No national study answers it directly -- so this page runs its own calculation on the nearest data that exists, and finds a pattern nobody else has published.
Book a Free ConsultationNo published UK dataset tracks GCSE mock exam results through to final grades. The nearest available evidence is a December 2025 study by FFT Education Datalab, covering 740 schools and around 140 thousand pupils, and it measures a different figure: teacher-submitted predicted grades, not mock exam accuracy. An original calculation on FFT's own public data, run for this page, finds that all 24 GCSE subjects over-predict the pass grade (4+), while 22 of 24 under-predict the top grade (9).
A mock grade is a snapshot, not a sentence -- and the summer result is still being written.
Our expert GCSE tutors build a subject-by-subject plan from exactly what a mock paper exposed, on any exam board, well before the real exam season begins.
Book a Free Consultation Message us on WhatsAppThe question of exactly when a mock happens is a separate one, covered with a school-by-school breakdown here; this page picks up once the result is in hand and looks at what it does, and does not, tell you.
The honest answer is no, and it is worth showing rather than asserting, because a specific, confident-sounding number for this question circulates in parent forums and on tutoring blogs with no source attached anywhere. Four organisations were checked directly, each the kind of body that would have published this if anyone had.
FFT Education Datalab is the obvious first stop: it publishes more UK exam-outcome research than any comparable body, including the prediction-accuracy study this page draws on below. Its own site search for "mock exam" returns a single result, a 2020 piece on A-level grade trends that has nothing to do with mock accuracy. Searching "mock" alone returns two further results, neither about mocks predicting anything either.
Ofqual, the exams regulator, publishes research on teacher-assessed grades, appeals and access arrangements. A search of its own website for mock exam grades and predictions surfaces guidance on retaining evidence for exam-day disruption and open letters to schools, not a predictive study.
The Education Endowment Foundation runs one of the UK's most-cited databases of what actually improves attainment. Searching its site for "mock exams" returns the plainest possible answer: "No results found for 'mock exams'."
The Sutton Trust, which studies exactly the kind of access-and-outcomes question this touches, returns ten results for the same search -- and every one of them dates from 2020 to 2022, the period when GCSEs were cancelled and grades were assessed by teachers instead of sat as exams. None addresses whether an ordinary mock predicts an ordinary final result.
A figure repeated confidently on more than one page in this space -- that students "typically improve by two to three grades" after a bad mock -- has, on this evidence, nowhere to come from.
If no study answers the mock question directly, the next best thing is a study that answers something adjacent -- provided the difference between the two is stated plainly every time it matters, not once in a footnote.
FFT Education Datalab published exactly this kind of study on 10 December 2025. The researchers report that "740 of our participating schools submitted both predicted grades and results" and that, in total, "our dataset contains around 1 million predicted-actual grade pairs for 140 thousand pupils." That is a genuinely large sample -- comparable in scale to a full year group spread across several hundred secondary schools.
The label is the whole point, though. This is not mock exam data. FFT is explicit that these are "predicted grades submitted to FFT for the research project during the May/June collection window" -- teachers' professional judgements, entered months after a Year 11 mock, not the mock result itself. A school's predicted grade may be informed by a mock, by coursework, by classroom performance across the year, or by all three; FFT's own data cannot separate those inputs. Every figure below describes that predicted-grade dataset, never a mock paper, and the distinction is repeated everywhere it appears on this page rather than assumed after the first mention.
Run against the full chart data FFT published alongside its December 2025 study -- the raw predicted and actual percentage for every grade in 24 GCSE subjects -- the pattern at the headline pass grade is unambiguous: every single subject over-predicts it.
Across all 24 subjects, the grade 4-and-above pass rate is over-predicted, by 5.45 percentage points on average. Design & Technology shows the widest gap: teachers' predictions put 74.4% of entries at grade 4 or above, against an actual 64.6% -- a 9.8-point gap. Music (+9.7 points), Computer Science (+8.2), Spanish (+8.1) and PE / Sports Studies (+7.8) sit close behind. At the other end, Biology's gap is the narrowest of any subject at +1.3 points (94.0% predicted against 92.7% actual), with Mathematics (+1.4), English Literature (+1.8), Chemistry (+2.0) and Sociology (+2.3) the next smallest.
