Puzzle and mystery caches

Hard is fine. Ambiguous is not.

The most useful distinction in puzzle design has nothing to do with how clever the mechanism is.

A hard puzzle asks something demanding and has one answer. You may not get it. If you do, you know you have. The numbers land, the coordinates are plausible, everything clicks.

An ambiguous puzzle asks something that admits several readings. You reach an answer, you cannot tell whether it is the answer, and confirming it means a drive. When you turn out to be right it feels like luck rather than accomplishment.

Solvers will happily spend hours on the first kind and abandon the second. Difficulty is not the complaint; ambiguity is.

The checker changes the economics

A solution checker, anything that tells the solver whether their coordinates are right before they travel, transforms what a puzzle can reasonably ask.

Without one, the cost of being wrong is an afternoon and a drive. Solvers respond rationally: they only attempt puzzles they are confident about, and they abandon anything where they cannot verify their reasoning.

With one, being wrong costs a click. That single change makes people willing to attempt harder puzzles, iterate on partial answers, and finish things they would otherwise have left.

If you build puzzles and want them solved, this is the highest-value thing you can add.

What experienced solvers actually do

Watching good solvers is instructive, because their first move is not to start decoding.

They look for what is deliberate. A listing is written by somebody who chose every element. An odd phrase, an image with no obvious purpose, a strange line break, a title that does not quite fit — these are the puzzle announcing itself.

They check the obvious encodings early, because a surprising share of puzzles use them.

They count things. Coordinates need a specific number of digits, and finding a set of values that is exactly the right length is usually confirmation you are on the right track.

They sanity-check the result. Coordinates that land in the sea, in a different county, or in the middle of a motorway are wrong regardless of how satisfying the derivation was.

Advice for owners, in one line each

Test it on somebody else. You cannot judge your own puzzle; you know the answer. Watch a volunteer and note where they stall.

Add a checker. See above.

Make the final container easy. A hard puzzle followed by a hard hide is two walls with no reward between them. Once someone has solved it, let them find it.

Say what kind of puzzle it is, at least roughly. Solvers do not want the answer; they want to know whether this is a cipher, a maths problem or an image puzzle so they can decide whether to start.

Keep it solvable over time. Puzzles that depend on an external website, a current event or a specific piece of software age badly. The ones still being solved in ten years depend only on what is in the listing.

Age it deliberately

The puzzles still being solved in ten years have one thing in common: everything needed to solve them is inside the listing.

Puzzles that depend on an external website, a piece of software, a current event or a service that might close have a shelf life, and the owner rarely notices it expiring. What they notice is the finds slowly stopping.

If a puzzle must reach outside the listing, choose the most durable thing available and expect to revisit it.

The hardest part is not the puzzle

Owners tend to spend their effort on the mechanism and almost none on the two things that decide whether people attempt it at all.

The first is knowing where to start. A listing that is deliberately cryptic about what kind of puzzle this is filters out everybody except the people who already enjoy that specific type. Saying plainly that this is a cipher, or a research task, or requires an image to be examined, costs nothing and does not make the solve easier. It converts a puzzle nobody opens into a puzzle people try.

The second is knowing when you are done. A solver who arrives at a coordinate pair and cannot tell whether it is right faces a drive to find out. That asymmetry (hours of work, then a gamble) is what keeps solve rates low, and it is the problem the coordinate checker exists to remove. A mystery cache is one of the listed types, and the type is set before anybody leaves home.

Puzzles that age badly

A puzzle is a piece of software with dependencies, and the dependencies fail quietly.

Anything that relies on an external website will eventually rely on a website that has changed, moved or gone. Anything that depends on a piece of contemporary popular culture stops being solvable by anyone who was not there. Anything requiring a specific tool assumes that tool still runs on current hardware.

The puzzles still being solved a decade later are almost all built from things that do not move: plain ciphers, the same letter shift the hint field uses is the gentlest of them — arithmetic, physical observation at a location, information printed on something permanent. That is a constraint on the puzzle and not on its difficulty, the hardest puzzles in the hobby are frequently the ones with no external dependency at all.

If your puzzle needs the outside world, check it on the same schedule you check the container. It is the same maintenance obligation, and it is the one owners forget because nothing about it gets wet.