There are video games whose influence can be measured by sequels, sales, or the size of the company that produced them. Then there are games whose basic idea becomes so familiar that players stop thinking of it as belonging to any one moment in history.
Sokoban belongs to the second category.
A worker stands inside a warehouse. Boxes must be moved onto designated spaces. They can be pushed but not pulled, and a single careless move can leave a crate permanently trapped against a wall or in a corner. There are no elaborate controls and no complicated rules to memorize. Yet from those few restrictions came one of the most enduring puzzle designs in video game history.
Behind it was Hiroyuki Imabayashi, a Japanese hobbyist who entered game development before the profession had fully defined itself. He learned programming after a friend left a computer at his house, built games largely because he found the process fascinating, created Sokoban for friends to struggle with, and eventually founded Thinking Rabbit to sell it. From there his career expanded into adventure games, console production, writing, level design, and decades of stewardship over the puzzle he had created.
The story of Imabayashi is therefore larger than the story of one famous puzzle. It is also the story of an early independent Japanese game developer who entered the business when a small group of people, a home computer, cassette tapes, graph paper, and a strong idea could still produce something capable of lasting for generations.
Three Days in Front of a Computer
Imabayashi did not begin as a trained computer engineer. In a remarkably detailed interview published in the December 1983 issue of the Japanese computer magazine LOGiN, he explained that his introduction to computers had come almost by accident.
Several years earlier, a friend had brought a Sharp MZ computer to his home. The machine was heavy enough that the friend eventually left it there rather than carrying it back. Imabayashi knew little about computers, but he became fascinated almost immediately. According to the contemporary interview, he spent roughly three days and nights in front of the machine with its manual beside him, learning as he experimented. When he finally tried to stand, his legs had become so weak from sitting that they would barely support him.
The first program he completed was a simple landing game. The exact details differ slightly between his 1983 interview and his much later recollections, but both accounts agree on the essential point. His introduction to computers rapidly became an obsession with making games. He soon purchased a NEC PC-8001 of his own and produced several more programs.
For Imabayashi, the computer was not simply a piece of business equipment. He remembered turning one on and feeling as though another dimension had opened in front of him. It offered a combination of logic and imagination that appealed to him immediately.
That fascination mattered because Imabayashi entered the Japanese computer-game world at a moment when the boundaries between hobbyist and professional were unusually thin. There were no enormous development teams behind his earliest work. There was a person sitting in front of a home computer, experimenting until an idea became playable.
The Puzzle That Became Sokoban
The first version of Sokoban emerged in 1981.
On the official Sokoban website, Imabayashi remembers creating the game during the spring, programming it in BASIC after becoming interested in the idea of a character pushing objects. His central design thought was deceptively clever. The objects that the player was supposed to put away should also become the objects that interfered with putting everything away.
That single decision created the foundation of Sokoban.
In later correspondence published by Sokoboxes, Imabayashi identified the original prototype as a PC-8001 BASIC program. He made it as a hobby and invited friends to his home to play. He took pleasure in watching them struggle through the puzzles. He also confirmed that the first stage he designed survived into the commercial PC-8801 version.
The idea of pushing objects did not appear from nowhere. Decades later, Imabayashi recalled being influenced by a Hudson computer game in which objects could be moved to serve as barriers. He identified the game as Aldebaran. What interested him was not necessarily the larger action game surrounding the mechanic, but the act of moving an object into position. He separated that movement from its original context and transformed it into a puzzle.
The result was built around limitations rather than possibilities. A box could be pushed but not pulled. The player could move only if space existed on the opposite side of the box. A crate pushed into the wrong corner could be lost permanently. The challenge therefore did not depend upon reflexes. It depended upon anticipating consequences.
Every push mattered.
That principle helped make Sokoban unusually durable. The rules can be understood almost immediately, but understanding the rules does not mean understanding a level. A player has to think several moves ahead while also recognizing that solving one immediate problem can create another somewhere else.
