Examination Conventions 2026-27

1. Introduction

Examination conventions are the formal record of the specific assessment standards for the course or courses to which they apply. They set out how examined work will be marked and how the resulting marks will be used to arrive at a final result and classification of an award. This document sets out the examination conventions for the M.Sc. in Mathematics and Foundations of Computer Science for the academic year 2026-27. These examination conventions are approved annually by the Supervisory Committee for the M.Sc. in Mathematics and Foundations of Computer Science and by the Graduate Studies Committee in the Mathematical Institute. The Board of Examiners may only make minor deviations from these conventions in exceptional circumstances and only after the consent of the Proctors. This document is in all ways subsidiary to the current: 

Examination regulations;

Policy and Guidance for Examiners and others involved in University Examinations.

2. Examiners

The Board of Examiners consists of at least three (currently four) members, with at least one (currently two) being external to the University. The examiners will appoint assessors to help with the assessment of the course.

The current Board of Examiners consists of Prof. Kobi Kremnitzer (Internal examiner and Chair - Maths), Professor Tom Melham (Internal examiner - CS), Prof. Anupam Das (External examiner - University of Birmingham) and Prof. Jan Grabowski (External Examiner - Lancaster University).

3. Requirements of the Course

Each candidate must obtain passes on at least five sat written examinations or mini projects that include two at the Schedule II level and for the dissertation. Candidates must offer at least four courses from among those available in Michaelmas and Hilary terms. No candidate may enter for more than four courses in one term.

An overall award of distinction may be made to candidates who have shown excellence over the whole examination. An overall award of merit may be made to candidates who have produced work of particularly high quality in the whole examination.  

For written examinations the Examiners translate the raw marks on each paper into University Standardised Marks (USMs) out of 100. See Section 4 for more information on USMs.

4. University Standardised Marks

• 70-100: Distinction
• 65-69: Merit
• 50-64: Pass
• 0-49: Fail

5. Written Examinations

Mathematics papers - each mathematics paper will examine one unit. Most mathematics units will be assessed by a closed book exam. These will consist of three questions, each worth 25 marks. You may submit answers to as many questions as you wish, but only the best two answers will count towards the final mark for the paper. 

In all papers the questions set should, as a whole, be fairly spread across the syllabus.

Questions will be broadly similar in style to previous years questions with an easy start examining material explicitly covered in the course, followed by a part which tests understanding. Each question will be set so that a sound student can produce a complete answer within 35- 40 minutes. 

Each question should be divided into two to four parts and an indication of the raw marks available for each part of each question should be given on the question paper.

Computer Science papers -  each Computer Science paper will be assessed by a written in-person exam. You will have three hours to complete each exam. The requirements for each exam will be listed on the front of the question booklet. 

5.1. Analysis of marks

The Board of Examiners will assign USMs for each paper taken may scale the raw marks to arrive at the USMs reported to candidates.

The examiners may choose to scale marks where in their academic judgement:

• a paper was more or less difficult than other papers taken by students in a given year, and/or
• a paper has generated a spread of marks which is not a fair reflection of student performance on the University's standard scale for the expression of agreed final marks, i.e. the marks do not reflect the qualitative marks descriptors.

Such scaling is used to ensure that all papers are fairly and equally rewarded.

When scaling the raw marks on a paper the examiners will consider the following:

• the relative difficulty of the paper compared to the other Part C papers;
• the report submitted by the assessor who set and marked the paper.

Examiners will use their academic judgement to ensure that appropriate USMs are awarded and may use further statistics to check that the marks assigned fairly reflect the students' performances on a paper.  Examiners may also review a sample of papers either side of the classification borderlines to ensure that the outcome of scaling is consistent with the qualitative marks descriptors.

5.2. Examination Conduct

6. Submissions

Mini-projects

Several of the courses are assessed by mini-projects. Such mini-projects are commonly used for courses involving elements of programming, computing packages or data sets where it is felt that a timed, paper-based assessment would not be appropriate. The format of the mini-project can vary between courses and might include a computation or written report.

The examiners will send out notices to candidates detailing any specific requirements for the mini-projects.

An electronic copy will need to be submitted and the precise details of this submission process will be included in the notice to candidates.

Dissertation

Each project will be double marked, with one of the two assessors being the supervisor. The two marks will be reconciled through discussion between the two assessors and overseen by the examiners.

Dissertations, identified by your candidate number only, should be submitted online by the stated deadline at the end Trinity term the precise details of this submission process will be included in the notice to candidates.

