About HatDojo
Ten people stand in a line wearing red or blue hats. Each can see every hat in front of them but not their own, hears every word said behind them, and says one colour, back of the line first. The group agrees a plan before the hats go on. How many can the plan always save?
The teen builds the plan from two rules: what the back person says, and what everyone else says. Then they test it. "Try one line" draws ten people and shows who was saved. "Run 10,000 lines" runs the plan on random lines, the same lines every time. "Check all 1,024 lines" runs it on every possible line and shows the worst case.
For a parent or teacher
Ask for the worst case, not the average. A plan that saves five on average can save nobody on a bad line. Under the plan there is a folded card with what to do, one question to ask, and a smaller step if they are stuck; the answer is in its own fold. The best plan treats the back person's word as a parity bit, the same idea used to catch errors in stored and sent data.
How it behaves
No score, no timer and no streak. Nothing is stored or sent. The parity plan's result is checked against our puzzle oracle: on all 1,024 lines it never saves fewer than nine.
What it does not claim
It is practice in designing a strategy and testing it on every case. We claim a good problem to think through, not a measured gain. A plan you type in your own words, the three-person guess-or-pass puzzle and the hundred boxes are planned. HatPals (ages 6–8) is the younger version of the same idea.