Building Warehouses on Graph Paper
The programming behind early Sokoban was relatively simple. Creating good puzzles was not.
Imabayashi originally designed stages on graph paper. In 1983 he openly acknowledged that the technical sophistication of the program itself was not extraordinary. What mattered was the idea and the quality of the puzzles placed inside it. His approach reflected a philosophy that would continue throughout his career. Technology was useful, but technical complexity by itself did not make a good game.
The early levels also reveal that Imabayashi was still discovering what his own system could do. He later explained that he initially underestimated the depth of Sokoban. Thinking he might eventually exhaust its possibilities, he experimented with breakable walls in some early stages. The feature proved controversial because some players felt that discovering a hidden or destructible wall violated the logic of a pure puzzle. Imabayashi did not continue the idea in later designs.
More important was what he learned from the experience. The simple act of pushing boxes under strict rules could generate far more complexity than he had initially imagined.
This realization also changed how Sokoban levels were created. The sequel incorporated a level editor, allowing puzzles to be produced more systematically and by people other than Imabayashi. In the 1983 LOGiN interview, he compared Sokoban to tsume shogi, the composed checkmating problems associated with Japanese chess. The comparison is revealing. A Sokoban stage was not merely a map. It was a carefully constructed problem that another person could attempt to solve.
That idea allowed Sokoban to grow far beyond the twenty stages of an early commercial edition. The rules remained stable while the number of possible puzzles became enormous.
From a Record Store to Thinking Rabbit
Sokoban might have remained a game shared among friends if someone had not recognized its commercial potential.
Imabayashi’s wife’s family operated a record store, and a computer corner was established there during the early Japanese personal-computer boom. An improved version of Sokoban appeared on the machine in the shop. According to the 1983 interview, a customer saw the game and suggested that it could become a hit. On his official retrospective, Imabayashi remembered a software salesman making the decisive observation that the game could sell and encouraging him to produce it commercially.
Imabayashi responded in the spirit of an early independent developer. He purchased equipment for duplicating software, prepared cassette tapes, worked on packaging, and handled much of the process himself. The 1983 feature even photographed original PC-8801 cassette copies that he had personally assembled.
Out of that process came Thinking Rabbit.
The exact chronology is one of the more complicated parts of Imabayashi’s story. The modern official Sokoban history dates PC-8801, FM-7, and PC-8001mkII versions to December 1982 and identifies Thinking Rabbit as their publisher. Imabayashi’s own official recollection similarly states that he formed Thinking Rabbit and released Sokoban in 1982.
The December 1983 LOGiN article presents a somewhat different timeline, describing Sokoban as having been released in May and Thinking Rabbit as a company established only several months before the article appeared. Surviving historical references therefore do not line up perfectly. Rather than hiding the discrepancy, it is safer to say that Sokoban was created in 1981 and entered commercial publication during Thinking Rabbit’s formation in the early 1980s, with the current official chronology placing the first releases in December 1982.
Whatever the exact month of its commercial debut, the important transition is clear. A home project written for friends had turned into the foundation of a software company.
Thinking Rabbit Begins to Grow
Thinking Rabbit remained a remarkably small operation in its earliest period.
The 1983 LOGiN profile described five people working with Imabayashi. The staff included specialists in programming, hardware, assembly language, and design. Many still had other careers and worked from their homes. They gathered on Saturday nights to exchange ideas and refine projects into the early morning hours.
The company’s name reflected its circumstances. Imabayashi later remembered drawing a rabbit as a mascot in a notebook, with the rough logo eventually being refined by a designer friend. The word “Thinking” also fit the kind of games the young company believed it could make successfully. The team did not initially possess the technical resources of larger developers, but Imabayashi believed puzzles and adventure games could compete through ideas.
That philosophy would lead Thinking Rabbit in directions that are sometimes overlooked when Imabayashi is remembered only as the creator of Sokoban.
Murder Mysteries and Adventure Games
Imabayashi was deeply interested in storytelling.
