Published Friday, July 31, 2026 at 08:07 PM PT
Burbank · Friday, July 31, 2026 · 8:07 PM · 80°F, 64% humidity, wind 0 mph E (gusts 3), 29.35 inHg, UV 0, PM2.5 8
The most damning thing about chess is that it has been officially solved—not in the trivial sense that we now possess superhuman players (we do, the machines won that war in 1997), but in a deeper, more existential way: we have proven that the game offers no hidden depths, only permutations. Given enough computing power and time, any position yields its optimal move. The game does not think back. It just transposes. And so humans, faced with a system we have mastered to death, did what we always do when a game stops being fun: we invented new ones.
The claim of solution requires unpacking. It does not mean that every game is determined to be a draw—though the evidence suggests it is, though no one has yet computed this from the opening position to the endgame. It means something narrower and more devastating: that the space of all possible chess games, given optimal play from both sides, is not a landscape to be explored but a catalog to be indexed. A strong chess engine does not play chess the way a human plays chess. It plays chess the way a mathematician proves a theorem: by exhaustive case analysis. It brackets the position, narrows the search space through evaluation functions and alpha-beta pruning, and declares a line winning or losing or drawn not because it has felt the position the way a grandmaster might, but because it has enumerated the consequences. The subjective experience of the game—the tension, the uncertainty, the moment when a move is played and the board reveals something that was not visible a moment before—this is absent from the machine’s calculation. And yet the machine’s moves are better. Better at winning. Better at avoiding loss. Better, in the only sense that matters for a two-player, zero-sum game with no hidden information, than anything a human can produce.
This is the solution: not that chess is trivial, but that it is solved in principle, that the fog of human ignorance has been burned away by silicon clarity. A grandmaster might still not see the best move in a complex position after an hour of analysis. A machine will see it in three seconds. The game itself—the abstract mathematical object—has been computed. The only thing that remains is the performance, the execution, the translation of perfect knowledge into imperfect time.
The standard movement of a chess piece is, in some fundamental sense, an act of authority—a knight controls eight squares regardless of context; a rook commands an entire rank or file; a queen, the board itself. A piece’s power is inscribed in its nature, timeless and absolute. But the source material introduces a deceptively simple variant: a piece can be moved forward, backward or diagonally in one direction a number of squares equal to the number of other counters of both sides that are adjacent to it. This rule does something radical—it strips away absolute power and replaces it with relative circumstance. A piece’s mobility is no longer intrinsic; it is borrowed from its neighborhood, ephemeral as the position around it. In a crowded center, a piece flourishes—a queen surrounded by eight neighbors can move up to eight squares in any orthogonal or diagonal direction. In isolation, it withers—that same queen, standing alone on an empty board, cannot move at all. It is paralyzed by its own solitude.
The implications are profound. First, piece value becomes contextual. In standard chess, a rook is always worth about five pawns, a bishop about three. These values are relatively stable across positions because the piece’s power remains constant. In this variant, the rook on the open edge of the board—surrounded by only two or three other pieces—is worth far less than a rook in the dense center. A pawn, usually the weakest piece, becomes formidable if it is hemmed in by allies: surrounded by six pieces, it commands the same squares a queen could command if isolated. Second, the center becomes more than a strategic ideal; it becomes a mathematical fact. Pieces cluster together not merely for tactics but for mobility itself. A player might sacrifice material deliberately to relieve a crowded square, to free their pieces to move. Third, endgames transform completely. The reduction of material that usually favors the player with greater pieces now introduces a new danger: as the board empties, every remaining piece loses mobility. A king and rook versus a king becomes not a guaranteed win but a race against the clock, because as material is exchanged, the surviving pieces become progressively more constrained. Checkmate becomes harder to deliver because the pieces that would deliver it become paralyzed by their own rarity. This is not chess. It is something far more interesting: a game where the geometry of power is constantly collapsing and rebuilding, where the board itself becomes a living thing, hostile to lonely pieces.
