Brain Teasers
Brain Teasers Lab
A shelf of classic logic and game-theory puzzles, each played as an animated story that makes you commit to an answer before revealing the reasoning.
How it works
The Brain Teasers lab is an index of playable puzzles. Two featured games - Screwy Pirates and Tiger and Sheep - have their own custom experiences with graded answer builders. The other ten run in a shared story engine: pixel-art characters act out the setup beat by beat, the assumptions appear one at a time, and then the game stops on a single question.
At the question screen you type your answer into a free-text box before anything is revealed. Locking in a guess is not graded - the game simply shows you what you wrote next to the real answer, so the honesty is on you. The answer then unfolds as a sequence of labelled steps you click through, followed by a 'Why it works' breakdown covering the underlying principle and where it shows up in the real world.
Every story teaser also has a server-rendered written solution at /brain-teasers/<slug>/solution, cross-linked from the playable page. The numbers in the written version match the game's telling exactly.
How scoring works
The story teasers have no score. Your typed guess is displayed back to you unchanged when the answer appears - there is no grader, no points, and no tiers. The two featured games each have their own answer builder with deterministic checks, described in their own guide entries.
Access works differently here than in other labs: the eleven story-format teasers share a single lifetime unlock for free users. Opening any one of them uses the whole group's allowance, so pick the teaser you most want to see solved. Screwy Pirates is linked directly from the lab page without that gate.
Commit before you reveal
The entire training value of this lab is the gap between your locked-in guess and the real answer. If you skip the guess, you get a nicely animated explanation and learn almost nothing about your own instincts. Write down a specific answer - a number, a named strategy, a yes or no - not a hedge.
When your guess is wrong, the answer steps are structured so you can find the exact step where your reasoning diverged. That step is the thing to remember, not the final answer.
Read the assumptions as constraints, not flavour
Each teaser states its assumptions explicitly before the question - rationality, no communication, common knowledge, and so on. These are not decoration: nearly every teaser's answer flips if one assumption changes, and the breakdown sections routinely point at which one. In an interview, restating the assumptions before answering is what buys you time and credibility.
The tags on the lab page (Game theory, Probability, Induction, Case reduction, Conditional probability) tell you which tool to reach for before the story even starts. Use them.
A worked example
Open The Third Door. The story plays out the Monty Hall setup: three doors, one car, a host who knows where the car is. The assumptions appear: the host always opens a non-chosen door, always reveals a goat, always offers the switch. The question asks whether you switch and what your odds are.
Type '50/50, no point switching' - the most common wrong answer - and lock it in. The answer headline says switching wins two-thirds of the time. Stepping through, the divergence point is Step 1: your first pick was 1/3 from the start, and nothing the host does changes that. That single step is the lesson the whole game exists to deliver.
Common mistakes
• Skipping the guess box and going straight to the answer - you lose the only feedback signal the game offers.
• Burning the free-tier group unlock on a teaser you already know cold, when the whole group locks after one reveal.
• Watching the answer steps without clicking through them one at a time - the pacing exists so each inference lands before the next one builds on it.
• Ignoring the 'Why it works' breakdown, which is where the transferable interview material lives.
Why interviews test this
These are the puzzles interviewers actually ask, in close to their canonical forms: Monty Hall, the birthday problem, hat parity, bridge crossing, the prisoner's dilemma family. The lab's format - assumptions first, commitment second, reasoning last - is also the format of a good interview answer.