How puzzles and answers are checked

A puzzle site is only worth your break if the answer at the bottom is right. This page sets out what Think Fanny actually does before a teaser, quiz, or walkthrough goes live, what it refuses to do, and how to tell us when something is wrong.

Nothing ships on a single solve

Writing a puzzle and solving a puzzle are different jobs, and the person who just wrote one is the worst judge of whether it works. So every teaser goes through a second, separate pass before it is published. The second pass starts cold: only the prompt as a reader will see it, no notes, no working from the first attempt.

• The prompt is re-solved from scratch against the published wording, not the draft wording.
• If the second pass lands on a different answer, the page does not ship. Either the wording is ambiguous or the answer is wrong, and both are our problem, not yours.
• The explanation has to reconstruct the answer on its own. If you cannot get from the explanation back to the answer without extra assumptions, the explanation gets rewritten.
• Quiz keys are checked against the rendered page, not the source file, because an off-by-one index is the easiest way to publish a quiz that marks the right answer wrong.

Pages get re-read on a rhythm too. Whenever a section gains new content, the rest of that section is re-read alongside it, and the full library gets a pass at least twice a year. The updated date on a page moves only when something on the page actually changed. It is not bumped to look fresh.

Where the facts behind trivia come from

Riddles are self-contained: the puzzle carries its own proof. Trivia is not. A trivia question asserts something about the world, so it needs a source that is not us. Questions are checked against independent reference works, and quiz pages name the ones used for that set.

• Film and television questions: the AFI Catalog and the British Film Institute for titles, release years, and credits.
• General knowledge: Encyclopaedia Britannica.
• Words and definitions: Merriam-Webster and the Oxford English Dictionary.
• Maths and number puzzles: Wolfram MathWorld and the On-Line Encyclopedia of Integer Sequences.
• Classic puzzle texts: public-domain collections on Project Gutenberg.
• Mobile game facts such as price, platform, and what a mode is called: the developer's own site and the store listing, which stay the authority because they change faster than a page here does.

Crowd-edited encyclopaedias are fine for finding a fact and useless for confirming one. They get used to locate the claim, never as the source that settles it.

Puzzles we wrote, and classics we retell

Two kinds of puzzle sit in the library, and they are not the same thing.

Most of the logic set is classic: folk problems that have circulated in puzzle books, maths lectures, and pub arguments for decades, some for centuries. The river crossing with the fox, the goose, and the grain is roughly twelve hundred years old. Nobody owns it. What can be owned is the wording, so classic setups are rewritten here from the underlying problem rather than lifted from whoever ranked first last week, and the explanations are worked out rather than paraphrased. If a puzzle feels familiar, that is the point. A classic survives because it keeps working.

The rest are ours: variations, number traps, and quiz sets written for this site. Neither kind gets dressed up as the other. We do not claim to have invented a puzzle older than the printing press, and we do not pad an original with a fake origin story to make it sound weightier.

When the answer is arguable

Some puzzles have exactly one defensible answer. Others have one intended answer and a second reading that also survives the wording, which is where almost every argument about riddles actually comes from. The missing-dollar hotel puzzle is the standard example: the numbers are not in dispute at all, the framing is, because the puzzle invites you to add two amounts that were never meant to be added.

Our position on those is fixed:

• If two readings both hold up, the wording is the bug. We tighten the prompt rather than argue that you read it wrong.
• Where an alternative answer is genuinely valid and worth knowing, it goes in the explanation instead of being quietly deleted.
• Where the puzzle is a known trick with a known trap, the explanation names the trap. Getting the answer without understanding why the obvious answer fails is not worth much.
• When a fix changes what the correct answer is, the page says what changed and the updated date moves with it. Silent rewrites are how a site quietly stops being trustworthy.

Who writes this

Everything here is published under one collective byline: the Think Fanny puzzle desk. That is a standing byline for the process described on this page, not a person. There is no invented author profile on this site, no stock headshot, and no biography claiming credentials that would not survive a check.

One byline is also the honest way to describe how the site is put together. Writing, the second pass, fact checks, and corrections all sit with the same desk, so the desk is what gets named and the desk is what answers your email. Authorship is recorded the same way in the page markup: the publisher is the site, not a fictional individual.

How the site pays for itself

Think Fanny is independent and pays its own way. A domain and static hosting are the entire running cost, and they come out of pocket. Nobody has bought a mention here.

• No game maker has paid to appear in a roundup, a walkthrough, or a quiz.
• There are no sponsored posts and no paid guest articles.
• There are no affiliate links in the game roundups as this page is written. If that ever changes, this page changes with it and the pages carrying those links will say so plainly.
• There is no account system, no newsletter wall, and no payment processing. There is nothing to buy, so nothing to sell you.

What this site will not do

• Hide an answer behind an email signup, a survey, a countdown, or five paginated clicks.
• Publish an answer nobody has actually worked through.
• Run "only 2% of people can solve this" bait. The claim is invented and the puzzle is usually easy.
• Copy another site's puzzle wording, explanation, or quiz set.
• Pad a two-paragraph answer into a two-thousand-word page to hold you on the tab longer.
• Invent statistics, review counts, ratings, or player numbers to look bigger than it is.

Tell us an answer is wrong

Corrections are the most useful mail this site gets, and they are read before anything new gets written. If a teaser answer is wrong, a quiz marks a correct choice as incorrect, a prompt is ambiguous, or a walkthrough describes a game feature that no longer exists, send it in.

What happens next: correction mail is triaged ahead of new content, and if you are right the page is fixed and its updated date moves. If we disagree, you get the reasoning rather than silence. We are a small operation without a support queue, so we will not promise you a reply within a set number of hours, but reports that name a specific page and a specific problem are the ones that get handled fastest.

For anything that is not a correction, use the contact page. To see what the site publishes and who it is for, read about Think Fanny.