Loading field map
Endgame tsumeshogi is the study and composition of tsume shogi: the formalized checkmate-problem genre of shogi (Japanese chess). A tsumeshogi problem presents a position in which one side (usually the attacker, playing the gold-general side) has a forced checkmate sequence against the opposing king, with every move by the defender being a legal response to check. The solver’s task is to find the unique or best sequence of checks that leads to mate, regardless of the defender’s choices. The subfield encompasses both the practical skill of solving these problems and the theoretical and compositional craft of constructing them.
A tsumeshogi position is governed by strict conventions that distinguish it from a mere endgame study or a tactical puzzle. The attacking side moves first and must give check on every move. The defending side must respond legally, but has no objective other than to prolong the mate or to choose the line that is most resistant. The attacker may use pieces already on the board and may also drop captured pieces from hand, a central mechanic unique to shogi. The defender, however, is not allowed to drop pieces; the defender may only move pieces already on the board or capture the checking piece. This asymmetry is fundamental: the attacker’s hand is a resource, while the defender’s hand is irrelevant.
The goal is not merely to checkmate in the abstract, but to do so under the tsume condition: every attacker move must be a check, and the defender must have no move that escapes the sequence. A position that is winning but requires a quiet (non-checking) move is not a valid tsumeshogi. This constraint makes the genre a pure exercise in forced mate, akin to a mathematical proof of a mating net.
The length of a tsumeshogi is counted in te (moves), where one te is a single move by either side. A "3-move" problem (3手詰) means three plies: attacker, defender, attacker, ending in mate. Problems range from the trivial (a few moves) to the monumental (hundreds of moves), with the latter often involving intricate dropping and re-capturing cycles.
The central intellectual question of tsumeshogi is deceptively simple: Given a position, is there a forced mate, and if so, what is the shortest or most elegant sequence? But beneath this lie deeper concerns.
For the solver, the stakes are practical: the ability to see forced checkmate sequences is a core skill in actual shogi play. Endgames in shogi are notoriously sharp, and the capacity to calculate a long checking sequence accurately—often with many pieces in hand—separates strong players from weak ones. Tsumeshogi solving is therefore a training method, a way to internalize patterns of mating nets, piece coordination, and the geometry of the board.
For the composer, the stakes are aesthetic and theoretical. A good tsumeshogi is not merely a valid sequence; it is a construction with a theme, a "trick" or idea that the solver must uncover. Composers aim for uniqueness (no alternative mating lines), economy (no superfluous pieces), and a satisfying logical structure. The theoretical interest lies in the boundaries of the form: how long can a forced mate be? What kinds of positions admit a pure checking sequence? What are the minimal material requirements for a mate against a king in a given corner?
The subfield also has a mathematical flavor. Because the rules are finite and deterministic, tsumeshogi positions are, in principle, decidable by exhaustive search. This has made them a natural domain for computer analysis. The question of whether a given position is a tsumeshogi is a finite search problem, and modern solvers can handle positions of enormous length. This computational angle has shifted some of the field's energy from pure human composition toward the verification and even generation of problems.
Tsumeshogi has a long and continuous history in Japan, where it developed as a distinct art form centuries ago. The earliest collections of shogi problems, known as tsume shogi or shogi zukō (将棋図巧), date to the Edo period (1603–1867). These were often composed by professional shogi players and served both as entertainment and as a demonstration of skill. The most famous early collection is the Shōgi Zukō (将棋図巧), a massive work containing hundreds of problems, some of extraordinary length. These early compositions were not "endgame tsumeshogi" in the modern sense—the term tsume simply meant a mating problem—but they established the formal conventions that the modern genre follows.
A crucial historical distinction is between tsume shogi (詰将棋), which are the formal problems described here, and hisshi (必至), which are "forced mate" positions in actual play where the attacker has a winning sequence but may include quiet moves. Tsumeshogi is a purified, artificial form; hisshi is a practical endgame concept. The modern subfield of endgame tsumeshogi is the direct descendant of the Edo-period problem collections, but it has been shaped by two later developments.
First, the rise of the shogi problem as a distinct compositional genre in the 20th century, with dedicated magazines, problem columns, and a community of composers. This professionalized the craft, leading to a greater emphasis on thematic content and technical correctness. Second, the advent of computer solving in the late 20th and early 21st centuries. Computer solvers have verified the correctness of many classical problems, sometimes finding flaws (cooks, or unintended alternative solutions) and have also been used to generate new problems. This has not replaced human composition but has added a layer of verification and a new tool for exploration.
Within the subfield, one can identify several distinct but overlapping traditions of practice and thought.
The most widespread engagement with tsumeshogi is as a training tool. This tradition is practical and pedagogical. Its central assumption is that repeated exposure to mating patterns builds the solver's intuition and calculation speed. The method is simple: solve problems of increasing length and complexity, starting with 3-move and 5-move problems and progressing to longer ones. The emphasis is on pattern recognition—the "gold and silver mating net," the "edge mate," the "drop on the head" (頭金, atama-kin)—and on the discipline of reading all defender responses.
This tradition does not theorize much about the nature of tsumeshogi; it treats the problems as a given corpus. Its limits are practical: solving many problems does not automatically make one a good composer, and the transfer of solving skill to actual play is real but imperfect, since real endgames rarely present the pure checking sequences of tsumeshogi. Nevertheless, this tradition is the most influential in terms of sheer numbers of practitioners, and it has shaped the modern canon of "standard" problems used in shogi training.
The compositional tradition is the artisanal core of the subfield. Its practitioners are concerned with creating new problems that are both valid and aesthetically pleasing. The central problem they address is: How does one construct a position that yields a unique, forced, and elegant mating sequence?
