Our goal is to get the learner to true mastery — being accurate, being fast, and with self-assessed “high” confidence.

There is a “cost” associated with the confidence selection which rewards learners by moving them up a mastery level when they are confident and proficient, or dropping them down if they are confident and not proficient.

By proficiency, we are referring not just to accuracy, but also whether the learner responded within a certain duration and other signals.
Square with four quadrants. Along the left axis is "Knowledge" from low (bottom) to high (top). Along the bottom axis is "Confidence" from low (left) to high (right). The top-left quadrant reads "Doubt (Hesitation)". The top-right quadrant reads "Mastery (Smart Action)". The bottom-right quadrant reads "Misinformed (Mistakes)". The bottom-left quadrant reads "Uninformed (Indecision)".


This is how a learner’s mastery level (ML) is affected by the combination of their proficiency and confidence:

Diagram showing "proficiency" and "confidence". When proficiency is sufficient and confidence is low, the learner's mastery level does not change. When proficiency is sufficient and confidence is high, the learner's mastery level increases by one. When proficiency is insufficient and confidence is low, the learner's mastery level drops by one. When proficiency is insufficient and confidence is high, the learner's mastery level drops to zero.