Hard cat puzzle tips: reason through crowded candidate sets

On harder cat puzzles, look for what several candidates force collectively rather than trying to pick one immediately. Reserve rows or columns, exclude cells touched by every possible placement, and return to single candidates after each change.

Refresh the candidate map first

Before searching for an advanced pattern, apply every direct exclusion from confirmed cats. Check the entire occupied colour region, the full row and column, and diagonal neighbours. One forgotten exclusion can hide a simple placement; one unjustified X can create a false one.

Then count candidates in each unfilled unit. Start with the smallest complete sets because their consequences are easier to compare. A region with only a few surviving cells may constrain a large part of the board even if none of its cells can yet be chosen individually.

Ask what is fixed across alternatives

Suppose a required cat can be at either of two neighbouring cells in one row. You cannot choose its column yet, but its row is certain. Exclude other regions' candidates in that row. Also look for cells that touch both alternatives, since those cells cannot hold another cat.

These conclusions do not require guessing the exact location. The key question is: if each remaining possibility were the true position, which outside cells would be impossible in every case? Exclude that common set and preserve everything that still has a supported possibility.

Combine independent region requirements

If two distinct regions can use only the same two columns, their cats must occupy those columns in some order. Other regions cannot use either column. This can remove candidates far from the pair and create a single in a row that previously looked unconstrained.

The same counting idea extends to a group of regions and an equally sized set of columns or rows. Every region must contribute a cat, and each selected line can hold only one. The restriction must cover every candidate of every chosen region; an outside option breaks the reservation argument.

Do not turn the reservation into a specific matching unless a further rule establishes that matching. Knowing that two columns are occupied is weaker than knowing which cat goes in each, and it is often already enough to make progress.

Check a candidate by contradiction carefully

When direct patterns stall, you can reason about a candidate as a temporary hypothesis without placing a cat. Imagine it were true, apply the resulting exclusions, and see whether an unfilled row, column or region loses every candidate. If that contradiction follows necessarily, the original candidate is impossible.

A contradiction is not merely that a route becomes awkward or that you cannot see the next move. You need a concrete rule failure. If every branch you explore feels difficult but remains consistent, you have not proved the candidate wrong.

Keep hypotheses out of the permanent marks

1. State the candidate being tested as a hypothetical cat.
2. Track its consequences separately from confirmed deductions.
3. Identify a specific impossible unit or rule conflict.
4. Remove only the original candidate if the contradiction is valid.
5. Discard all temporary consequences and rescan the real board.

The distinction matters because an exclusion derived under a false hypothesis is not automatically true on the actual board. Carrying those temporary Xs forward can corrupt later reasoning even after you correctly reject the starting candidate.

Diagnose mistakes before spending another life

If you reach an empty candidate set, revisit recent permanent actions first. Ask whether a locked region really had no outside candidate, whether a pair used distinct required cats, and whether a supposed shared neighbour was forbidden under every alternative.

Wrong cat placements cost lives, so use reasoning rather than a placement to test an idea. When a hint is needed, compare its explanation with your candidate map. It may reveal a missing basic exclusion rather than a new advanced technique. That is useful information about the scan you should improve.

Keep the solve moving after an advanced step

One difficult exclusion may reveal a simple chain. Recheck the affected units immediately instead of continuing to hunt for another elaborate structure. Hard boards still contain ordinary singles once the necessary restrictions have been exposed.

Progress can be an exclusion rather than a placement. If a step is logically sound and reduces the candidate set, it has done useful work. Do not force an exact cat position merely to make the board look more complete.

Where to play

哈基米数独 (Hachimi) offers HarmonyOS cat-region puzzles and explanatory hints. Use a hint to study a deduction, then return to your own complete candidate scan.