Swordfish and jellyfish, when they actually help

Anyone who's read about Sudoku techniques past the X-wing has run into the larger members of the fish family — swordfish across three rows or columns, jellyfish across four. The naming is whimsical, the underlying logic is a clean generalisation of the X-wing, and most of the technique guides that mention them describe them with the same enthusiasm they reserve for any other named pattern. The implicit framing is that these are simply more advanced X-wings, useful in the same way and worth practising for the same reasons.

That framing isn't quite right. Swordfish and jellyfish are real techniques that work when their conditions are met. They're also rare, eye-expensive to spot, and almost never the highest-leverage move available on a stuck puzzle. The honest version of where they fit in a solver's toolkit is more modest than the technique guides imply.

What they are, briefly

A swordfish is the X-wing's three-row generalisation. A digit has at most three candidate cells in each of three rows, and those candidates fall in only three specific columns across all three rows. The conclusion: the digit must occupy three of those nine cells (one per row, in three different columns), and therefore can't appear in any other cell of those three columns. The eliminations run vertically, the same way they do for an X-wing, just across three columns instead of two.

A jellyfish extends this to four rows and four columns. The mechanism is identical — the digit's candidate cells across four rows fall in four columns, the digit must occupy four of the sixteen cells, eliminations happen in the four affected columns. Both techniques have row-based and column-based versions; same shape rotated.

The X-wing piece is the right read for the perceptual habits that surface fish techniques generally — paired-unit scanning, digit-by-digit pencil-marking. Once those habits are in place, swordfish and jellyfish become spottable in principle. In practice is where the difficulty starts.

Why they appear less often than people expect

Fish techniques require a specific configuration that gets exponentially less likely as the fish gets bigger. An X-wing needs two rows where a digit has exactly two candidates, with those candidates lined up in two specific columns. A swordfish needs three rows where a digit has at most three candidates, with all candidates falling in three specific columns. A jellyfish needs four rows where a digit has at most four candidates, with all candidates in four specific columns.

The number of grids that contain a swordfish is much smaller than the number that contain an X-wing, and the number that contain a jellyfish is smaller still. In ordinary expert and master tier Sudoku, swordfish appear in maybe one in fifteen puzzles. Jellyfish appear in something closer to one in fifty. These aren't precise figures — they vary by puzzle source — but the order of magnitude is the point. Most stuck puzzles don't have a fish bigger than an X-wing in them.

What stuck puzzles do have, much more often, is a Y-wing, an XY-chain, a pointing pair you missed, or a pencil-marking error in the back half of the grid. The probability that the move you're missing is a swordfish is small enough that scanning for one first is usually a misallocation of attention.

The eye-time cost

X-wings are already eye-expensive — the wide-grid scanning pattern they require is one of the things that makes them famously hard to spot. Swordfish and jellyfish are worse on this dimension by the same factor the technique is bigger.

A swordfish scan asks you to consider three rows simultaneously, comparing their candidate columns across all three. The combinatorial space is meaningfully larger than the X-wing's two-row scan; the pattern matching takes longer and produces more false starts. Jellyfish are worse again — four rows compared at once is at the edge of what most solvers can hold in working memory without dropping back to writing things down.

If you're spending three or four minutes of a stalled solve scanning for a jellyfish, you're spending those minutes on the technique that's least likely to actually be in the puzzle. That's the trade-off the technique guides usually don't surface.

When they actually earn the time

Three specific cases where reaching for the bigger fish is the right call.

You've already eliminated everything else. If a puzzle has stalled, you've checked for naked and hidden pairs, scanned for X-wings, looked for Y-wing configurations, and the grid is genuinely silent on every easier technique, then a swordfish scan is the next reasonable step. The technique is more likely than not to find something at that point, simply because the easier patterns have already been ruled out.

The puzzle source is known to favour fish techniques. Some constructors deliberately design puzzles where fish techniques are the load-bearing move. If you regularly solve from a source that favours fish patterns, the base-rate shifts and scanning for them earlier in the solve becomes rational.

You're competing or training for speed. At competitive solving speeds, the few puzzles that genuinely require a swordfish are the ones that distinguish very fast solvers from merely fast ones. We've covered the broader question in the stopwatch problem; the fish-specific version is that competitive solvers learn to recognise these patterns reflexively, even though the per-puzzle utility is low, because the rare puzzle that needs one is unsolvable without the technique in your toolkit.

For the rest of us, untimed, on a typical hard or expert grid, the bigger fish are usually a curiosity rather than the next step.

What to do if you don't see one

The honest move on a stuck hard or expert puzzle is usually to scan for things in this order: missed pencil-marks, then naked and hidden pairs you might have moved past, then pointing pairs, then X-wings on the slowly-resolving digits, then Y-wings on bivalent cells. Swordfish and jellyfish come after all of those, not before. If the puzzle is still genuinely silent after that scan, the bigger fish becomes a reasonable next thing to look for. Often the puzzle yields to one of the earlier passes, and the fish were never in the puzzle in the first place.

That sequencing isn't elegant or exotic, but it matches what's actually in the puzzles solvers get stuck on. The exotic techniques get more attention than they deserve precisely because they sound exotic; the boring techniques keep doing the work.