Learn Better With AI

Methods · Assessment & diagnosis

H6

Grading simulation

Have your work marked the way an examiner would mark it, strictly, against the scheme, with the marks you did not get explained.

When to use this
When checking yourself
Learning impact
Reliable gain
Learning evidence
Credible with caveats
Where you are with the topic
Checking readiness
How often to use it
Worth repeating often
What AI is doing (1)Evaluate your work
What AI is doing (2)Play examiner, peer, or client

When to use it

  • An exam is close and you want itemisation, not a vibe check.
  • You finished a practice answer and need to know which available marks you left on the table.
  • You want a dress rehearsal that feels strict, not reassuring.

What to ask AI

Paste the real mark scheme and your answer. Ask for a strict examiner mark with no benefit of the doubt.

Demand mark-by-mark itemisation: which marks you earned and where, which available marks you did not earn and why.

Ask: “Where did I write something true that earned no marks?”, that catches command-word failures.

Ask for the single highest-value fix if you can only change one thing.

How it helps

Feedback often arrives too late and too rarely. An essay comes back in three weeks with a grade and four comments.

By then you have written two more, and the comments land on nothing in particular.

The gap this fills is not the grade, you can estimate that. It is the marks you did not know were available.

Mark schemes reward things students miss stating: assumptions, signposting, definitions before use, intermediate steps, limitations.

Those are invisible until someone marks you for missing them. Rubric self-assessment puts you in the marker seat; this one has marking done to you, strictly, for the itemisation.

How to run it

  1. Supply the real mark scheme. Without it you get a plausible number with no basis.
  2. Ask for strictness explicitly: “no benefit of the doubt.”
  3. Ask for mark-by-mark itemisation, not a grade with comments.
  4. Ask where you wrote something true that earned no marks.
  5. Ask for the single highest-value fix.
  6. Compare against your own predicted mark. The gap is calibration information.

Example exchange

You

Here's the mark scheme and my 12-mark evaluate answer. Can you mark it like a strict examiner and itemise what I earned versus what I missed?

AI

Paste both. I'll itemise only, no soft total until the lines are listed.

You

Okay, scheme and answer pasted.

AI

You earned 5 of 12. Missed: evaluation moves, a definition before use, a limitation, and structure signposts. Your description of the study is accurate but unmarked. “evaluate” doesn't pay for that. One fix: turn two descriptive sentences into weigh-and-judge moves.

Copyable prompt

Here's the mark scheme [paste] and my answer [paste]. Mark it as a strict
examiner. Itemise: which marks I earned and where, which available marks I
didn't earn and why. No benefit of the doubt: I need the real number.

The Tell

Your simulated marks are consistently higher than your real ones.

The strictness instruction is not landing. Stop trusting the number and use only the itemisation.

Taking the grade and ignoring which lines earned nothing also fails the method.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Being marked “as an examiner would,” with justification, is formative rehearsal of standards (examiner simulation / assessment literacy). It builds awareness of how marks are allocated.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI marks are not official marks. Treat them as speculative rehearsal. Risky if you trust the number as a predicted grade; useful for where marks go.

Related methods