Chinese... · SunChang Go: Korea's Unique...Missing: variants | Show results with:variants
[4]
[PDF] Rules of First Capture Go Strategy & Tactics
Here are some guidelines for playing First. Capture Go more effectively: 1. Keep your stones connected; it makes them more difficult to capture. 2. Avoid making ...
[5]
Rengo - Go Magic
Rengo is a collaborative Go format where two teams, each comprising two players, compete in a single game. This format promotes teamwork and shared strategy.
[6]
Toroidal Go at Sensei's Library
Toroidal Go (henceforth referred to as “t-Go”) is a Go variant which is played on a graph that can be seen as a normal board where the edges have been “removed” ...Missing: rules | Show results with:rules
[7]
[PDF] the game of go in ancient and modern tibet - American Go Association
The difference in stone size provides additional evidence for the separation of Chinese and Tibetan go. In a very incomplete set of full size stones (about 20 ...
[8]
Tibetan Go at Sensei's Library
Apr 14, 2017 · Tibetan Ko Rule: The ko rule was: may not play a stone on an intersection where the opponent has just removed a stone.Missing: origin | Show results with:origin
[9]
Tibetan Go - BabelStone Blog
Mar 4, 2006 · It is a board game of enormous popularity in China, Korea and Japan that is played by placing black and white "stones" (rdel in Tibetan) on a 19 x 19 grid of ...
[10]
Sunjang baduk
Sunjang baduk. Sunjang baduk (Korean: 순장바둑) is a variant of the game go (known as baduk in Korean) played in Korea which dates to the 16th century.<|control11|><|separator|>
[11]
Haengma and Sunjang Baduk - General Go Discussion
Mar 11, 2023 · Sunjang Baduk is underrated and underused as a teaching technique, especially for new players making the transition from 9x9 and/or 13x13 to 19x19.Variant Ideas - Online Go ForumAncient Go variants - OGS Development - Online Go ForumMore results from forums.online-go.com
[12]
Sunjang Baduk at Sensei's Library
Oct 12, 2025 · Sunjang Baduk is one of the forms of Go, which was common in Korea through the 16th century and into the 20th century.
[13]
Sunjang Baduk Counting at Sensei's Library
Sep 10, 2024 · The old Korean variant of Go, called Sunjang Baduk, used a different kind of territory scoring and a special counting procedure:.
[14]
Area or territory – A brief history of go rules - Nordic Go Dojo
Aug 7, 2020 · Closer to the end of the 20th century, more variations of the go rules started popping up. ... (American Go Association) rules. Now, in the present ...<|separator|>
[15]
Go Rule Dialects
Feb 18, 1998 · The players like to play a Game of Go and agree mutually of a komi which is an integer or half of an integer. If this agreement doesn't take ...
[16]
Full text of "Sherpas of Nepal: Buddhist highlanders" - Internet Archive
A wealthy man described os ChetrbSherpa^ is die sen uf a poor Sherpa wn mat i who, during the winter when employment and food were scarce, used to go and ...
[17]
Batoo - Korean variant of the board game Go
Boards in Batoo usually have plus-point and minus-point spots. When a player plays on these spots, they will either gain or lose five points. Different "maps" ...
[18]
Batoo on OGS? - General Go Discussion
May 19, 2020 · Batoo. Batoo is a Korean variant of the board game Go. The name stems from a combination of the Korean words baduk (“Go”) and juntoo (“battle”).Missing: Mongolia history
[19]
Batoo at Sensei's Library
Apr 3, 2021 · Rules. Batoo is played on a 11x11 board and features plus-point and minus-point locations, three stones each played simultaneously before the ...Missing: Mongolia | Show results with:Mongolia
[20]
Luis Bolaños Mures | Board Game Designer - BoardGameGeek
Stoical Go and Incursion Go, Go variants with alternative ko rules that render all known forced cycles impossible. Kingo, a Go variant where each player has two ...
[21]
Stoical Go at Sensei's Library
May 27, 2023 · Stoical Go. Keywords: Variant. In this Go variant, standard ko rules are not used. Instead, it is illegal to make a capture if your opponent ...
[22]
Incursion Go (soft-finite variant without repetition rules ...
All known forced Go cycles are impossible in this variant. Unlike the related variant Stoical Go , Incursion Go preserves snapbacks. If you place a stone on ...