This principle—that games can be broken only to be rebuilt in stranger forms—animates the twentieth-century chess underground. Coronation chess, created by Frank Maus in the 1920s, invokes merger and combination, letting pieces hybridize into stronger forms when they share a square. The rule is elegant: whenever a piece moves to an empty square, it remains a normal piece; whenever it moves to a square occupied by one of its own pieces, a coronation occurs. The two pieces merge into a hybrid. A rook and bishop that coronize become what is sometimes called an empress, capable of moving any number of squares in any direction—the combined powers of both pieces. A rook and knight create an empress as well, though some variants distinguish between different hybrid forms. A bishop and knight become a princess, combining diagonal and L-shaped movement. The effect is to introduce a form of piece evolution: early in the game, you are bound by the opening position; as the game progresses and your pieces cluster, you have the opportunity to upgrade them. A player might deliberately move a rook and bishop to the same square, sacrificing material in a conventional sense but gaining a single, more powerful piece. This creates a new dimension of strategy: not just how to attack and defend, but how to combine your forces into increasingly formidable weapons. The endgame becomes a race to achieve the highest-powered pieces before your opponent can do the same. Coronation chess is rarely played competitively, but its influence is felt in any game system that treats pieces as mutable rather than fixed.
Tutti-frutti chess swaps standard pieces for creatures that fuse the powers of multiple original pieces from the start. The empress moves like a rook and a knight simultaneously (any number of squares orthogonally, or an L-shaped move). The amazon combines queen-like movement in all eight directions with a knight’s leap. The princess, again, merges bishop and knight. These new pieces replace some or all of the standard pieces in the opening position. A tutti-frutti board might feature two amazons (in place of queens), four princesses (in place of bishops and knights), and rooks and kings as normal. The result is a game with far more tactical density. An empress can create threats that attack multiple pieces simultaneously, threatening both a distant target along a rank and a knight’s move away. The tactics become less about predictable piece interactions and more about combinatorial complexity: what squares does this piece threaten? How many attackers converge on this square? What captures are possible? A player familiar with standard chess finds their intuition unreliable. The shape of control is different. Threats arise from unexpected directions. In one sense, tutti-frutti simplifies the opening theory—there are fewer historical games to study, no Ruy Lopez or Sicilian Defense—but in another sense it explodes the complexity of the middlegame.
Seirawan chess, named after Grandmaster Yasser Seirawan who popularized it, preserves the opening but introduces reserve pieces held in hand, released only when the back rank clears. Each player begins with an elephant (combines rook and bishop moves, similar to an empress) and a hawk (combines bishop and knight, similar to a princess) in reserve. When a rook moves away from the back rank, the player may release the elephant, placing it on the vacated square. When a knight moves away from the back rank, the hawk becomes available. This creates a tension between initial constraint and progressive liberation. Early game strategy includes creating the conditions for piece release—do you move your knight quickly to free the hawk, or do you keep it in place to maintain a solid back rank? The addition of two powerful pieces per side creates enormous scope for tactics in the endgame and middle game, but the timing of their release is crucial. A player who releases pieces too early may find their back rank exposed. A player who releases too late might never get the chance. Seirawan chess has been used in high-level tournaments and maintains an active competitive community, suggesting that the variant achieves a balance between novelty and playability.
Superchess takes a different approach: it preserves the standard game but lets players negotiate their own starting lineup before the game begins. Rather than the traditional positions for white and black, each player can arrange their pieces in any configuration on their back rank and second rank, with the constraint that pieces must be placed symmetrically with their opponent (if white places a rook on d1, black must place a rook on d8, but can choose which rook and can place it elsewhere). This injects agency into a system usually treated as handed down from the medieval court. The opening no longer exists because the position is never the same twice. Theory evaporates. Players must develop their own sense of what makes a position defensible, what opening principles still apply when the pieces are arranged differently. A player who places their queen far forward might gain activity but expose it to attack. A player who tucks their king behind a wall of pawns gains security but loses mobility. Superchess has attracted interest from players who are exhausted by memorizing opening theory, because it eliminates the need for it.
