Two percentile formulas, both correct
Percentile rank has more than one definition and they disagree whenever scores are tied, which in a class of marked exams is most of the time. The strict version counts only the scores strictly below yours and divides by the class size. The midpoint version counts those below plus half of those equal to yours. Standardised tests generally publish the strict form. Statistics texts and anything dealing with heavily tied data generally use the midpoint form, because it has the useful property that the median score lands at the 50th percentile instead of somewhere below it.
This page shows both, alongside a third figure that people usually mean when they say percentile in a school context: the top share, which is the fraction of the class at or above the score. Being told you are in the top 10 percent is a statement about that last figure, and it is not the same number as either percentile rank. When a scholarship or an admissions form asks for a percentile, it is worth checking which of the three it wants.
Ties, and what they do to a rank
When several students share a score, standard practice is that they share the highest rank the group occupies and the following ranks are skipped. Three students tied at the top are all rank 1, and the next student is rank 4, not rank 2. This calculator reports the shared rank and the span of ranks the tied group occupies.
Ties produce results that look wrong until you see the mechanism. A perfect score on an easy exam can leave a student outside the top 10 percent, if more than 10 percent of the class also scored perfectly — no error has occurred, there is simply no room above them. This is one of the strongest arguments against reading a rank as a measure of ability: it is bounded by the resolution of the assessment, and an exam that cannot separate the top students will not rank them.
Rank measures the cohort
The number this page produces depends on two things, and only one of them is the student. Change the other twenty-nine people and the rank changes without a single answer being written differently. That is not a flaw in the arithmetic; it is what a relative measure is. The consequences show up in real decisions:
| Situation | What rank does |
|---|---|
| Small class, under about 15 | Each place is worth several percentile points, so rank is noisy |
| Strong cohort | Depresses the rank of students performing well in absolute terms |
| Easy assessment | Compresses everyone near the ceiling and ranks on noise |
| Mixed course levels | Rank across different courses compares scores that are not comparable |
Why fewer schools report it
Roughly half of US high schools have stopped putting class rank on transcripts, and the share has been falling for years. The stated reasons are consistent: rank penalises students at academically strong schools, it converts differences of a few hundredths of a GPA point into visibly different labels, and it creates incentives to avoid demanding courses that might cost a place. Many now report a decile or quintile band instead, or nothing at all.
Where rank still has teeth is in automatic-admission and scholarship rules that reference a hard threshold — the top 10 percent provisions some state systems operate, for instance. There the number is doing real work and is worth calculating precisely. Everywhere else, the score and the distribution behind it carry more information than the rank does, and this page prints both so you can see the difference between them.
Questions people ask
Is percentile the same as percentage score?
No, and conflating the two is the most common error here. A percentage score is how many marks you earned out of the available marks — it is a property of your paper alone. A percentile is the share of the group you outscored — it is a property of the group. A student can score 95 percent and sit at the 40th percentile on an easy exam where most of the class also scored above 90, and can score 62 percent and sit at the 95th percentile on a brutal one. When both figures are available, read them together: the percentage says how much of the material was mastered, the percentile says how that compared.
How do I work out class rank across several subjects?
You do not rank the subjects separately and average the ranks — that produces a number with no meaning, because a rank of 5 in a class of 12 and a rank of 5 in a class of 300 are incomparable. The standard approach is to compute a credit-weighted GPA across all courses first, then rank students on that single figure. Schools that weight honours and AP courses apply the weighting before ranking, which is exactly why weighting policy and rank are so tightly linked and why changing one changes the other.
The list has 20 scores but 22 students. Does that matter?
It changes every figure on the page, so it is worth fixing. Percentile and rank are both computed against the list you paste and nothing else, so missing scores shrink the denominator and inflate the top-share figure. Absent students, incompletes and late sittings all cause this. If some scores genuinely do not exist yet, the result is provisional and should be read as such rather than reported as a rank.
Can I use this for times rather than marks, where lower is better?
Yes — set the direction field so the lowest value takes rank 1. That handles race times, lap times, error counts, golf scores and anything else where smaller wins. Everything else on the page behaves identically, except that the mean and standard deviation are direction-neutral by definition and read the same either way. The top-share figure still means the leading share of the field, which for times is the fastest.