[23]
None
Nothing is retrieved...<|control11|><|separator|>
[24]
Go variants - Wikipedia
There are many variations of the simple rules of Go. Some are ancient digressions, while other are modern deviations. They are often side events at tournaments.Missing: sources | Show results with:sources
[25]
Capture Game Teaching Method - American Go Association
... Capture Game." This approach was developed by Yasuda Yasutoshi, an 8 dan professional player and one of the most devoted teachers in Japan. Yasuda sensei ...
[26]
[PDF] Go Complexities - Hal-Inria
Jan 15, 2016 · Atari Go is another variant, in which the first capture leads to a victory. Therefore, Komi has no impact. Atari Go is usually played as a ...
[27]
Go as Communication - British Go Association
Jan 17, 2022 · Yasuda is well known as the inventor of Capture Go, and his book says a lot about this. What he says came as a bit of a surprise to me at ...
[28]
[PDF] GO – A TOOL FOR LEARNING - Zen Machine
I began by showing him how to play First. Capture Go and he caught on quickly and ... George's homework was to teach his foster- father the rules of Atari Go and ...
[29]
In the Dark? Pro Biographies - British Go Association
Oct 14, 2017 · Her speciality in recent years has been touring western countries teaching children using the Yasuda method of capture Go. She has visited ...Michael Redmond · Shigeno Yuki · Kato Masao
[30]
Tenuki Go at Sensei's Library
Jul 26, 2008 · A Go Variant popularized (invented?) by ChessWhiz. Same as go, but with the additional restriction that no moves are allowed that are ...
[31]
https://www.eurogofed.org/pairgo/pairrules.htm
[32]
No Contact Go at Sensei's Library
Feb 14, 2022 · No Contact Go is arguably the simplest possible variant of Go. It introduces students to the Go board, the stones, the alternation of moves, ...
[33]
The 25th International Amateur Pair Go Championship
' The Professional Pair Go Championship is ranked among professionals as an important tournament, and this year we have held it for the twentieth time.
[34]
Pair Go - European Go Federation
The first European Pair Go Championship (EPGC) was held in Amsterdam in 1995. Since 1997 it has been a regular event, hosted by a different country each year.Missing: history origin
[35]
Pair Go Rules - European Go Federation
Pair Go is a fun and interesting way to play Go. Pair Go is played by teams of two partners - a male and a female player - the "Mixed Doubles" of Go.Missing: pass passes
[36]
https://www.britgo.org/files/london/docs/RenGoRules.pdf
[37]
International Amateur Pair Go Championship – IGF
1999, 21, Goto Naoko, Taga Bungo, Japan. 9th, 1998, 20, Kim Se-Sil, Park Sung-Kyun, S ... World Amateur Go Championship · World Mind Sports Games · Blog · Members.
[38]
Pair Go | British Go Association
Apr 12, 2018 · Pair Go is a form of "Rengo" - Go with more than one player per side where no conferring is allowed between the players.
[39]
https://www.dragongoserver.net/
[40]
Consultation Games - General Go Discussion - Online Go Forum
Mar 23, 2023 · Rengo implies no discussion, however, if you organize a group of players that all agree to playing a variant with discussion, beforehand, then ...
[41]
Zen Go at Sensei's Library
### Summary of Zengo Go Variant
[42]
Dragon Go Server: DGS
The Dragon Go Server (DGS) is a place where you can play turn-based Go (aka Baduk or Weichi) with other players from around the world in different time ...Missing: Zengo | Show results with:Zengo
[43]
Variants at Sensei's Library
Inventors have developed a dazzling variety of games with rules similar to Go in intuitiveness, wording, territorial objective, simplicity etc.Non-Variants · Variants restricting valid moves · Alternative rules of capture...Missing: credible | Show results with:credible
[44]
Rules - Go Variants
Missing: Micaël variant
[45]
Pair Go at Sensei's Library
Jun 29, 2023 · Arbitrary team-sizes are possible with N vs M players (e.g. 1:3, 2:2, 8:8, etc) · Rengo is a special case of Team-Go with 2 vs 2 players.
