What are predicted grades, and how are they produced?
A teacher's professional judgement of the grade a student is most likely to achieve, used by universities to make conditional offers before results exist. They are estimates, not promises, and their known inaccuracy is the central problem with the system built on them.
How they are actually produced:
Evidence from mock exams, classwork, coursework and past assessments.
Comparison with previous cohorts who performed similarly at the same stage.
Prior attainment data, including projections from earlier qualifications.
Departmental and senior moderation, so an individual teacher's optimism is checked.
Why they are unreliable, and this is well documented. Research has repeatedly found that a minority of predictions match the final grade — with over-prediction being far more common than under-prediction. Predicting a whole year ahead, on incomplete evidence, for teenagers whose performance is still changing, is genuinely difficult.
The equity problem. Studies have found systematic differences in accuracy for students from different backgrounds, with high-attaining disadvantaged students more likely to be under-predicted — which affects the universities they apply to, and therefore where they end up. This finding is the main driver of reform proposals.
Why the system persists. The UK applies before results because the admissions cycle runs across the academic year. Post-qualification admissions — applying after results — has been repeatedly proposed and repeatedly not adopted, because it would compress the timetable severely for universities, schools and students.
What this means practically:
Predictions are not a ceiling. Missing a prediction is common, and exceeding one happens too.
Universities know they are unreliable and interpret them in context.
Clearing and adjustment exist precisely because results differ from predictions.
You can ask how yours was reached, and schools should explain the evidence.
Challenging one is possible through the school, but must rest on evidence rather than preference.
Contextual admissions use background data alongside grades, which mitigates some of the effect.