How adaptive difficulty actually works

Most exercise libraries are static: a tutor assigns worksheet #4, the student does worksheet #4, and difficulty only changes when the tutor manually picks worksheet #5. That's a lot of manual sequencing for something that should follow from how the student is actually doing.

In Billy, every notion — a single skill or concept — has its own adaptive engine. As a student answers, the engine adjusts the difficulty of what comes next based on their actual performance on that notion, not a fixed script. Two students assigned the same notion can end up on completely different exercises within a few attempts, because they're not solving the same problem at the same pace.

The tutor's job shrinks to what only a tutor can do: deciding which notion to assign, and when. Sequencing and difficulty inside that notion are the platform's job — which is also why the dashboard reports mastery per notion rather than a single overall score. A student who's strong on one notion and shaky on another needs to be treated as exactly that, not averaged into a number that hides both facts.