Eight of the twelve techniques taught on this site — Naked and Hidden Pairs, Triples, and Quads — are a single mathematical idea wearing eight costumes. If you understand the idea once, properly, you stop memorizing rules and start recognizing the one pattern behind all of them. The idea is the pigeonhole principle, and the proof takes a paragraph.
The setup: units are bijections
Consider any single unit — one row, column, or box. When the puzzle is finished, its nine cells will contain the nine digits exactly once each. In the language of mathematics, the finished unit is a bijection, a perfect one-to-one matching between cells and digits: every cell gets exactly one digit, every digit gets exactly one cell.
Every subset technique is a statement about what partial information forces in a one-to-one matching. Here is the master lemma:
If k digits have all their possible homes within the same k cells of a unit, then those k cells contain exactly those k digits — and nothing else.
The proof is pigeonhole counting. The k digits must all be placed somewhere in the unit, and by assumption their only available cells are those k. No two digits can share a cell, so k digits filling at most k cells must fill exactly those k cells, one each. There is no room left over — so no other digit can occupy any of the k cells, and none of the k digits can appear anywhere else in the unit.
Hidden subsets: the lemma read forward
Read the lemma as stated and you get the hidden family. Suppose in some row the digits 4 and 9 each have only two possible cells — and they are the same two cells. That is k = 2 digits confined to 2 cells: a Hidden Pair. The lemma says those cells contain 4 and 9 and nothing else, so every other candidate written in those two cells is false and can be erased. The pattern is “hidden” because the extra candidates camouflage it — the two cells might each show five candidates, and only per-digit bookkeeping reveals that 4 and 9 have nowhere else to go.
Hidden Triples and Hidden Quads are the same statement with k = 3 and k = 4, with one subtlety: the three digits need not each be possible in all three cells. Digit A might fit in cells 1 and 2, digit B in cells 2 and 3, digit C in cells 1 and 3. Collectively they are still three digits confined to three cells, and the lemma fires.
Naked subsets: the lemma read in reverse
Now flip the roles of cells and digits. Suppose k cells of a unit collectively show only k distinct candidates — two cells both marked {3, 7}, say: a Naked Pair. Those two cells must receive two different digits, and their only options are 3 and 7 — so between them they consume both. Every other cell in the unit can drop 3 and 7 from its candidates. The same counting gives Naked Triples and Quads, again with the union subtlety: three cells marked {1,2}, {2,6}, {1,6} form a perfectly good triple even though no cell shows all three digits.
Notice the symmetry. Hidden subsets say: k digits locked into k cells → clear the other candidates from those cells. Naked subsets say: k cells locked onto k digits → clear those digits from the other cells. It is one lemma applied to the two sides of the bijection.
The duality theorem: every naked subset has a hidden shadow
The symmetry runs deeper than analogy. Consider a unit with U unsolved cells, and suppose it contains a naked subset of size k. The U − k remaining cells must contain the U − k remaining digits — which means those U − k digits are confined to those U − k cells: a hidden subset of size U − k, in the same unit, yielding exactly the same eliminations.
Concretely: in a row with six empty cells, a Naked Pair coexists with a Hidden Quad on the other four cells, and either one, applied, produces the identical result. This has a practical consequence: you never need subsets larger than four. In the worst case — a unit with nine empty cells — any subset of size five or more has a complementary subset of size four or less. It also explains a solving heuristic: naked subsets are easier to see (the candidates are right there in the cells), hidden subsets are easier when the unit is crowded, and which of the pair you find first is a matter of viewpoint, not substance.
Why subset moves are “safe”
Subset eliminations are locally provable: the justification lives entirely inside one unit, requires no assumption about the rest of the grid, and does not depend on the puzzle having a unique solution. Contrast this with uniqueness-based techniques, which borrow the setter's promise as a premise — a genuinely different kind of reasoning. If you apply a correctly-identified Naked Triple, the eliminations are true in every solution the grid could possibly have. When a deduction goes wrong in practice, the cause is almost always corrupt pencil marks rather than faulty logic — which is why the maintenance discipline matters so much to subset play.
One idea, one skill
Seen this way, the subset ladder is not eight techniques but one skill exercised at increasing perceptual difficulty: find k things confined to k places. Pairs are easy because k = 2 patterns are visually loud. Triples are harder because the union condition lets each cell show an incomplete fragment of the pattern. Hidden variants are harder than naked ones because the pattern is obscured by extra candidates rather than displayed. Nothing new is being proved as you climb — your eyes are just being asked to do more.
That is also the argument for drilling each rung separately: the logic transfers instantly, but the pattern recognition does not. Generate a puzzle that specifically requires a Hidden Triple and your eyes learn what the camouflage looks like on a real board — the walkthroughs show worked examples, and the technique picker will hand you as many fresh ones as you can solve. And when you are ready for the same pigeonhole idea stretched across multiple rows and columns at once, that is exactly what the fish family is.