A wrong answer is not enough data
Most systems overreact to isolated mistakes. But one incorrect answer does not always indicate a true weak point. Learners can guess, misread, rush, or simply lose focus momentarily.
Real weakness emerges when patterns repeat across related questions, concepts, and review intervals. That is where meaningful diagnosis begins.
Why topic-level patterns matter
A learner can score reasonably overall while still carrying a few fragile topics that repeatedly break under pressure. Those local weak zones matter because they create hidden volatility.
By tracking answers at the topic level, the system can spot where confidence looks higher than actual retention.
Prediction only matters if it changes the next step
Analytics becomes useful only when it tells the learner what to do next. A warning without an action path is just noise.
That is why weak-point detection should connect directly to review timing, practice selection, and explanation depth.