DESIGN NOTE / 20 AUGUST 2026
Correct is not the same as unaided.
This design note asks how an attempt record can preserve rough latency, open notes, hints, answer reveal, and declared AI help without treating one correct answer as evidence of mastery or transfer.
01 / CONDITIONS
“Correct” collapses the attempt.
A record that stores only “correct” loses the conditions that produced the response. An answer retrieved with notes closed, one composed with notes open, one reached after a hint, and one checked after an answer reveal can end with the same words. They are not the same attempt.
Rough latency adds context, but not a score. The clock includes reading, thinking, and typing. A hint changes the retrieval conditions. A reference answer shown before the visitor commits to an attempt changes the task again. Declared AI help should remain attached to the record rather than disappear behind a correct result.
These fields preserve a claim about conditions. They do not verify attention, honesty, understanding, or independence.
02 / RECORD
A compact attempt record.
This human-readable example keeps the conditions beside the outcome. It is an illustration, not visitor data.
- Attempt
- Active recall
- Result
- Correct
- Notes
- Open
- Hint
- Used
- Reference answer
- Revealed after attempt
- AI help
- Declared
- Elapsed
- 01:42, rough
- Self-assessment
- Partial
- Project transfer
- Not observed
LOCAL INTERACTION / EXPLAINER
Try the attempt evidence model.
This is an explainer, not the private prototype or a learning assessment. It uses React memory only. There is no answer input, saved history, or attempt-state analytics.
Without looking at the reference answer, why is a correct response alone insufficient evidence of an unaided recall attempt?
Mark an answer to freeze the declared conditions in this local trace.
One attempt under declared conditions is not proof of mastery or project transfer.
Prompt ready. Declare the conditions before marking your attempt.
03 / TRANSFER
Recall and project transfer are different evidence.
A recall attempt asks what someone can reconstruct under declared conditions. Project transfer asks whether the idea can be applied in a different, representative task with its own constraints and definition of done.
The first record cannot stand in for the second. A correct recall response may justify another attempt or a separate application task. It does not establish that the person can use the idea in project work.
04 / LIMITATIONS
What this does not show.
Research supports retrieval practice and spacing on average across studied comparisons. That evidence does not validate this record format, AI-help attribution, or the private prototype.
This evidence model and the private prototype have not demonstrated improved retention, transfer, speed, or capability.
- Declarations about notes, hints, and AI help may be incomplete.
- Self-assessment is not an independent measure.
- Rough latency includes reading and typing and is not a speed or capability score.
- One attempt cannot establish mastery.
- Project transfer needs separate evidence from a representative application task.
05 / SOURCES
Research boundary.
These reviews support a careful average claim about retrieval practice and distributed practice in the contexts they studied. They do not test this evidence model.
- Yang, C., Luo, L., Vadillo, M. A., Yu, R., & Shanks, D. R. (2021). “Testing (quizzing) boosts classroom learning: A systematic and meta-analytic review.” Psychological Bulletin, 147(4), 399–435. https://doi.org/10.1037/bul0000309
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). “Distributed practice in verbal recall tasks: A review and quantitative synthesis.” Psychological Bulletin, 132(3), 354–380. https://doi.org/10.1037/0033-2909.132.3.354
TALK TO US
Building something where the conditions of an attempt matter?
This is the discipline Virtuoso is built on. If you are working on learning, assessment or human + AI collaboration and want to compare notes, start a conversation.