1. Record the surviving evidence
We begin with the object itself or the material already in the archive: photographs, packaging, dimensions, piece count, old filenames, product codes and the wording used in catalogues.
Small clues matter. A filename may contain an older product name, while a distinctive notch pattern can show that two visually similar burr puzzles are not the same design.
2. Check live sources first
Manufacturer pages, official manuals and verified solution videos are preferred because they provide the clearest provenance. Alternative names are checked separately because one puzzle may have been rebranded several times.
3. Use the Wayback Machine as historical evidence
The Wayback Machine can reveal former product pages, catalogue descriptions and links to instruction files that have disappeared from the live web. A capture proves that a page existed at a particular address and date, but it does not make the archived text or artwork free to republish.
Crafty Puzzles uses archived pages to identify and verify puzzles, then writes new summaries or links to the original source where appropriate.
4. Compare the mechanism, not only the shape
Two puzzles may both look like a cube, ball or chest while using different internal constructions. We compare piece count, grooves, slots, orientation and the order of movement before treating a source as a match.
5. Publish the confidence level
Every recovered page should state whether it is supported by an original Crafty Puzzles sheet, an official manufacturer source, a strong multi-source match or only a provisional visual identification. Uncertainty is useful information and should never be hidden.
6. Cross-check solution libraries and historical retailers
Large puzzle catalogues can establish that a solution exists even when the final instruction file is delivered through an email or temporary link. Crafty Puzzles uses those catalogues to recover product names, makers, dimensions, piece counts and likely aliases, then verifies the mechanism against the surviving puzzle before treating a solution as a match.
Useful research sources include manufacturer libraries such as SiamMandalay, retailer catalogues such as Puzzle Master, specialist collector sites and historic product reviews. A matching name alone is never enough: the Hercules and Neutron records in this archive demonstrate why size and component count must also agree.
7. Use Pinterest and saved images as leads, not automatic permission
Old pins can preserve product titles, thumbnails, destination URLs and filenames long after the original page disappears. They are valuable clues for finding a lost Wayback capture or identifying an alternative product name. A pin does not itself grant permission to republish the underlying photograph or diagram, so third-party images are used as research evidence unless reuse rights are clear.
8. Reconstruct the explanation instead of copying the artwork
When an outside solution verifies the geometry, Crafty Puzzles can write a new explanation and, where the piece shapes are sufficiently known, create a new diagram or 3D reconstruction. The reconstruction must be based on verified component geometry and should be labelled as a reconstruction rather than presented as an original product photograph.