[46]
Computational and Human Intelligence in Blind Go. - ResearchGate
... rules of blind Go for making it interesting and pedagogically efficient. ... meaning of moves and. some patterns. Maybe there are nonetheless plenty of.<|control11|><|separator|>
[47]
Wayback Machine
### Summary of Shadow Go from https://web.archive.org/web/20080725130631/http://www.usgo.org/resources/downloads/deviantgo.pdf
[48]
Phantom Go at Sensei's Library
Jul 24, 2025 · Phantom Go is a variant of Go for two players and a referee. The players only see their own stones, but they don't know where the opponent plays.Missing: invention | Show results with:invention
[49]
Phantom Go - Chessprogramming wiki
Phantom Go is a Go variant where players see only their own board, and a referee sees all, using it to verify moves.Missing: rules invention
[50]
[PDF] A Comparison of Monte-Carlo Methods for Phantom Go
Phantom Go is a variant of Go which is complicated by the condi- tion of imperfect information. This article compares four Monte-Carlo methods for Phantom. Go ...
[51]
[PDF] Deviant Go - American Go Association
various topological variants, most commonly “toroidal go”, in which opposite ... plan Go”, which is played on an INFINITE board! Two recent fantasy novels ...
[52]
Toroid Go 101 - Nordic Go Dojo
Dec 20, 2020 · Because there are no edges in toroid go, forming territory (and even eyes) is a lot more complicated than in regular go.Missing: rules history
[53]
Little Golem - FAQ - Ed Collins
Apr 11, 2020 · Go has several variants including a vey large 37x37 size board, a variation known as Toroidal, a 19x19 Hahn point system, and a 19x19 Random 100 ...
[54]
https://senseis.xmp.net/?Keil
[55]
History of Go in Norway | Go i Norge
1983: In the spring Knut Roll-Lund introduces 3D Go. He builds a 9x9x9 `board' and it is tested in Kristiansand Go center. Dag ...
[56]
Freed Go - The Game of Go in 3D - Lewcid.com
Jul 3, 2004 · It is a complete Go application with all the rules and features you need to play a real game (undo, dead stone removal, KO checking, territory counting, etc).Missing: surrounding | Show results with:surrounding
[57]
[PDF] Basics - Cameron Browne
I'd wanted to devise a workable 3D Go variant for years. The biggest challenge was to find a way to reduce 3D freedom (so that groups can't perpetually ...
[58]
[PDF] My title - British Go Association
Games played this year included Four Colour Go, Blind Go and. Kriegspeil Rengo, but they avoided Zengo (Rengo with odd numbers of players). ... Joker Go as a ...<|separator|>
[59]
Hannoverscher Go-Verein e.V.
### Summary of Rund-Go Content
[60]
Hexagonal Go variants | BoardGameGeek
Apr 5, 2022 · Hexago is a variant of Go played on a hex grid. Stones are played on the spaces, not the intersections. The rules are different from those of Go.Hexagonal Go? - BoardGameGeekWhat are people's favorite Go variants or multiplayer versions of Go?More results from boardgamegeek.com
[61]
Keil at Sensei's Library
Oct 9, 2023 · Keil is a Go variant played on the intersections (points) of a hexagonal grid of triangles (or, equivalently, on the spaces of a [ext] hex grid).
[62]
TriGo :: igGameCenter
### Rules of TriGo
[63]
TriGo at Sensei's Library
Jan 25, 2010 · One difference to Go is that, since one is allowed to place two stones in one turn, it is necessary for groups to have three eyes to be alive.
[64]
[PDF] Abstract Games Magazine
Now AG18 contains a first analysis of Luis. Bolaños Mures' brilliant revisioning of hexagonal Go, the new game Keil. In my opinion, Keil has outstanding ...
[65]
Go - Other Pieces
There are two known variants based on this concept: (i) Omino Go (or Tetromino Go) by R. Wayne Schmittberger, explained in his book "New Rules for Classic ...
[66]
New Rules for Classic Games - Google Books
Title, New Rules for Classic Games ; Authors, R Wayne Schmittberger, R. Wayne Schmitterger ; Publisher, Random House Value Publishing, 1994 ; ISBN, 0517129558, ...
[67]
Coin Go at Sensei's Library
Feb 24, 2022 · In 'Coin Go', a stone cannot be played on certain intersections of the board by either player. A coin may be placed on these intersections ...
[68]
Pillar Go (or Coin Go) - Go Variants - Online Go Forum
Feb 19, 2022 · Pillar Go (or Coin Go) · Each player starts with 3 coins. · On a player's turn, instead of playing a stone, they may spend a coin to either place ...
[69]
Go Variants | BoardGameGeek
Jul 14, 2025 · The variants here are limited to two players taking alternating turns, with perfect information and no luck, and played on regularly tiled ...
