Bolt Series · Human Performance Calibration · Article 18
There is a kind of confidence that says, “I do not know yet.”
We often assume that strong learners should sound certain.
They should know.
They should answer quickly.
They should not hesitate.
But hesitation can contain information.
A learner who says, “I can solve the routine version, but I am not yet sure I can handle a new form of this problem,” may understand their current capability more accurately than someone who says, “Yes, I know it,” after recognising the chapter heading.
Knowing a limit can be a high-resolution achievement.
Metacognition is partly about monitoring your own certainty
Research on metacognition examines how people judge their own knowledge, decisions and performance.
Modern work distinguishes several related ideas: overall confidence bias, the ability to discriminate between one’s own correct and incorrect responses, and how efficiently confidence tracks performance.
That distinction matters because a person can have high average confidence while being poor at identifying which individual answers are actually secure.
The useful skill is not simply feeling confident.
It is being able to tell when confidence is warranted and when uncertainty is real.
Uncertainty can guide exploration
Research with young children offers a particularly interesting clue.
Preschoolers can be overconfident when explicitly rating what they know, yet their exploration choices can still show sensitivity to uncertainty.
In one study, children preferentially explored situations where information was missing, and their explicit confidence judgements improved after they encountered disconfirming evidence.
That suggests something hopeful.
The machinery for noticing uncertainty can develop through contact with evidence.
Not knowing can become useful when it tells you where to look next.
A student who knows their limit can choose better
Imagine two students.
Student A says:
“I understand everything.”
But independent testing shows large gaps.
Student B says:
“I can do the direct questions. I still need help when the problem hides the method.”
Student B sounds less confident.
But Student B may be in the stronger learning position.
They know what to practise.
They know what kind of help to ask for.
They know which future performance will test whether the weakness has changed.
Limits should be specific, not identities
There is a dangerous difference between:
“I cannot do Mathematics.”
and:
“I cannot yet reliably select the correct method when simultaneous-equation questions are presented indirectly.”
The first sentence collapses an entire person into a global judgement.
The second defines a current boundary that can be tested and changed.
Good calibration makes limits narrower and more useful.
Why experts may be better at describing what they cannot yet do
Expertise often brings more resolution.
A beginner says:
“My sprint is bad.”
An experienced sprinter may say:
“My maximum velocity is strong, but my first 30 metres are not yet where they need to be.”
The second athlete is not thinking less of themselves.
They are decomposing the performance more accurately.
Usain Bolt’s public discussions of his start and acceleration are useful for exactly this reason: exceptional overall performance did not require pretending that every component was already equally strong.
Confidence can interfere with learning if it shuts feedback out
Recent research on probabilistic learning found that higher confidence was associated with reduced processing of feedback and less updating after feedback arrived.
That makes intuitive sense.
If I am completely certain I already know, contradictory evidence may receive less attention.
Confidence can therefore regulate whether I continue searching.
Well-calibrated uncertainty keeps the channel open when more information is genuinely needed.
But self-assessment is not enough by itself
There is an important safeguard.
A 2022 scoping review of self-assessment research warned that improving self-assessment accuracy should not be confused with improving actual competence, and that self-reported gains are poor substitutes for objective evidence of learning.
That means Bolt should never become a programme in endlessly thinking about yourself.
Self-knowledge is useful because it helps choose better action.
The action still has to develop capability.
Knowing the limit is valuable only if it helps you train beyond the limit.
A practical classroom language
Instead of asking only:
“Do you understand?”
ask:
- Which part can you do without help?
- Which part can you do only when the method is obvious?
- Where would you want a hint?
- Which answer are you least certain about?
- What could I change in the question that might break your current method?
- What would convince you that this skill is secure?
Those questions do not lower standards.
They increase resolution.
Why this matters for education
A learner who accurately knows their current limits can make better decisions about practice, difficulty, help and readiness.
They can say:
“I cannot do this yet — and I know exactly what ‘this’ means.”
That sentence is not defeat.
It is a map.
And a map of the current boundary is one of the best starting points for moving it.
eduKate Sengkang 