FFT's own published cohort figure -- "across all subjects, 78% of predicted grades were at grade 4 or above vs 74% of results" -- describes the same direction at a national, entry-weighted level. The subject-by-subject calculation above, run for this page rather than published anywhere by FFT itself, shows that this is not one or two subjects pulling a national average out of line: it is all 24, every time, to varying degrees.
At the very top of the grade scale, the pattern runs the other way. Of the same 24 subjects, 22 under-predict grade 9 -- teachers' predictions, on average, are slightly more conservative about who reaches the very top grade than the results turned out to be.
The gap is smaller than at the pass grade -- 1.09 percentage points on average, against 5.45 -- but the direction flips consistently. Religious Studies shows the widest gap (5.0% predicted against 7.2% actual, a 2.2-point under-prediction), followed by Art and Design (Fine Art) at -2.1, Art and Design at -2.0, French at -1.9 and History at -1.8. Only two subjects buck the pattern: Mathematics, where predictions were marginally too high (4.2% predicted against 4.1% actual, +0.1), and PE / Sports Studies (3.8% predicted against 3.3% actual, +0.5).
At grade 7 and above -- the midpoint between a pass and the top grade -- the system is close to neutral. Thirteen of the 24 subjects over-predict, eleven under-predict, and the average gap is just 0.98 percentage points, against FFT's own published cohort figure that "25% of both predicted and actual grades were at grade 7 or above." Put the three grade thresholds together and a single shape appears: predictions run optimistic about clearing the pass grade, are roughly even-handed in the middle, and turn slightly pessimistic about the very top.
Knowing the national pattern is one thing. Knowing where your own child sits inside it is another.
Our GCSE tutors work from a student's actual mock paper and school reports to build a subject-specific plan, rather than relying on a national average that may not describe this particular child.
Book a Free ConsultationThe size of the gap varies a great deal by subject, and knowing where a particular subject sits is more useful than the national average on its own. The tables below list the subjects with the largest and smallest gaps at each grade threshold, drawn directly from FFT's published chart data.
| Subject | Predicted 4+ | Actual 4+ | Gap |
|---|---|---|---|
| Design & Technology | 74.4% | 64.6% | +9.8 |
| Music | 82.8% | 73.1% | +9.7 |
| Computer Science | 76.0% | 67.8% | +8.2 |
| Spanish | 78.2% | 70.1% | +8.1 |
| PE / Sports Studies | 79.7% | 71.9% | +7.8 |
| Biology | 94.0% | 92.7% | +1.3 |
| Mathematics | 75.6% | 74.2% | +1.4 |
| English Literature | 76.5% | 74.7% | +1.8 |
| Chemistry | 94.4% | 92.4% | +2.0 |
| Sociology | 72.3% | 70.0% | +2.3 |
Source: FFT Education Datalab, prediction-accuracy chart data, December 2025 -- five widest and five narrowest gaps at grade 4 and above, original calculation for this page.
| Subject | Predicted 9 | Actual 9 | Gap |
|---|---|---|---|
| Religious Studies | 5.0% | 7.2% | -2.2 |
| Art and Design (Fine Art) | 3.1% | 5.2% | -2.1 |
| Art and Design | 2.7% | 4.7% | -2.0 |
| French | 5.0% | 6.9% | -1.9 |
| History | 3.9% | 5.7% | -1.8 |
| Mathematics | 4.2% | 4.1% | +0.1 |
| PE / Sports Studies | 3.8% | 3.3% | +0.5 |
Source: FFT Education Datalab, prediction-accuracy chart data, December 2025 -- five widest grade-9 under-predictions plus the two subjects that buck the pattern, original calculation for this page.