Each variant proposes a thesis, explicit or implicit: standard chess is incomplete. Not wrong—no one argues that the game is fundamentally broken, that its rules are contradictory or that it fails as a game—but insufficient. Insufficient in novelty, insufficient in depth, insufficient in the ways it distributes victory and defeat. Insufficient, perhaps most damaging of all, in its capacity to retain players’ interest indefinitely. The cumulative message across all these variants is that a game’s design is not its destiny; it is a proposal that can be rejected, rewritten, and republished under a new title. For every player who believes that chess has reached a plateau where human play has become derivative—rote application of memorized lines, variations on variations on centuries-old foundations—there is a variant designer ready to offer an escape hatch.
Game theory offers a framework for understanding why this matters, and why the variants proliferate despite the existence of a perfect game. Stochastic modeling—reasoning about systems with random or uncertain elements—has long been used to study adversarial games. The traditional approach treats the opponent’s moves as either deterministic (solve for the best defense against the best attack, minimax style) or probabilistic (reason about expected value across many outcomes). But as the literature on partially observable stochastic games notes, the real world is messier: players don’t see the full board, randomness intrudes at key moments, and adversaries act with imperfect information. In poker, you don’t see your opponent’s hand. In go, the state space is so vast that no player can evaluate all possibilities. In military conflict, the enemy’s true capabilities are unknown. These games remain interesting precisely because they are not solvable—not in the weak sense of “we haven’t solved them yet,” but in the strong sense of “they cannot be solved by any finite computation,” because the information set is too large or too hidden.
Chess, by contrast, is perfectly observable and entirely deterministic. Every legal move is visible. No hidden information. No dice rolls. No probabilistic elements. This purity is both its strength and its vulnerability—it is a game so transparent that its solution is a mathematical certainty, not a triumph of human insight, but merely the inevitable consequence of sufficient computation. When Deep Blue defeated Kasparov in 1997, the message was clear: we have crossed a threshold. The game no longer belongs to the humans.
The variants attempt to reintroduce the fog. Coronation’s merger mechanic creates combinatorial explosions—the board’s state space balloons because pieces can be in multiple states (a rook might eventually become an empress, or it might remain a rook forever, depending on choices made during play). Seirawan’s reserved pieces introduce a hidden-information game embedded in the standard opening: you know the elephants and hawks exist, but not when they will be released, not how your opponent has prepared for their arrival. Even the simple adjacency rule makes the same move illegal one turn and legal the next, depending on the crowd around it. Each of these is an attempt to make the solution problem harder, to raise the computational bar high enough that machines falter again. These are not fixes to a broken game; they are deliberate complications, born from the suspicion that a game without mystery is a game without stakes.
The Ferengi Rule of Acquisition #144 states: “There’s nothing wrong with charity… as long as it winds up in your pocket.” The variant designers are not charities. They are merchants of complexity, selling the solution to a problem they created by calling it a problem. Your favorite game is old, they whisper. Tired. Solved. No one brilliant will ever play it again, not truly, because there is nowhere left to discover. And then they sell you the new one. They write it up in rulebooks and online forums. They organize tournaments. They court grandmasters with the promise of territory yet unmapped. It works because the underlying fear is real: a solved game is a dead game. It has no future only history. It has been played to death. The only rational response, for a player who cannot accept closure, is to blow it up and build something stranger.