[70]
Go variant with pieces placed as stacks to the height of their liberties
Jun 27, 2020 · A player on his turn places a stack on an empty intersection that has liberties, or a stone on an intersection that is surrounded on all four ...Missing: blocks | Show results with:blocks
[71]
Rules of Go - introductory at Sensei's Library
Dec 15, 2024 · Every stone or string of stones must have at least one adjacent point that is unoccupied. This unoccupied point is called a liberty.<|control11|><|separator|>
[72]
Go Variant Server - Nordic Go Dojo
Aug 19, 2020 · It is best for Black and White to leave the blue stones untouched – that is, unless all the black and white stones are alive and Blue is winning ...Missing: Europe | Show results with:Europe
[73]
Pixel Go Tournament Details - Nordic Go Dojo
Oct 27, 2020 · The tournament uses Variant Go Server rules, namely Chinese scoring (neutral intersections are worth points) and positional superko. White ...
[74]
Variant Ideas - #124 by yebellz - Go Variants - Online Go Forum
Apr 9, 2025 · shinuito: Pixel go you place 2x2 blocks, which is a Tetris shape. You could expand it to playing any Tetris shaped blob. 3 Likes.Page 7Page 6More results from forums.online-go.com<|control11|><|separator|>
[75]
Environmental Go at Sensei's Library
Jan 17, 2019 · Environmental Go, also called "coupon go", devised by Elwyn Berlekamp, adds an environment of simple gote to regular go.
[76]
Environmental Go Game | British Go Association
May 5, 2017 · Environmental Go. Environmental Go is Go played in the normal way, but with a stack of forty cards, numbered ½, 1, 1½, .., 19½, 20 as well as ...
[77]
[PDF] Go Thermography: The 4/21/98 Jiang–Rui Endgame
The Jiang–Rui Environmental Go Game. On April 21, 1998, Jiang Zhujiu and Rui Nai Wei, 9-dan professional players, played the first environmental go game.
[78]
https://senseis.xmp.net/?Sygo
[79]
Bid Go at Sensei's Library
In Bid Go the players bid for the right to move. There are a couple of variants of Bid Go: one by Bill Taylor and Alternate Bid Go by ChipUni.
[80]
Alternate Bid Go at Sensei's Library
### Summary of Alternate Bid Go
[81]
Go - Other Players
### Summary of Multi-player Go Rules from http://www.di.fc.ul.pt/~jpn/gv/players.htm
[82]
(PDF) Multi-player Go. - ResearchGate
Aug 7, 2025 · Multi-player Go is Go played with more than two colors. Monte-Carlo Tree Search is an adequate algorithm to program the game of Go with two ...
[83]
Diary, November 2010 - I write, therefore I am
Nov 1, 2010 · SyGo · Win Condition: winner is player with majority stones on board when board is full. · Capture by Conversion: stones convert to the opposite ...
[84]
Sygo at Sensei's Library
Mar 13, 2012 · Sygo is a territory game invented by Christian Freeling in november 2010, based on the 'Symple' move protocol and othelloanian' capture.Missing: C. | Show results with:C.
[85]
Moving forward looking back - MindSports
And mind, I don't even dislike draws if they keep a low profile. Sygo is a Go variant that uses the Symple move protocol. I remember inventing it a couple of ...Missing: C. | Show results with:C.
[86]
http://mindsports.nl/index.php/arena/sygo/
[87]
SyGo | Board Game - BoardGameGeek
Rating 6.6/10 (5) SyGo or Symple Go is a multi-move abstract strategy game for two players. There is a parallel to Go, because players try to control area with living groups.Missing: length | Show results with:length
[88]
Novels and Other Books Featuring Go | British Go Association
"Walking on Glass" by Iain Banks (1985) features three stories of converging ... One of these games is Open-Plan Go, played on an infinite board and ...
[89]
Infinite Board at Sensei's Library
Apr 14, 2022 · My first thought is to generalise the super-ko rule, such that the infinite goban is split into 19x19 areas which cannot be repeated anywhere ...
[90]
What are people's favorite Go variants or multiplayer versions of Go?
Apr 28, 2020 · Mitropia is essentially a multi-player version of Go, played on a hex board, with Cards that limit your movement options (similar to Onitama).Missing: Nepal Bangneki
[91]
Reversi Go - igGameCenter
The objective of Reversi Go to have a higher score than your opponent at the end of the game. Your score depends on the amount of territory you have surrounded ...
[92]
Othelloanian capture in Go variants - MindSports
In Go and its variants 'groups' are captured when they run out of liberties. Groups are defined as 'connected stones', which in square variants usually ...