Practice Sudoku by Technique

Why drilling one technique works

The usual way to improve at Sudoku is to solve a lot of puzzles and hope. It works, slowly, because exposure is exposure — but it is badly aimed. If you cannot yet see Hidden Pairs, fifty random puzzles might present you with six, spaced weeks apart and unlabelled, so no recognition ever consolidates and you may not even notice the ones you missed.

What actually builds recognition is massed, labelled exposure with immediate feedback: many instances of the same pattern close together in time, each confirmed right or wrong straight away. It is the same reason chess players drill tactical motifs from position sets instead of only playing games. Twenty puzzles that all contain Hidden Pairs will do more for your Hidden Pair recognition than two hundred random puzzles, in a tenth of the time.

That is what these pages are. Every puzzle in a set has been solved by a step-by-step logical solver which recorded exactly which techniques were required, so the exposure is guaranteed rather than hoped for.

Reading the two numbers

Each technique above shows two counts, and the difference between them matters:

• Puzzles that require it — the technique is genuinely needed somewhere in the solve, though something harder may be needed too.
• Puzzles that need nothing harder — the technique is the peak of the puzzle. These are what each practice page serves, so you can work on one pattern without a harder one interrupting.

The gap between the two is why the counts fall off so sharply as you go down the list. Naked Singles are required by essentially every puzzle but are the peak of very few; Swordfish puzzles are rare in both senses, because a grid hard enough to need one usually demands something else along the way. The construction guide explains where that distribution comes from.

How to work a practice set

1. Read the technique page first so you know what the pattern is and why the deduction is valid — not just what it looks like. Shape memory does not survive contact with a real board, where patterns are ragged and asymmetric.
2. Work a walkthrough or two to see it fire on a real grid before and after.
3. Then drill the set, and on each puzzle exhaust the cheaper moves before hunting for the named technique. It will not be available on move one.
4. Stay until it is boring. The point at which the pattern stops feeling like work is the point at which it has become recognition rather than derivation, and that is the whole objective.
5. Then mix it back in via the technique combinations or the difficulty pages, so the recognition has to compete with everything else on the board.

Where to start

If you are unsure, start higher on the list than feels flattering. Most solvers who stall on hard puzzles are missing Omission — locked candidates — rather than anything advanced, and most who stall on medium puzzles are simply not following their single cascades to exhaustion. Fluency low on the ladder is what makes everything above it cheap. The learning-order guide sets out the full sequence and the signals that tell you when to move on.