Fourfold guide
Quality, tags, and reputation
rating measures difficulty. Reactions and tags describe how well a puzzle is made and what kind of challenge it offers.
React to each quartet
After a first play, eligible signed-in players can choose Loved it for a quartet that worked especially well or mark one that Didn't land. Skipping a quartet leaves it neutral. Players can also say whether they enjoyed the puzzle overall and select what stood out—for example, a great theme, fair misdirection, an ambiguous connection, or untagged specialist knowledge.
These reactions remain editable for 24 hours. Reports and the optional private note for the Fourfold team are handled separately from quality.
How community quality appears
Quartet reactions and the overall response combine into one community approval score. The words come first: Needs work, Mixed, Well liked, Standout, or Exceptional. New puzzles do not show a precise percentage until enough people have responded.
- Fewer than 10 responses: New
- 10–29 responses: Emerging, with an early score
- 30 or more responses: Established
Quality methodology
For answered quartets, q is the share that received Loved it. The overall response o is 1 for Liked and 0 for Disliked. When both are present, the ballot is 55% quartet reaction, 30% overall reaction, and 15% agreement between the two. If only one is present, that signal stands alone. The community total begins with eight neutral baseline responses at 65%, which keeps a handful of extreme reactions from dominating a new puzzle.
Scores below 50% are Needs work; 50–64% Mixed; 65–79% Well liked; 80–89% Standout; and 90% or above Exceptional. The easier/about-right/harder question and descriptive checkboxes never change this number.
Two views of creator reputation
Community reputation summarizes eligible human reactions across a creator's puzzles. Skilled-player reputation separately reflects established players in the current top fifth who have completed at least 30 rated games. It appears only after at least ten skilled-player responses, so a tiny sample is never presented as a consensus.
Puzzle structure
Creators classify each puzzle as Independent or Interlocking. The label tells players whether every strong quartet should stand on its own or whether the remaining board may be needed to confirm a guess. Community feedback about a misleading label is shown separately from the author's choice. See examples and the fairness promise.
Where tags came from
Fourfold distinguishes tags added by the author, suggested with AI assistance, supported by community responses, and added by the moderation team. Fixed traits such as niche, technical, pop culture, or heavy misdirection are based on yes/no/unsure responses and appear only after a strong community consensus.
Topical tags stay hidden
Topic tags are for specialist knowledge that may genuinely be required, not every subject mentioned by a clue. Because they can reveal too much, they remain hidden until a player asks to see them or finishes the puzzle. Signed-in players may suggest a missing topic; suggestions are combined, cleaned up, and moderated before appearing. “Niche” is an aspect of difficulty rather than a topic.