No single subject group explains the pattern on its own. Practical and creative subjects -- Design & Technology, Music, Computer Science, PE -- cluster at the wide end for the pass grade, while the core sciences and Mathematics cluster at the narrow end. At grade 9, humanities and languages -- Religious Studies, Art and Design, French, History -- show the widest under-prediction, while Mathematics and PE are the only two subjects where the top grade was, if anything, slightly over-predicted rather than under-predicted.
FFT's data cannot answer this directly -- it measures the size of the gap, not its cause -- but the shape is consistent with how a predicted grade actually gets set. A teacher predicting a borderline grade 4 candidate has a strong professional and pastoral incentive to give the benefit of the doubt: a grade 3 changes a pupil's sixth-form and college options in a way a grade 4 does not, and a "nearly there" judgement tends to round up. At the very top, the incentive runs the other way -- a grade 9 is deliberately calibrated to be rare, and a teacher predicting it for a pupil who has never quite produced grade-9 work under exam conditions has more to lose professionally by over-claiming than by holding back.
This is a plausible mechanism, not a finding FFT itself makes, and it should be read as such. It sits alongside a separate, better-evidenced explanation for why a mock grade specifically -- as opposed to a predicted grade -- can diverge from a final result: mock papers are marked against boundaries set months before the real exam's own boundaries exist, a mechanism covered in full separately.
My son was low in confidence after COVID and the gaps in learning were building up. Using this service helped plug those gaps, rebuild confidence, and improve his ability to learn properly again.
We arranged maths and English tuition ahead of GCSEs and both tutors were brilliant. My daughter's grades improved, and by results day she achieved two grades higher than in her mock exams.
I attended an English bootcamp with Leading Tuition and the experience was excellent. The teacher gave focused attention to the small group while keeping the lessons interactive and engaging.
The subject-level gaps above describe a national average, and averages are exactly the wrong tool for judging one child's own result. FFT's own study is explicit about this, in terms stronger than most summaries of it: "predicted grades tend to match results quite closely at the cohort-level, but not at the pupil-level."
The pupil-level figures explain why. FFT reports that "in all but four subjects, the majority of predictions were different from results, that is, predictions matched results for fewer than 50% of pupils." English literature is one of the starkest cases: despite having an average predicted grade almost identical to the average result, only 34% of individual predictions matched the actual grade achieved. Combined science matches even less often -- FFT records it as having "the lowest percentage of predictions which matched results, around 30%" -- though the subject's larger number of possible grade outcomes makes that particular comparison harder to read cleanly. FFT names only Mathematics and the three art and design subjects as the ones where "more than 50% of predictions matched results"; every other subject falls below that bar.
In other words: the national pattern above -- pass grades over-predicted, top grades under-predicted -- is real and measurable at a population level, and a prediction system built by teachers, working from a full year of coursework, classroom performance and mock results together, still gets any individual pupil's exact grade wrong more often than it gets it right, in most subjects. A mock result sitting above or below the national pattern for its subject says something about that subject's typical prediction behaviour. It says very little, on its own, about which side of the coin flip a specific child will land on.
No, and the difference matters for anyone reading this page while also preparing a sixth-form or college application. Everything above describes GCSE predicted grades in FFT's own research dataset -- a different exercise, a different qualification and a different purpose from the predicted grade a school submits for A-level or IB study, which is covered separately, drawing on its own accuracy research rather than FFT's GCSE dataset.
If a mock result has raised a real question about tier, subject choice or exam board, that's worth a direct conversation.
Our GCSE tutors review a mock paper subject by subject and turn it into a specific, board-matched plan for the months before the real exam.
Book a Free Consultation Message us on WhatsAppA mock grade, a predicted grade and a target grade are also three different numbers within GCSE itself, produced by three different processes -- a distinction set out in full elsewhere for a reader trying to work out which figure a specific form is actually asking for.
One further question sits just outside this page's own scope but comes up often alongside it: whether a sixth form is legally entitled to use a mock or predicted grade to decide who gets a place at all. That is a School Admissions Code question, not a statistical one, and it has its own answer, including the one Schools Adjudicator ruling on record that touches it directly, addressed in full here.