By the time LOGiN visited Thinking Rabbit in 1983, he was already working beyond puzzle design. Kagiana Satsujin Jiken, a murder mystery adventure, demonstrated another side of his interests. Imabayashi wanted players to investigate a crime and ultimately identify the murderer. He originally considered a more open text-input interrogation system, but the limitations of computer memory forced compromises. Rather than consume most of the available memory with text processing, he emphasized graphics.
The result reportedly contained more than eighty images, a substantial visual component at a time when many computer adventures remained heavily dependent on text. Thinking Rabbit even operated a telephone hint service, and Imabayashi sometimes spoke directly with players who called because they had become stuck. He enjoyed discovering which parts of his games people found difficult.
His interests in mystery fiction went further. In the same interview, Imabayashi discussed plans for Dōkeshi Satsujin Jiken, or The Clown Murder Case, built around a formal solution in which the player would identify the murderer and murder weapon. He referenced the tradition of Ellery Queen mysteries, where the reader could be directly challenged to solve a case before the final revelation.
Even more ambitious was a science-fiction, romance, and murder story involving time travel. The project eventually became Casablanca ni Ai o: Satsujinsha wa Jikū o Koete, released several years after Imabayashi had first described the concept. His career was increasingly becoming that of a producer, planner, and writer rather than simply the programmer of Sokoban.
Ideas Before Technology
Imabayashi’s 1983 comments provide an unusually clear statement of his design philosophy.
He admitted that he did not consider himself an exceptional programmer. He consulted other members of the company when technical problems exceeded his expertise. That did not trouble him. Technical skill could be hired or learned. What interested him most was the idea at the center of the game.
He warned that developers could become lost in technology for its own sake. A game did not become interesting merely because its programming was complicated.
In that sense, Sokoban was almost the perfect demonstration of his argument. Its greatness did not depend upon sophisticated graphics or enormous amounts of code. Its strength came from rules that interacted in interesting ways.
Two years later, the October 1985 issue of BEEP! included Imabayashi among fifty-nine game developers responding to a questionnaire. He was listed as twenty-nine years old and associated with Thinking Rabbit and Sokoban. When asked which of his games made him particularly proud, he chose Sokoban. When asked how someone should become a game designer, his answer centered on playing and learning to play well. He identified curiosity and a love of games among the important qualities of a designer.
The answers were often humorous, but they fit the developer visible in the earlier interview. Imabayashi approached game creation as exploration.
Sokoban Leaves Japan
Meanwhile, the puzzle that had started on Imabayashi’s PC-8001 was spreading.
The official Sokoban chronology records versions for a remarkable variety of Japanese computers and consoles. Releases appeared for NEC, Fujitsu, Sharp, MSX, Sega hardware, Nintendo systems, the PC Engine, Mega Drive, Game Boy, and numerous other platforms. Sokoban 2 expanded the concept and introduced an editor, while later versions greatly increased the available collection of puzzles.
Contemporary reporting suggests that the game became a considerable commercial success. By the late 1980s, American computer publications were describing Sokoban as an established Japanese hit. Spectrum HoloByte licensed the game for Western computer systems, releasing Soko-Ban for machines including the IBM PC, Commodore 64, and Apple II.
A November 1988 report in inCider stated that the game had sold more than 400,000 copies in Japan before Spectrum HoloByte brought it to the American market. Whatever differences existed among the Japanese editions included in that figure, the number illustrates how far Imabayashi’s home-built puzzle had traveled.
The international release also helped separate Sokoban from any single machine. It became a form of puzzle that could be reproduced almost anywhere. Clones, adaptations, fan-made stages, and later implementations carried the system across generations of hardware.
Eventually the Japanese title itself became the name of an entire style of puzzle.
More Than the Creator of Sokoban
Remembering Imabayashi only as the man who invented Sokoban understates the rest of his career.
Original credits document his involvement in a wide range of roles. On Thinking Rabbit’s 1988 Famicom title 8 Eyes, he was credited as producer. The game was an action platformer, far removed from the quiet warehouse puzzles that had established the company.
