Small Group Tutorials

Here to help students catch up, keep up, and move ahead. Book a consultation here.

Chat on WhatsApp

Bolt 05 — Predict Before You Perform

Bolt Series · Human Performance Calibration · Article 05

What did you think would happen before it happened?

A student finishes a test.

A week later, the teacher returns it.

73%.

Now everybody knows something about the performance.

But we may have thrown away another piece of information that could have been even more useful.

Before the student saw the mark, what did the student think the mark would be?

And why?

“About 75. I think I understood most of it, but I lost time on the final problem and I’m not sure about two algebra steps.”

Then the paper returns at 73.

Something more than a score has appeared.

The learner has made a prediction about their own performance, supplied reasons for it, and received a receipt from the world.

That gap between prediction and outcome is educational information.

Sport makes prediction visible

Elite sport is full of forecasts.

How fast am I ready to run? How much can I lift? How hard was that session? Can I sustain this pace? Is my start improving? How close am I to competition condition?

The athlete has an internal estimate.

Then the body performs.

Then the stopwatch, distance, load, video or competition gives another reading.

Usain Bolt gave a particularly useful example before racing in London in July 2015. After an injury-disrupted season and months of training, he said he would not really know how ready he was until he raced. He knew what he had been working on, especially his first 40 metres, but competition would provide evidence that training alone could not fully supply.

That is not lack of confidence.

It is a sophisticated recognition that an estimate remains an estimate until performance tests it.

Prediction turns assessment into a loop

Ordinary assessment often looks like this:

Attempt → score

That is useful, but incomplete.

A calibration-oriented assessment can look like this:

Predict → attempt → score → compare → explain → update

Now the student is not only learning the subject.

The student is learning whether their own internal judgement can be trusted.

Researchers actually measure this

Metacognitive calibration research compares what people predict with what they subsequently do.

In physical education, researchers asked students to predict their performance on sport tasks and compared those predictions with actual results. Across three experiments involving 388 students, overconfidence appeared repeatedly, higher performers were generally more accurate, and the size of calibration error depended partly on task characteristics.

Classroom studies show a similar reason to care. Some students repeatedly predict substantially higher examination grades than they later earn, and those prediction errors can persist even after earlier assessments provide corrective evidence.

The important point is not that students are bad at knowing themselves.

The important point is that knowing your own performance is itself something that can be measured, examined and improved.

Do not ask only for a number

If we ask only, “What score do you think you got?”, students can simply guess.

The better question is:

“What do you predict, and what evidence are you using?”

A learner might say:

“I think I’m around 80 because I could explain every answer, but I may have two calculation errors.”
“Probably 55. I recognised the topics, but I did not know how to start the unfamiliar questions.”
“I have no idea. I could not tell which answers were secure.”
“About 70, but I’m only moderately confident because this paper was very different from our practice.”

Those statements contain much more information than the predicted score alone.

They expose what cues the learner is using to judge themselves.

The wrong prediction is not a failure

Suppose a student predicts 85 and receives 52.

That gap looks bad.

But if we use it properly, it can be extraordinarily useful.

Ask:

What made 85 feel plausible?
Which answers did you think were correct but were not?
Did familiarity with the question feel like mastery?
Did you overlook errors because the method looked familiar?
Were you judging effort rather than correctness?
Were you comparing yourself with previous work rather than the demands of this paper?

Now the 33-point prediction error has become a diagnostic object.

We are not humiliating the learner for being wrong.

We are locating the reason their internal measuring instrument gave a poor reading.

And a correct prediction can still be shallow

Suppose another student predicts 70 and receives 70.

Perfect?

Not necessarily.

Perhaps they always predict 70.

Perhaps they guessed.

Perhaps their reasoning about which parts were strong and weak was completely wrong even though the total happened to land on the right number.

Good calibration is not one lucky match.

It is repeated, explainable correspondence between judgement and performance across time and conditions.

Why this matters for education

Students make predictions about themselves constantly, even when nobody asks them to write the prediction down.

“I know this already.”

“I need another three hours.”

“I’m not ready for the harder question.”

“This answer is definitely right.”

“I should ask for help.”

Every one of those judgements can alter what the learner does next.

If the internal estimate is poor, study decisions can be poor even when the learner has plenty of ability.

If the estimate improves, the learner can allocate effort, seek help, choose difficulty and prepare for assessment more intelligently.

Do not wait for the score to begin learning from the score. Make a prediction first.

That simple habit creates something education rarely captures: evidence about how accurately the learner understands their own performance.