Given all of the above, the most useful response to a mock grade is not to search for a national multiplier that does not exist, and not to treat a single result as a verdict at either end of the scale -- reassuringly high or worryingly low. Two more specific, genuinely useful steps follow instead.
The first is to use the mock paper itself, not the grade it produced. A grade is one number; the paper underneath it usually shows exactly which topics, question types or technique gaps cost the marks, which is diagnostic information a national dataset can never provide about one specific child -- the basis of a proper mock paper review rather than a grade read in isolation.
The second is to check what is still procedurally open, separate from the grade itself -- tier of entry, access arrangements, and whether a second mock round, where a school runs one, can plausibly inform either decision before the relevant deadline. Those are calendar questions with real board-set dates behind them, covered in full here, alongside dedicated post-mock GCSE tuition for the months in between.
What a mock grade cannot tell a family, on the evidence this page has assembled, is a fixed number of grades a student will move by before the summer. Nobody has published that figure because, at an individual level, it does not exist as a stable, transferable statistic -- only a subject-specific starting point does.
No national dataset answers this directly for mock exams specifically. The nearest evidence, FFT Education Datalab's December 2025 study of 740 schools, measures a related but different figure: teacher-submitted predicted grades, not mock results. Even there, FFT's own conclusion is that predictions "match results quite closely at the cohort-level, but not at the pupil-level" -- in most subjects, fewer than half of individual predictions matched the grade a pupil went on to achieve. A single mock grade, on its own, is not a reliable forecast of one child's final result.
Both, depending which grade is being predicted. An original calculation on FFT's published data found that all 24 GCSE subjects over-predict the grade 4-and-above pass rate, by 5.45 percentage points on average, while 22 of 24 subjects under-predict the top grade 9, by 1.09 percentage points on average. At grade 7 and above, the two effects roughly cancel out, with predictions running close to neutral. The two subjects that buck the grade-9 pattern are Mathematics and PE / Sports Studies, where predictions were marginally too high rather than too low.
Measured by how often an individual pupil's predicted grade matched their actual result, rather than by the size of the average gap, FFT's December 2025 study names English literature (34% of predictions matched) and combined science (around 30%, the lowest of any subject) as the least accurate at the pupil level. FFT names only Mathematics and the three art and design subjects as the subjects where more than 50% of individual predictions matched the final result -- every other subject falls below that threshold, including several where the average predicted grade was almost exactly right.
No. A mock grade comes from an actual internal exam, usually sat in November or December of Year 11. A predicted grade is a teacher's professional judgement of the grade a pupil is likely to achieve, which may draw on the mock result alongside coursework and classroom performance across the year, and is typically submitted later, in May or June. The dataset this page relies on measures predicted grades specifically, not mock results, and the two should not be read as interchangeable. A full comparison of mock, predicted and target grades is set out separately.
No, and this page will not invent one. Our two published, owner-confirmed outcome figures are that 91% of our students achieve their desired grades, and that 95%+ of our selective-school students receive an offer -- neither of which measures the specific gap between a mock result and a final grade, which is what this page is actually about. No credible source, including FFT's own research, publishes a reliable figure for how many grades an individual student typically moves between a mock and the summer exam.
Start with the paper itself rather than the grade alone -- it usually shows exactly which topics or techniques cost the marks, which is more useful than any national average. Then check what is still genuinely open procedurally: exam tier, access arrangements, and whether a second mock round could realistically inform either decision before the relevant board deadline. Those calendar questions are covered in full separately, alongside guidance on turning a specific mock paper into a subject-by-subject revision plan before the real exam.
Whatever your child's mock grade showed, our GCSE tutors turn the actual paper into a subject-specific plan, on any exam board, ahead of the real exam.
Book a Free ConsultationRated 4.8 out of 5 on Trustpilot from 58 reviews, and 91% of our students achieve their desired grades.