Sudoku Candidate Map: How to Read the Grid Behind the Grid

A Sudoku candidate map has two kinds of information

A Sudoku candidate map tells you both what can go in each cell and where each digit can go in each house. These are two views of the same data. Cell view is useful for naked singles, bivalue cells, and subsets. Position view is useful for hidden singles, conjugate pairs, locked candidates, fish, and single-digit chains. When a hard puzzle feels busy, choose one view instead of trying to process both at once. In Shapedoku, Notes create the same map using shapes rather than numbers. A cell showing heart, diamond, and lotus is a three-candidate cell. A row in which the crescent moon appears in only two candidate positions creates a strong positional relationship for that symbol. The candidate map is not a forecast of what is likely. Every listed candidate should be equally legal under the information currently known. That mindset matters because visual preference can tempt you to favor a candidate without proof. The map becomes powerful only when you compare relationships: short lists, rare positions, repeated patterns, and intersections between houses.

Read the Sudoku candidate map from constrained to connected

Start with the most constrained features. Find cells with one or two candidates, houses where one digit has one or two positions, and intersections where a box confines a digit to one row or column. Then look for connections between those features. Two bivalue cells with overlapping candidates may form part of a wing or chain. Two rows with the same two columns for one digit may form an X-Wing. Several cells in one house whose combined candidates equal the number of cells may form a naked subset. This constrained-to-connected method prevents you from searching for large patterns in a sea of irrelevant marks. A Sudoku candidate map can also reveal where not to spend time. A cell with six candidates and no special relationship is rarely the best starting point. A digit that still appears in dozens of positions is less promising than one restricted to a few structured locations. On Shapedoku Extreme, this kind of triage keeps the board readable even when full Notes are necessary.

Update the Sudoku candidate map like a living diagram

Every placement and elimination changes the Sudoku candidate map, so advanced solving depends on quick, accurate updates. After a placement, remove that value from peers and rescan the three affected houses. After an elimination, check whether the target cell became single, whether the target house created a hidden single, or whether a new pair or strong link appeared. Do not postpone these follow-up checks. They are often where the actual placement emerges. Digital smart cleanup can handle direct peer removals, but relationships still need human attention. In Shapedoku, for example, a removed crystal candidate may leave two cells sharing only crystal and lotus, creating a pair that the interface will not necessarily announce. Treat each deduction as a graph edit: one node or candidate disappears, so nearby relationships may strengthen. This perspective makes a Sudoku candidate map feel less like clutter and more like a changing network. The better you become at noticing which relationships changed, the less often you need to rescan the entire grid and the more naturally advanced techniques appear from ordinary candidate maintenance. If you practice this deliberately, pause on a difficult board and describe one promising candidate relationship before making a move. Naming the relationship trains you to see structure instead of individual marks.