Further information on disseration advice can be found on the course website: https://www.maths.ox.ac.uk/members/students/postgraduate-courses/msc-mfocs/information-current-students

6.1. 3.3 Penalties for Late or Non-Submission of Work

Late Submission of Coursework for the M.Sc. in Mathematics and Foundations of Computer Science (this includes mini-projects and the dissertation) is a serious matter and will usually result in academic penalties unless prior permission for late submission has been given by the Proctors. In the absence of such Proctorial permission, academic penalties will be as set out in Table 1 below. (Note that if the late submission penalty is higher than the mark for the submission, a candidate will be awarded a mark of zero.)

Where no work is submitted, a mark of zero will be awarded for that particular piece of work. The mark for any resit of the assessment will be capped at 50. Such a resit is only available once to candidates who initially fail the whole M.Sc. course.

6.2. 3.4 Penalties for poor academic practice in submitted work

The Examination Board shall deal wholly with cases of poor academic practice in submitted work and open-book online examinations where the material under review is small and does not exceed 10% of the whole.

Assessors should mark work on its academic merit with the board responsible for deducting marks for derivative or poor referencing. 

Determined by the extent of poor academic practice, the board shall deduct between 1% and 10% of the marks available for cases of poor referencing where material is widely available factual information or a technical description that could not be paraphrased easily; where passage(s) draw on a variety of sources, either verbatim or derivative, in patchwork fashion (and examiners consider that this represents poor academic practice rather than an attempt to deceive); where some attempt has been made to provide references, however incomplete (e.g. footnotes but no quotation marks, Harvard-style references at the end of a paragraph, inclusion in bibliography); or where passage(s) are ‘grey literature’ i.e. a web source with no clear owner.

If a student has previously had marks deducted for poor academic practice or has been referred to the Proctors for suspected plagiarism the case must always be referred to the Proctors. 

In addition, any more serious cases of poor academic practice than described above should also always be referred to the Proctors.]

7. Practicals

Practicals are offered as part of the Distributed processes, types, and programming and Geometric Deep Learning courses in 2026-27, instead of class teaching. Practicals play no part in the overall classification, but do appear on the student transcript. Where a student takes more than one course using practicals, the overall result will be an average of the marks for all courses.

Reports on practicals are marked by the demonstrating staff as each practical has been completed, and the Examiners receive these marks, together with the practical reports themselves. The demonstrating staff are not appointed as Assessors for the purpose of marking practicals, and it is therefore Examiners’ responsibility to determine what credit is given for each piece of practical work. The marks given by the demonstrating staff will serve as a guide, using the table below.

The Examiners will give no credit for practical work that was not submitted for marking by the deadline and signed by a demonstrator, unless there are extenuating circumstances. The following numerical procedure is suggested for processing the marks. Each practical is marked on a scale S+, S, S- that is explained in the Course Handbook. These marks will be converted to numbers using the following scale: 

The borderlines for passing the practicals are 50 for a Pass and 70 for a Distinction.

8. Approved Subjects

Candidates may, with the permission of the Course Director, substitute no more than one of the five required written assignments with an appropriate course of study from outside the published list, which need not be examined by written assignment. Where approved, this will be given a Schedule listing (Schedule I or II) and used in place of a mini-project of the appropriate schedule in the classification criteria for the course. Where an approved subject is taken, examination for the course is the responsibility of the exam board responsible for the setting of the paper (eg. the Part C Examiners where a Mathematics Part C course is taken), and follows the rules set out in the conventions governing the course offering the assessment.

Rules governing non-attendance at examinations and any consequent penalties are set out in full in the Examination Regulations (Regulations for the Conduct of University Examinations, Part 14). If a student will be prevented by illness or other urgent cause from attending one of their examinations they should contact their college office or college advisor as soon as possible. Any case of non-attendance at an examination involving illness or other medical condition will require written medical evidence and will usually be referred by the college to the Proctors. If the Proctors do not believe there are satisfactory reasons for non-attendance, or an application to the Proctors has not been submitted, a candidate will be awarded a mark of zero for that examination.

9. Vivas

All students shall give an oral examination on their dissertation and its background material, and you will be expected to give a short 10-15 minute presentation on the dissertation, after which there will be an opportunity for the Examiners to ask questions. You are not permitted to use presentation slides although use of a whiteboard is permitted. You must wear sub-fusc and are advised to take a copy of your dissertation with you to your oral examination. Students are permitted one A4 sheet of notes in the viva. 

10. Marking Conventions

Written examinations are marked by a single assessor or examiner according to a pre-agreed mark scheme which is strictly adhered to.  The examination scripts are then checked by an independent checker to ensure that all work has been marked, and that the marks have been correctly totalled and recorded.

