Why Grade Boundaries Change Every Year

The order in which things happen

Nothing about a boundary exists while you are sitting the exam. The papers are taken, they are marked, and only then does awarding begin. Senior examiners look at real scripts around the likely boundary marks and compare them with work from previous years held at the same standard, alongside statistical evidence about the cohort.

This is the fact that makes most boundary anxiety pointless. There is no number being kept from you in May, because in May there is no number. What exists is a range you can infer from previous series, and the awarding decision that will settle it has not been made.

One board's column, two consecutive summers

Here is Edexcel's Higher tier for GCSE Mathematics, June 2024 against June 2025, out of 240 marks. Read it as a column rather than as individual grades.

Every boundary moved up, and it moved up most at the top: twenty marks at grade 9, nineteen at grades 8 and 7, eleven at grade 4. That is the signature of a paper that gave candidates more access than the previous year's. Awarding responded by raising the price of every grade so that a grade 8 in 2025 meant the same thing a grade 8 meant in 2024.

Now put AQA beside it. AQA's Higher tier grade 9 was 219 in both years. Its grade 8 was 191 in both years. Grade 7 moved by one mark, from 163 to 164. Same subject, same regulator, same summer — and one board's papers behaved very differently from the other's.

• Grade 9 — 197 → 217
• Grade 8 — 167 → 186
• Grade 7 — 137 → 156
• Grade 6 — 105 → 121
• Grade 5 — 73 → 87
• Grade 4 — 42 → 53

What is being held constant is the grade, not the mark

The principle behind awarding is comparable outcomes. Ofqual's framing is that if the ability of students stays the same and nothing else that affects performance changes, results should be stable over time — and the system exists to accommodate the fact that the mix of students taking a qualification does change from year to year.

So predictions built from prior attainment sit alongside examiner judgement, and the boundary lands where those two agree. It follows that a national rise in raw marks does not automatically produce a national rise in grades. If a cohort scores higher because the paper was more accessible, the boundary moves up and outcomes hold steady, which is exactly what the Edexcel column above shows.

What this means for mocks and predicted grades

A mock marked against last year's boundaries produces a grade that describes last year. When the same board's grade 7 has moved by nineteen marks in one step, a student two marks above the 2024 boundary has been told something that would have been false in 2025.

The habit we teach is to work in marks and topics rather than in percentages and grades. A student who knows they lose eleven marks a paper to careless algebra and nine to unattempted geometry has something to act on. A student who knows they are "a borderline 6" has a label that will move under them.

There is one honest exception. A student who is fifteen marks clear of the highest boundary that grade has taken in recent series can stop worrying about awarding, because no plausible movement reaches them. Getting to that position is the whole purpose of starting early.