University of Oxford
degree classification explained

The conventions are set per Final Honour School, not per university

At most universities one document governs every undergraduate. At Oxford, each Final Honour School has its own examination conventions, agreed by the relevant faculty or teaching committee and published annually. They set out how papers are marked, how marks are combined and what is required for each class — and they differ from school to school. A board of examiners may make only minor deviations from its conventions, and then only in exceptional circumstances and with the consent of the teaching committee or the Proctors.

It can be more granular still. The Department of Computer Science publishes one conventions document covering three different Final Honour Schools — Computer Science, Mathematics and Computer Science, and Computer Science and Philosophy — and that document states that papers falling under the Mathematical Institute or the Faculty of Philosophy are governed by those faculties’ conventions instead. A single joint-school candidate’s result can therefore be governed by three sets of rules at once.

Your marks are USMs, and a USM is not a raw mark

Oxford reports marks as University Standardised Marks on a 0–100 scale, and USMs used in classification are always whole numbers. The stated purpose of the USM is to allow direct comparison between examinations in different subjects: examiners recalibrate raw marks to arrive at the USMs, and are explicitly permitted to scale different papers differently to compensate for differences in difficulty. The recalibration must not disturb the order of candidates on a paper, and examiners are told to note the distribution of USMs above 70 in a normal year and not to depart from it without good reason.

The practical consequence is the one that defeats a calculator: the raw mark you think you scored is not the number that gets classified, and the relationship between the two is set by the examiners for that paper in that year.

Parts are weighted differently in different schools

There is no Oxford equivalent of “year two counts 40%, year three counts 60%”. In Mathematics and Philosophy, for instance, candidates are classified on Parts A and B together, with the two Parts scaled to a ratio of 2 : 3 — and before that, USMs for individual Mathematics papers are weighted to reflect the proportion of the course examined in each subject, while no weighting at all is applied to the Philosophy papers. Another school will do something else. The weighting is a property of your Final Honour School, and you have to read its conventions to find it.

Some schools classify on papers, not only on an average

This is the part a weighted-average calculator cannot express even in principle. Mathematics Part B classifies on the average USM or on a strong paper rule: a candidate satisfies the First Class rule if at least six units from Parts A and B lie in that class or better. A student below the average threshold can still be classified on the profile of their papers.

The conventions also contain rules that hand the decision to people rather than to arithmetic. In Mathematics and Philosophy, a candidate at one specified borderline is classified II(i) only after the examiners have carefully gone over that candidate’s marks. No formula models “the examiners looked at it”.

What to read instead

Find the examination conventions for your own Final Honour School and your own year, published on your faculty or department’s website. They are the authoritative document, they are reissued annually, and they will tell you the weighting, the classification rules and the qualitative class descriptors that apply to you. A figure from any other source — including this page — is a general description, not your algorithm.

Why this site does not offer a calculator for either

Every other UK page here takes the marks you enter, applies the algorithm printed in that university’s own regulations, and tells you the class. That needs three things to be true: a weighting that is fixed and published, boundaries that are fixed and published, and an algorithm that depends only on your marks. Oxford has none of the three. Cambridge now has the first, and does not have the other two. A calculator built anyway would produce a number that looks authoritative and is not, which is worse than no page at all — so this is a page instead.

Sources

Everything above is read from Oxford’s own published conventions:

• Mathematics Part B Examination Conventions 2021–22, Mathematical Institute — the definition and purpose of the USM, recalibration of raw marks without disturbing candidate order, the aggregation of Part A and Part B USMs, and the strong paper rule requiring at least six units in a class.
• Final Honour School of Mathematics and Philosophy, Examination Conventions 2024–25 — classification on Parts A and B together, the 2 : 3 scaling between the Parts, the weighting of Mathematics papers by proportion of the course with no weighting applied to Philosophy papers, symmetric rounding, and the borderline at which examiners go over a candidate’s marks before classifying.
• Prelims Mathematics Examination Conventions 2023–24 — the instruction to examiners on the distribution of USMs above 70, individual consideration for candidates outside the standard pattern, and the status of conventions as subsidiary to the Examination Regulations.
• Department of Computer Science, Examination Conventions Parts A and B 2022–23 — one conventions document covering three Final Honour Schools, and its deferral to the Mathematical Institute and Faculty of Philosophy conventions for papers under their responsibility.

Conventions are reissued every year and differ by school. Your own school’s current document is always authoritative.