Method beats hope in a good sudoku.
Sudoku has a reputation for pure logic, which is fair enough, but the useful distinction is between methods that place a digit directly and methods that only remove possibilities. If you do not separate those jobs, the puzzle starts to feel like fog. You make neat pencil marks, admire them briefly, and then somehow end up exactly where you started.
The foundation is still simple. A classic sudoku is a 9×9 grid split into nine 3×3 boxes. Each row, column and box must contain the digits 1 to 9 once each. The first productive techniques are singles. A naked single appears when one cell has only one candidate left. A hidden single appears when a row, column or box has only one possible location for a digit, even if the cell still shows several pencil marks. These are not glamorous, but they are the engine of many solves. If your notation is messy, you miss them and start reaching for advanced names too early.
Candidate notation matters more than some solvers admit. Pencil marks are not decoration. They are a working model of the puzzle. Good notation is compact, consistent and updated immediately after each elimination. If you delay, the grid lies to you. Digital apps often automate this, which can be convenient, but it also hides the reason a move is strong. On paper, you can feel the structure changing. That makes later techniques easier to trust.
Locked candidates and subsets
Once singles slow down, locked candidates are often the first real step up. If all possible locations for a digit in a 3×3 box lie in one row, that digit can be removed from the rest of the row outside the box. The same works with columns. This technique is satisfying because it links local and global structure. You are not just tidying one box. You are pushing information outward.
Pairs and triples do something similar with sets. A naked pair means two cells in a unit contain the same two candidates and no others, so those digits can be removed from the remaining cells in that row, column or box. Hidden pairs reverse the emphasis: if two digits can only go in two cells, other candidates can be stripped from those cells. In practice, hidden pairs are easy to miss because they ask you to scan digit by digit rather than cell by cell. That slower view is sometimes the one that breaks a stubborn grid open.
The trade-off with subset techniques is time. They are reliable, but visual search can become expensive. In an easy newspaper sudoku, careful singles and the occasional locked candidate may be enough. In harder grids from specialist collections, subsets often become routine. The puzzle is not cheating. It is simply asking you to manage information more deliberately.
When pattern techniques earn their keep
X-Wing is one of the first named advanced patterns many solvers learn. If a digit appears in exactly two positions in each of two rows, and those positions line up in the same two columns, that digit can be removed from other cells in those columns. The same idea works with rows and columns swapped. On paper, this can feel slightly theatrical the first time you spot it. The important point is not the name. It is that the grid has created a stable rectangle that limits where the digit can live.
Simple colouring and chain-based methods push the same logic further. You assume alternating truth values across linked candidates and use contradiction or convergence to make eliminations. This is powerful, but it also introduces cognitive overhead. For many solvers, colouring is worth learning because it remains readable. Very long forcing chains, by contrast, can be technically valid while being unpleasant to execute without software support.
That is a good place to be honest about solving goals. If you want a calm daily habit, you do not need every advanced pattern in the encyclopaedia. If you enjoy extreme hand-set puzzles or the more demanding collections associated with specialist sudoku magazines and dedicated constructors, then learning fish patterns, colouring and chain logic becomes more relevant. The right toolkit depends on what sort of friction you actually enjoy.
A practical solving order
A sensible routine begins with scanning for hidden singles after every placed digit. Then check for naked singles, locked candidates and obvious pairs. Only after those passes dry up is it worth hunting for larger patterns such as X-Wing. This order saves time because basic moves often expose the same eliminations more cheaply. There is little glory in finding an advanced pattern that a missed hidden single would have made unnecessary.
Also, avoid the trap of trial and error disguised as bravery. Some hard sudokus do invite bifurcation methods, where you test an assumption and follow consequences. That can be a legitimate solving style. But in standard editorial puzzles, extensive guessing usually signals either a missed logical step or a puzzle whose intended route is less elegant than advertised.
The best habit, then, is modest but effective: keep candidates clean, use a stable solving order, and treat technique names as labels for relationships already present in the grid. Sudoku becomes much less mysterious once you stop asking for inspiration and start asking what the structure already forbids.