How scores are compared

1. Better than X% of children

Each paper is compared only with children in the same year group who sat the same paper. For each of those children we take their best genuine attempt. “Better than 74%” means 74% of those children scored lower.

A sitting of a full paper counts as genuine once at least a quarter of the questions are answered and it lasts five minutes or more. A paper opened and submitted a minute later, or one where the timer ran out with nothing chosen, still has its own result page but is left out of every comparison, average and progress chart. Nothing is deleted and no score is changed. Mini-tests count as soon as one question is answered.

The five-band chart appears once 15 children have sat the paper, because a chart of eight children would look precise and mean little. Exact head-counts and a position (“6th of 63”) are shown once more than 50 children have sat it.

2. The age-adjusted score

The real 11+ standardises by age. A child born in September is almost a year older than a classmate born in August, and on the same paper the older child scores higher on average. So the exam gives the younger child a little more credit for the same raw mark.

GrammarMock does the same with one fixed rule. Each child’s birth month is placed in the school year, September first and August last, and the raw percentage is moved by 0.75 points per month away from the middle of the year. The youngest gain about four points; the eldest lose about four.

The 0.75 was measured from GrammarMock’s own attempts on 8 September 2026: across 1,116 genuine attempts, each month of age was worth roughly one percentage point on the same paper, and the youngest in a year group sat about seven points below the eldest. The figure is fixed and published here rather than tuned quietly; when it is re-measured, this page will change.

The comparison in section 1 is made on age-adjusted scores, for everyone. The raw score stays exactly as it was.

3. Why there is no standardised score yet

Exam results are reported on a scale with an average of 100, where most children fall between about 69 and 141, and SW Herts adds the two papers together. That scale needs a large sitting: it is built from the spread of everyone’s marks, and with twenty children it would move several points every time one more child sat the paper.

GrammarMock will show a score on that scale once enough children have sat a subject, and it will be labelled as GrammarMock’s scale, not the exam’s. Until then the percentile is the honest number.

4. Who the comparison is against

The other children are families who use GrammarMock, not every child sitting the real exam. Families who seek out practice papers tend to be further along, so “better than half” here is probably better than half of a stronger group. Treat the comparison as a guide to what to work on next, not as a prediction.