Hold the numbers. The numbers do not hold still.
Every run starts at level 1 on the same schedule. Difficulty is a simulated number, not a hand-set one — see .
A few named slots hold numbers. You see them once. Then instructions rewrite them, one at a time, and you must keep up in your head. At the end you are asked about the result — with nothing on screen.
A wrong answer — or running out of time — costs one strike, however many questions you missed in that round. Three strikes end the run. Failing does not send you backwards: the same level runs again with freshly generated instructions, because replaying the identical round would just test whether you memorised it.
Level 1 is a single instruction over three slots. It stays that way for level 2, goes to two instructions for levels 3 and 4, and reaches three at level 5. New vocabulary arrives one idea at a time — swaps, then largest and smallest, then conditions — and the first time you meet anything new, it appears in the simplest possible sentence.
8 12 3 becomes 4 9 3.Your score is the highest level you passed. Ties break on strikes remaining, then on speed. Answering inside par scores full speed; past par it decays, and well past it the question closes.
No level here was authored. Each one is generated, simulated, and kept only if its predicted difficulty lands within 10% of the target for that level.
You are modelled as a store plus a processor. The store holds items: slot values, half-finished arithmetic, the question you are currently holding. Mistakes happen only at accesses — the moments you reach into the store or compute something. Each access carries a failure probability that rises with exactly three things:
Nothing else counts. Every instruction and every question compiles to a
trace of primitive events, the trace is priced, and the round's total is a
single number: strain. Your predicted chance of getting the round
right is e−strain.
Because it is falsifiable. The model claims a scan late in a round costs more than the same scan early, that a global operation is secretly rehearsal, and that rotating the slots makes every value read as stale. Those are predictions, and the game records enough to check them. If the numbers disagree with players, the numbers are wrong.
Nothing here changes reading time, processing time, gaps, or par. Difficulty is not configurable inside a scored run.
Practice mode is unscored and lets you pick a level and slow the tempo.
Unscored. Strikes still end the session, but nothing is recorded.