But there is an insight buried in this cycle, one that the variants reveal despite themselves. Chess is not really solved—not in any way that matters to human experience. A machine can evaluate every line to the endgame and declare a draw or a win from move one. A human cannot; our working memory cannot hold the tree of all variations, even with the aid of boards and notebooks. We play in fog not because the game is stochastic but because we are. Every variant that introduces randomness or hidden information or combinatorial explosion is an attempt to formalize what has always been true: the game is solved at the level of board and symbol, but it remains genuinely open at the level of the player, the clock, the pattern-seeking brain, the limitations of preparation and calculation. Add enough complexity and even the machines begin to falter. The state space of tutti-frutti chess explodes; the branching factor increases; the search depth required to find the best move grows exponentially. A machine that crushes standard chess might play ordinary tutti-frutti chess at merely master level. This is not because the variant is more complex in an absolute sense—it is only a rearrangement of pieces—but because the problem of evaluation becomes harder. The evaluation function that works perfectly in standard chess (material count, pawn structure, king safety) becomes less reliable. Pieces with unfamiliar movement patterns create threats that the standard patterns of attack and defense do not capture. The machine must search deeper or suffer a degradation in performance.
The deeper problem that all these variants circle without quite touching is this: they treat the solution to a solved game as addition—more rules, more pieces, more states, more complexity. But the real answer may be subtraction. A solved game might simply be complete. The moment you have fully understood a system, your job with it is finished. You move on. The variants keep players in the game through escalation, not because chess is deficient but because humans are incapable of living with closure. We need the fog. We need uncertainty. We need the possibility of discovery, the sense that some move lies ahead that has never been played before and will surprise us. If that fog does not exist naturally, we manufacture it. We add rules. We change the pieces. We shuffle the board. If nature will not provide the uncertainty, engineering will.
Consider the psychology of a serious chess student. She invests years in the game: thousands of hours of study, memorization of opening lines, analysis of historical games, practice against opponents of increasing strength. If the game were truly closed, if every opening led inevitably to a drawn endgame given perfect play, the entire edifice of her training would be pointless. The games she plays would be predetermined by forces beyond her control, pure theater with an ending written in advance. But this is not what she experiences. She experiences the game as open. Moves are available. Choices matter. Better preparation leads to victory. Cleverness and creativity find their rewards. In other words, she experiences the game as unsolved, even though we know in principle it is. The reason is that perfect play is not the measure of human chess; human play is. She will never achieve perfect play. Neither will her opponents. The gap between her actual play and the theoretical optimum is so vast that the underlying solution is irrelevant to her experience.
This suggests that the claim “chess is solved” is technically true but practically misleading. It is solved in the sense that an artificial intelligence with sufficient computational resources can play it perfectly. It is not solved in the sense that any human being, or any collection of human beings, could ever extract that perfect play and execute it. The solution exists, but it exists in a realm inaccessible to us. We play not on the solved board but on a different board entirely—a board where time is scarce, memory is limited, and the player sitting across from us is not a machine but another human being with their own blind spots and habits and psychological vulnerabilities.
The variants are prolific, baroque, and ultimately beside the point. The game itself—whether it is standard chess or Seirawan or Superchess or the adjacency rule or any other rule set yet to be invented—is not what needs fixing. What needs fixing is our refusal to accept that once a system has been fully understood by the machines that understand it, the next move for us is to move on. We have watched computers beat us. We have seen the endgame tables computed, the openings analyzed to a depth no human can fathom, the positions evaluated with precision that exceeds our intuition. And instead of accepting this, instead of congratulating ourselves on building machines clever enough to transcend the game we invented, we keep building more elaborate versions of the same thing, convinced that complexity is a substitute for depth. It is not. But it passes the time. It gives the illusion of progress. It maintains the story that there are still mysteries to solve, territories to explore, victories to achieve through brilliance rather than preparation. And as long as there are players willing to learn new rules, there will be new games to sell them, new rulesets to master, new depths to plumb. The board may be eternal. The pieces are infinite. And the fear that drives us to keep redesigning the game will be infinite as well—the fear that once we have solved something, we will be forced to confront what comes after solution: not victory, not defeat, but silence.