Word calculations
Letter counts and position comparisons operate on normalized English alphabetic input. Edit distance uses the Levenshtein method. Anagrams compare letter inventory, while position games preserve index. These measurements describe spelling structure, not meaning.
Dataset provenance
Word challenges are authored for this site as finite mechanic-specific sets, then checked for length, characters, duplicates, paths, and expected answers. They are not copied from a private publisher schedule and are not presented as a complete English dictionary. Geography entries keep the checked city name, country, region, coordinate, clues, and reveal fact together so a correction can update the full item.
Game validation
Each game uses a finite challenge set with rules stated on its page. Authoring checks confirm expected length, allowed characters, board paths, clue answers, and duplicate entries where those rules apply. A rejected word may still exist outside the current challenge set.
City coordinates and distance
The city game uses a curated coordinate near each city center. Distance uses the Haversine formula and a mean Earth radius of 6,371.0088 km, then rounds for a learning-friendly result. It is a surface-distance estimate, not a driving route, flight plan, border measurement, or surveying result.
Clue review
Geography clues move from orientation to distinction. Stable facts are preferred. Changing facts require a date and source. A reveal must connect the clues to the answer, and another reviewer should be able to follow that route without relying on the writer's private association.
Duel ratings
Every player starts at 1,000. An Elo-style calculation uses both pre-match ratings, the actual win, loss, or draw, and a factor of 24. Only completed server-scored Duel rooms update the public board. The rating describes results inside this game and is not a general language assessment.
Computer difficulty
Computer levels change candidate quality and consistency within the game's checked data. Easy is intentionally less consistent, while higher levels choose stronger legal moves more often. The labels do not claim human equivalence or use hidden personal data.
Content review
Every published article needs a distinct purpose, a matching title and H1, a useful opening answer, clear limitations, accurate internal anchors, and sources for claims that are not stable or directly calculated. Material that fails the originality or sourcing check remains held back from publication.
Corrections and updates
Game data, prose, links, and generated visuals are checked separately. A factual correction updates the affected explanation and its visible updated date. Reports can be sent through the contact page with the URL and evidence.
Use the method on a real page
Apply these standards in the educational word tools where each result states its rule. free word games where every mechanic defines valid play. Wordmap questions and answers for common limits and edge cases.