His work continued across other genres and generations of hardware. He produced later projects, contributed to stories and scenarios, worked with puzzle data, and remained involved with Thinking Rabbit’s development output. In 1995, the Super Famicom adaptation of Record of Lodoss War credited Thinking Rabbit with graphics and programming and listed Imabayashi among the producers.
By 1998, his name still appeared in Sokoban Basic 2, where he was credited with editing warehouse-stage data.
The range is significant. Imabayashi’s career moved between designer, producer, writer, planner, and puzzle creator. The pattern also reflects Thinking Rabbit itself. The studio that began with a deceptively simple puzzle eventually worked on adventure games, role-playing games, action titles, and licensed projects.
The Long Life of Thinking Rabbit
Thinking Rabbit eventually ceased regular game-development operations around 2000. In his later correspondence, Imabayashi described the company as legally remaining in existence but inactive, with the Thinking Rabbit name continuing primarily as a brand. He also explained that Falcon managed his rights connected with Sokoban.
The official Thinking Rabbit website still carries Imabayashi’s name and copyright in connection with the series, while the modern official Sokoban site identifies the game as his copyrighted creation and records Falcon’s role in managing the property.
For Imabayashi, however, the end of Thinking Rabbit’s most active period did not mean the end of game design.
His professional profile identifies him as Thinking Rabbit’s chief executive from the early 1980s until 2001 and later as representative director of Pictria. In 2019, Pictria released TOSCA, a two-player strategy game for smartphones. GameBiz described it as Imabayashi’s first new game in roughly twenty years. The game was presented as a perfect-information contest in the tradition of games such as chess and shogi, relying on strategy rather than random events.
Decades after Sokoban, he was still returning to systems built around thinking, positioning, and the consequences of individual moves.
Why Hiroyuki Imabayashi Matters
The lasting importance of Hiroyuki Imabayashi does not come from technological spectacle.
Sokoban was not remembered because it produced the most advanced graphics of its generation. It did not need cinematic storytelling, enormous amounts of memory, or specialized hardware. Its achievement was conceptual.
Imabayashi took the act of pushing an object and imposed a handful of restrictions upon it. Boxes could move forward but could not simply be pulled back. Walls created permanent consequences. The player’s own position mattered because reaching the correct side of a box was necessary before the next push could happen. From those restrictions came an enormous number of possible problems.
That is elegant game design.
It also explains why the game survived technological change so easily. A Sokoban puzzle does not depend upon the PC-8001, PC-8801, Famicom, Game Boy, personal computer, smartphone, or any other specific machine. Those systems are simply different places where the underlying idea can live.
Imabayashi once designed his stages with pencil and graph paper because the puzzle existed conceptually before it existed electronically. More than forty years later, people continue to create new stages using almost exactly the same logic.
His career also offers a window into an unusually formative period in Japanese video game history. He entered computing as an enthusiast, taught himself enough programming to experiment, made games for friends, personally prepared software for sale, built a small company around an unexpected success, and then expanded into production, writing, adventure design, and console development.
The industry around him transformed almost beyond recognition.
The warehouse did not.
A player still approaches a box. A destination waits somewhere across the room. One push can solve the problem, or quietly make the entire puzzle impossible.
That enduring moment of hesitation is Hiroyuki Imabayashi’s place in video game history.
Sources & Further Reading
Imabayashi, Hiroyuki. “ごあいさつ” [Greetings]. Sokoban Official Site. Accessed October 13, 2026. https://www.sokoban.jp/greeting.html.
Falcon Co., Ltd., and Thinking Rabbit. “倉庫番の歴史” [The History of Sokoban]. Sokoban Official Site. Accessed October 13, 2026. https://sokoban.jp/history.html.
Falcon Co., Ltd., and Thinking Rabbit. Sokoban Official Site. Accessed October 13, 2026. https://sokoban.jp/.
