Mastery is a belief, not a checkbox

One witness is never proof

A single Evidence Check result is one witness. It might be a lucky guess, a careless slip, or an outsourced answer. So Maestro never records mastery from one result — it maintains, for every knowledge component, a mastery belief: an evolving probability in the spirit of Bayesian knowledge tracing, not a checkbox, that climbs and dips with evidence.

A clean check moves the belief up; it never jumps it to “mastered.” Coherent evidence — practice and check agreeing, confidence calibrated to correctness, reasoning quality matching answer quality — moves it strongly. Incoherent evidence moves it weakly, and is held for confirmation.

The four mechanisms of triangulation

• Prediction versus result. The learner's engagement in the node — practice performance, time, help requests — predicts an expected outcome. A result far above what their own history predicts is held as unconfirmed, pending more evidence, not believed at full strength.
• Gradual belief. Belief moves in steps sized by coherence. There is no single moment where one answer flips “unknown” to “mastered” — which is precisely what makes gaming one moment pointless.
• Survival of re-encounter. Knowledge components resurface in later nodes and as spaced re-checks — a different variant over the same tested substance. Real mastery survives; outsourced mastery collapses, and the collapse retroactively discredits the earlier result.
• The milestone anchor. The controlled-conditions milestone is the strongest witness. Checks that said “secure” followed by a milestone that disagrees is the loudest incoherence — and since the grade attaches to the verified belief, the faked checks bought nothing anyway.

Self-correcting, and kind about it

Because the same concept recurs across many nodes, the picture self-corrects over time: a single wrong reading is recoverable, and mastery is only believed when the evidence agrees with itself across encounters. A low belief is treated as a temporary state to support — never a label to pin on someone.

For a learner who wanted to cheat, the consequence is stark: they do not beat the system once, they inherit the obligation to fake coherently, forever — timing, confidence, reasoning quality, every re-encounter, and a controlled-conditions exam. The only reliable way to fake mastery is to have it.