Computer scientists have a formal framework for the kinds of problems that puzzle solvers tackle every day. They call it Constraint Satisfaction — and thinking explicitly in these terms, even informally, makes a wide range of puzzles significantly more tractable. You don't need a computer science degree to use this mindset. You just need to understand its three components and how they map onto the puzzles you already love.
The Three Components
Every constraint satisfaction problem has three elements: variables (the things you're trying to determine), domains (the possible values each variable can take), and constraints (the rules limiting which combinations of values are valid). In Sudoku, the variables are the empty cells, the domains are the digits 1–9, and the constraints are the rules that no digit repeats in any row, column, or box. In a crossword, the variables are the answer slots, the domains are valid words of the right length, and the constraints are that crossing letters must match.
Naming these elements makes the problem structure visible in a way that "just solve it" doesn't. Once you've identified your variables, domains, and constraints, the next steps become clearer: reduce domains by applying constraints, look for variables with only one value left in their domain (forced assignments), and propagate the consequences of each assignment through all connected variables.
Constraint Propagation: The Engine of Solving
Constraint propagation is the process of using a known or assumed value to eliminate options from other variables' domains. In Sudoku, placing a 7 in a cell propagates through its row, column, and box: every other cell in those units has 7 removed from its candidate list. This is exactly what you do when you maintain pencil marks and update them after each placement. Thinking of it explicitly as constraint propagation makes the process feel systematic rather than intuitive — which helps when the propagation is less obvious.
The most powerful insight from constraint propagation thinking is that it's always worth doing before anything more complex. Before trying an advanced technique, ask: have I fully propagated all known values through all constraints? In practice, most intermediate Sudoku errors come from incomplete propagation — a known value that wasn't fully eliminated from all affected cells. Explicit propagation sweeps prevent these errors entirely.
Arc Consistency: Finding Hidden Constraints
Arc consistency is a more advanced propagation concept. Two variables are arc-consistent if, for every value in one variable's domain, there exists at least one compatible value in the other's domain. If a value in one domain has no compatible partner in any connected domain, that value can be eliminated even though no direct constraint explicitly forbids it.
In Sudoku terms, this is equivalent to the logic behind Hidden Singles and Box-Line Reduction. When you notice that a digit can only go in one cell within a unit — not because that cell is forced directly, but because all other cells in the unit have eliminated it — you're applying arc consistency reasoning. Naming it helps you apply it more deliberately and consistently.
Backtracking: When to Commit and When to Retreat
In computer science, when propagation alone doesn't solve a problem, the algorithm makes a choice — assigns a value to a variable — and continues propagating. If a contradiction arises, it backtracks: undoes the choice and tries the next value. This is exactly what bifurcation is in Sudoku, and what "try an answer and check the crossings" is in crosswords.
The constraint satisfaction mindset makes backtracking feel principled rather than like giving up on logic. You're not guessing — you're making a controlled, reversible hypothesis and testing it systematically. The key is choosing which variable to bifurcate on carefully: always choose the variable with the smallest remaining domain (the cell with the fewest candidates, the word slot with the fewest valid options). This minimises the branching factor and keeps the search efficient.
Applying the constraint satisfaction mindset to your puzzle practice doesn't require any new solving techniques. It just reframes the techniques you already know as instances of a coherent general framework — which makes them easier to apply deliberately and harder to forget.