Composers work with a set of aesthetic principles that have evolved over time. Uniqueness is paramount: a problem with two different mating lines is considered flawed (a "cook"). Economy means that every piece on the board must serve a purpose; a piece that merely blocks a square without being captured or forcing a specific move is considered wasteful. Thematic content refers to the presence of a recognizable idea—a "trick" such as a sacrifice, a deflection, a pin, or a "waiting move" that forces the defender into a zugzwang-like state. The most celebrated problems are those that combine a surprising key move (the first move, which is often counterintuitive) with a long and intricate sequence.
The compositional tradition has its own internal debates. One long-standing tension is between "orthodox" problems, which adhere to strict economy and natural-looking positions, and "artistic" or "task" problems, which prioritize a single extreme feature (e.g., the longest mate, the fewest pieces, the most captures) even at the cost of naturalness. Another tension is between the classical style, which favors clear thematic ideas, and a modern style that embraces complexity and length for their own sake. These are not rival schools in a strict sense—many composers work in both modes—but they represent different values within the craft.
The computational approach treats tsumeshogi as a formal system to be analyzed algorithmically. Its central question is: Given a position, can we determine whether it is a tsumeshogi, and if so, what is the optimal sequence?
This approach became practically relevant with the development of strong shogi programs. Because tsumeshogi positions are finite and the checking condition restricts the search tree, they are tractable for modern computers even at lengths of hundreds of moves. The methods used are variants of depth-first search with transposition tables, often specialized to exploit the structure of checking sequences. The key insight is that the attacker's hand is a finite resource, and the defender's options are limited by the need to respond to check, so the search space, while large, is manageable.
The computational approach has several important consequences. It has been used to verify the correctness of classical problems, sometimes overturning long-held beliefs about a problem's validity. It has also been used to generate new problems, either by random search or by guided construction, though the aesthetic quality of computer-generated problems is generally considered inferior to that of human compositions. More theoretically, computer analysis has established upper bounds on the length of tsumeshogi for given material, and has explored the boundaries of what is possible under the strict checking rule. This approach does not replace the human traditions; rather, it provides a tool that both solvers and composers can use, and it has opened a new frontier of "extreme" tsumeshogi that are too long or too complex for any human to solve unaided.
A smaller but important tradition is the study of tsumeshogi as a historical and cultural artifact. This approach examines the development of the genre over time, the attribution and dating of classical problems, the evolution of compositional conventions, and the social context of problem composition (e.g., the role of problem collections in the professional shogi world). Its central question is: How did tsumeshogi become what it is, and what do the problems tell us about the people who made them?
This tradition is primarily scholarly. It has established, for example, that many classical problems were transmitted orally or in manuscript before being printed, and that the modern canon of "standard" problems is a relatively recent construction. It also documents the shift from problems as practical training aids to problems as an autonomous art form. This approach is not in conflict with the others; it provides the historical context that the solving and compositional traditions often take for granted.
These four traditions are not mutually exclusive, and in practice they overlap considerably. A strong player who solves tsumeshogi for training may also appreciate the aesthetic qualities of a well-composed problem. A composer may use a computer to verify that a new problem is sound. A historian may also be a solver. The relationships are best understood as complementary rather than competitive.
The most significant interaction in recent decades has been between the compositional and computational traditions. Computer verification has become a standard step in the composition process, and the discovery of a cook in a classical problem is a notable event that can change the canon. Conversely, the aesthetic principles of the compositional tradition have influenced how computational results are presented: a computer-generated problem that is valid but ugly is unlikely to be published in a problem magazine.
The solving tradition and the computational approach also interact, though less directly. Computer solvers have made it possible to create problems of a length that no human could ever solve, and these "monster" tsumeshogi exist in a strange space: they are valid problems, but they are not usable for training. This has led to a de facto division of labor: short problems (up to roughly 15 moves) are for human solving; longer problems are for computer analysis or for the aesthetic appreciation of their construction.
The present landscape of endgame tsumeshogi is characterized by a healthy coexistence of all four traditions. The solving tradition remains the most popular entry point, with countless books, websites, and mobile apps offering graded problems. The compositional tradition is sustained by a dedicated community of composers who publish in specialized magazines and online forums, and who continue to push the boundaries of length and thematic complexity. The computational approach is now a standard tool, both for verification and for exploring the theoretical limits of the form. The historical-philological approach is a smaller but active field, with ongoing work on the transmission and interpretation of classical collections.
One notable feature of the current scene is the globalization of a previously Japan-centric activity. While the terminology and the classical canon remain Japanese, the internet has allowed non-Japanese speakers to participate in solving and even composing. This has led to some adaptation of conventions (e.g., the use of Western piece symbols) but has not fundamentally altered the rules of the genre.
Another feature is the continued vitality of the "long problem" as a frontier. The longest known tsumeshogi, composed in the classical era, runs to over a thousand moves. Modern composers, aided by computers, have created problems of comparable or greater length, though the aesthetic value of such monsters is debated. The question of what makes a tsumeshogi good—as opposed to merely valid—remains a matter of human judgment, and this is unlikely to change.
In sum, endgame tsumeshogi is a small but richly structured subfield. It is defined by a strict formal constraint—the pure checking sequence—that gives it a mathematical clarity, while its aesthetic and pedagogical dimensions give it a human depth. The four approaches described here are not rivals but facets of a single, multifaceted practice. For the educated newcomer, the most useful map is to understand the rules, appreciate the history, and then engage with the problems themselves—whether as a solver, a composer, or a student of the form.