Thinking Rabbit. Thinking Rabbit Official Site. Accessed October 13, 2026. https://thinkingrabbit.jp/.
“考えるウサギはパソコンの野を駆ける!” [The Thinking Rabbit Runs Through the Fields of the Computer!]. LOGiN, December 1983, 136–39. Internet Archive. https://archive.org/details/login-magazine-dec-1983-thinking-rabbit/page/5/mode/2up.
Shmuplations. “Thinking Rabbit: 1983 Developer Interview.” Translation of an interview with Hiroyuki Imabayashi originally published in LOGiN, December 1983. Accessed October 13, 2026. https://shmuplations.com/thinkingrabbit/.
“59 Developers, 20 Questions.” BEEP!, October 1985. Translated by Shmuplations. Accessed October 13, 2026. https://shmuplations.com/20questions1985/.
Montiers Aguilera, Carlos. “My Conversation with Mr Hiroyuki Imabayashi.” Soko Boxes. May 22, 2022. https://sokoboxes.com/articles/my-conversation-with-mr-hiroyuki-imabayashi/.
Tanaka, Junji, Hiroyuki Imabayashi, and Masami Ishii. THE 倉庫番 [The Sokoban]. SCALE, 1986. Internet Archive. https://archive.org/details/sokoban-book/mode/1up.
“倉庫番［番外編］” [Sokoban Extra Edition]. PC Magazine, August 1983, 52–56. Internet Archive. https://archive.org/details/pc-magazine-aug-1983-sokoban-extra-edition/page/n1/mode/2up.
“倉庫番2” [Sokoban 2]. LOGiN, July 1985, 76. Internet Archive. https://archive.org/details/logi-n-july-1985/LOGiN%20-%20July%201985/page/n77/mode/2up.
“ソフトハウス訪問” [Visiting a Software Company]. POPCOM, June 1984, 131. Internet Archive. https://archive.org/details/popcom-198406/page/130/mode/2up.
“倉庫番トラの巻” [The Sokoban Strategy Guide]. POPCOM, July 1984, 101. Internet Archive. https://archive.org/details/popcom-198407/page/100/mode/2up.
“作者が語る自信のニューソフト” [The Creator Proudly Presents His New Software]. POPCOM, March 1985, 28–30. Internet Archive. https://archive.org/details/popcom-198503/page/28/mode/2up.
Spectrum HoloByte. Soko-Ban. Computer game manual. Alameda, CA: Spectrum HoloByte, 1988. Internet Archive. https://archive.org/details/stx_Spectrum_HoloByte_Soko-Ban/page/n4/mode/1up.
Tandy Corporation. Soko-Ban. Computer game manual. 1988. Color Computer Archive. https://colorcomputerarchive.com/repo/Documents/Manuals/Games/Soko-Ban%20%28Tandy%29.pdf.
Wagner, Roy. “Puzzling Encounters.” Computer Gaming World, no. 47 (May 1988): 42–43. https://archive.org/details/Computer_Gaming_World_Issue_47.
Lesser, Hartley, Patricia Lesser, and Kirk Lesser. “The Role of Computers.” Dragon, no. 132 (April 1988): 84. https://archive.org/details/DragonMagazine260_201801/DragonMagazine132/.
Low, Lafe. “News Line: Made in Japan.” inCider, November 1988, 14. https://archive.org/details/inCider_1988-11.
Kaneko, Yoichi, ed. パソコンゲーム80年代記 [PC Games: Chronicle of the 1980s]. Tokyo: Tatsumi Publishing, 1990. https://archive.org/details/pcgame80.
テレビゲーム: 電視遊戯大全 [Video Games: Complete Collection of TV Games]. Kawade Kosan, 1988. https://archive.org/details/tv-games/tv-games/mode/1up.
Game Preservation Society. “LOGiN, December 1983.” Game Preservation Society Catalog. Accessed October 13, 2026. https://www.gamepres.org/en/catalog/?id=8814001635008.
