How We Build and Test WordSQD Puzzles

A source word guarantees a route

The round generator starts with a dictionary word long enough to supply three positions. It chooses three distinct positions, sorts them, and displays the letters found there. The source word therefore proves that at least one ordered answer exists.

Players do not have to find that source word. Any dictionary word that preserves the displayed order can pass.

Seeds make boards repeatable

A seed initializes the round generator. The same seed, difficulty, dictionary version, and round count produce the same sequence. Daily boards use the date in their seed; friend challenges store their seed so both players face the same rounds.

The server rebuilds the game

When an official result arrives, the server rebuilds the expected rounds and validates each submitted word. It recalculates score, speed bonus, combos, and longest word. A client cannot create a leaderboard score by posting a total alone.

Tests protect the rules

Automated tests cover letter selection, ordered matching, duplicate letters, normalization, dictionary membership, scoring, skips, Forbidden letters, deterministic Daily dates, tie breaks, and server verification. Live checks then confirm the production route, database, and browser agree.

A rule change needs both a test and a clear public explanation. Otherwise players can face one rule in the interface and another in the score ledger.

Why archive data has limits

The original Daily score table stores date, account, difficulty, seed, score, best combo, longest-word length, and completion time. It does not store each Daily submission. The first archive pages can therefore publish score statistics and deterministic letter sequences, but not highest-scoring answers or notable word lists.

Future archive detail should come from new server-verified records, not from browser claims or retrospective guesses. That boundary matters because an attractive statistic is not useful if WordSQD cannot show where it came from.