Constraint Propagation Sudoku: How One Move Cascades

Two engines: search and propagation

There are really two engines inside any Sudoku solver, human or machine. One is search: guessing a value and seeing if it leads to a contradiction. The other is propagation: using the rules to eliminate impossible options until the truth is forced. Good solvers lean on propagation first and only search when they must. Constraint propagation sudoku methods are the reason skilled players fill so many cells with certainty, never touching a guess on easier boards. The idea is that you are not adding information from outside; you are only drawing out what the symbols already on the grid quietly imply.

What constraint propagation sudoku does

Every empty cell starts with a list of candidates, the symbols it could still legally hold. Constraint propagation sudoku is the process of trimming those lists. When you place a symbol in a cell, you remove it from the candidate lists of every cell in the same row, column, and box. Those removals sometimes force new conclusions, which trigger more removals, and the effect ripples across the grid. The two simplest moves are wonderfully powerful. A naked single is a cell whose candidate list has shrunk to just one symbol, so that symbol must go there. A hidden single is a symbol that can legally appear in only one cell within a row, column, or box, even if that cell still lists other candidates.

Why it cascades and when it stalls

The satisfying cascade happens because the grid is densely connected. One filled cell touches up to twenty others, so resolving it can shrink twenty candidate lists at once; any of those might drop to a single option, which resolves that cell, which trims twenty more lists, and so on. This chain reaction is the beating heart of constraint propagation sudoku, and it is why a board can feel stuck and then suddenly open up in a rush. On harder puzzles, pure propagation eventually stalls, with every remaining cell holding two or more candidates and no single available. At that point a solver reaches for deeper patterns or falls back on careful search, trying a candidate and propagating its consequences to see if a contradiction appears. Extreme grids are designed to push you into these longer chains.

Notes make propagation visible

You can run constraint propagation sudoku in your head, but writing candidates down helps enormously, and this is where notes shine. Shapedoku lets you pencil in the possible shapes for each cell, and its smart notes cleanup automatically removes a candidate from peers when you place that shape, doing the bookkeeping for you. Watching stray candidates vanish as you commit a shape is propagation made visible, and the first hint is free too, often explaining why a shape fits, which is really a propagation lesson in disguise. The big mindset shift is trusting propagation over guessing: before you gamble on a cell, ask what the current placements already force, and follow those forced moves as far as they go. Most of the time that carries you further than you expect, and guessing becomes a last resort rather than a strategy.