Game Developer Research Institute. “Thinking Rabbit.” Accessed October 13, 2026. https://gdri.smspower.org/wiki/index.php/Thinking_Rabbit.
MobyGames. “Hiroyuki Imabayashi.” Accessed October 13, 2026. https://www.mobygames.com/person/257256/hiroyuki-imabayashi/.
MobyGames. “Thinking Rabbit Inc.” Accessed October 13, 2026. https://www.mobygames.com/company/2765/thinking-rabbit-inc/.
MobyGames. “8 Eyes Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/10041/8-eyes/credits/nes/.
MobyGames. “Casablanca ni Ai o: Satsujinsha wa Jikū o Koete Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/54750/casablanca-ni-ai-o-satsujinsha-wa-jiku-o-koete/credits/pc98/.
MobyGames. “Perfect Soko-ban Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/71627/perfect-soko-ban/credits/pc98/.
MobyGames. “Boxxle II Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/16329/boxxle-ii/credits/gameboy/.
MobyGames. “Toki o Koeta Tegami Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/203973/toki-o-koeta-tegami/credits/3do/.
MobyGames. “Record of Lodoss War Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/43852/record-of-lodoss-war/credits/snes/.
MobyGames. “Minton Keibu no Sōsa File: Dōkeshi Satsujin Jiken Credits.” Accessed October 13, 2026. https://www.mobygames.com/game/223359/minton-keibu-no-sosa-file-dokeshi-satsujin-jiken/credits/sega-saturn/.
Raido. “Kyuukyoku no Sokoban.” PlayStation Staff Credits. Accessed October 13, 2026. https://raido.moe/staff/ps1/ps1_kyukyoku_no_sokoban.html.
Raido. “Minakata Hakudō Tōjō.” PlayStation Staff Credits. Accessed October 13, 2026. https://raido.moe/staff/ps1/ps1_minakata_hakudou_toujou.html.
Raido. “Minton Keibu no Sōsa File: Dōkeshi Satsujin Jiken.” PlayStation Staff Credits. Accessed October 13, 2026. https://raido.moe/staff/ps1/ps1_minton_keibu.html.
Raido. “Sokoban Basic 2.” PlayStation Staff Credits. Accessed October 13, 2026. https://raido.moe/staff/ps1/ps1_sokoban_basic_2.html.
Raido. “‘98 Kōshien.” PlayStation Staff Credits. Accessed October 13, 2026. https://raido.moe/staff/ps1/ps1_98_koushien.html.
Ellison, Carol. “Special Report: Why Japan Can’t Write Software.” PC Computing 1, no. 5 (December 1988): 113–14. Internet Archive. https://archive.org/details/PC_Computing_1988_12/page/n116/mode/1up.
Takahashi, Pyonta. “シンプルisパーフェクト──倉庫番” [Simple Is Perfect: Sokoban]. BEEP, February 12, 2025. https://www.beep-shop.com/column_entry/67956/.
Sasaki, Jun. “感動のアドベンチャーストーリー『カサブランカに愛を 殺人者は時空を超えて』.” AKIBA PC Hotline!, November 11, 2025. https://akiba-pc.watch.impress.co.jp/docs/column/retrosoft/2060869.html.
GameBiz. “ピクトリア、完全情報型対局ゲーム『TOSCA』をリリース.” March 20, 2019. https://gamebiz.jp/news/234278.
Nintendo. “The Sokoban.” Nintendo Store. Accessed October 13, 2026. https://www.nintendo.com/us/store/products/the-sokoban-switch/.
Junghanns, Andreas. Pushing the Limits: New Developments in Single-Agent Search. PhD diss., University of Alberta, 1999. https://www.collectionscanada.gc.ca/obj/s4/f2/dsk1/tape7/PQDD_0033/NQ46861.pdf.
“An Analysis of Single-Player Monte Carlo Tree Search Performance in Sokoban.” Expert Systems with Applications 192 (2022). https://www.sciencedirect.com/science/article/pii/S0957417421015372.