Mini projects are set by those giving the course and are double-blind marked by that person and one other assessor (these two mark the work independently of each other). Each proposes a USM for the work. If the assessors’ marks do not differ by more than 10, and do not cross the pass/fail boundary, the final mark will usually be the average of the two marks, rounded to the nearest whole number (.5 is rounded up). However, if the marks differ by more than 10 marks, or cross the pass/fail boundary, the assessors will be required to have a further discussion about the assessment in order to agree on a final USM.

The exception is mini projects which have a model solution and marking scheme approved by the examiners. In such cases each script is marked by an assessor and this marking is checked independently to ensure that all parts have been marked and the part-marks have been correctly totalled and recorded.

The mini projects which are set are submitted to the Examiners for prior vetting, and the Examiners may moderate the marks given by assessors, in particular to achieve parity across subjects. The results will usually be published soon after the next meeting of the examiners and candidates will be advised of the USMs awarded. (Meetings take place in Hilary and Trinity terms, and in September after the dissertation vivas.)

Dissertations are marked independently by the dissertation supervisor and by a second assessor. These two marks are averaged to produce a provisional USM following the same procedures given for mini projects which are double-blind marked. Reconciliation is required if the two assessors’ marks differ by 10 points, or if the range crosses the pass/fail boundary. Each Dissertation will also be seen by at least one Examiner. The second assessor of the dissertation will normally be present at the oral examination, and the Examiners will determine the USM only after the oral examination has been held, taking into account all the evidence from the double-marking, the supervisor’s additional input, and from the oral examination.

11. Plagarism

You are reminded of the importance of avoiding any plagiarism, please see  http://www.ox.ac.uk/students/academic/guidance/skills/plagiarism for further guidance. Depending on their severity, cases of suspected plagiarism may be referred to the Proctors for investigation or may be dealt with by the board of examiners. This includes cases of self-plagiarism. If dealt with by the board of examiners as a case of poor academic practice, the examiners may deduct marks (for lack of adequate referencing, poor use of citation conventions etc) of up to 10% of the marks available for the assessment. Where the consequence of the marks deduction would result in both the failure of the assessment and of the programme, the case must be referred to the Proctors.

We will reserve the right to conduct follow-up viva voce exams to check candidates' understanding of the examined material, even where these are not currently specified in the Examination Regulations. 

The Mathematical Institute's policy for students on the use of generative AI in summatively assessed work is available via https://www.maths.ox.ac.uk/members/students/departmental-and-university-regulations.

12. Qualitative Marking Criteria

Written examination Qualitative Descriptors

Assessors setting questions should be asked to provide complete model solutions indicating everything that a candidate would be expected to write to answer the question fully.  The model solutions and marking scheme need to be sufficiently clear and comprehensive to be meaningful to an external examiner.  Those setting questions should be aware that solutions may be released to students in the future. 

The model solution for each question should be accompanied by a marking scheme out of 25.  The marking scheme should aim to ensure that the following qualitative criteria hold (see also the class descriptors given in appendix K):

20-25 marks A completely, or almost completely, correct answer, showing excellent understanding of the concepts and skill in carrying through the arguments and/or calculations; minor slips or omissions only.

13-19 marks A good though not complete answer, showing understanding of the concepts and competence in handling the arguments and/or calculations, and some evidence of problem-solving ability. Such an answer might consist of an excellent answer to a substantial part of the question, or a good answer to the whole question which nevertheless shows some flaws in calculation or in understanding or in both..

7-12 marks Standard material has been substantially and correctly answered with some possible minor progress on to other parts of the question.

0-6 marks Some progress has been made with elementary, accessible material.

Assessors should classify the parts of each question under the headings: 

B1: bookwork material: explicitly seen before;
B2: routine material: easily synthesized from material explicitly seen before;
S: similar to material seen before;
N: new rider, demanding good command of concepts and/or methods.

Mini-Projects Qualitative Descriptors

Mini-projects will be assessed with reference to the following qualitative descriptors.

70–100   The candidate has demonstrated an excellent understanding of almost all of the material covered with a commensurate quality of presentation and has completed almost all of the assignment satisfactorily, further subdivided by:

90–100 The candidate has shown considerable originality and insight going well beyond the straightfor­ward completion of the task set.

80–89 The work submitted shows a near-perfect completion of the task at hand, but does not meet the additional requirements above, or does but has some defects in presentation.

70–79 The work submitted is of a generally high order, but may have minor errors in content and/or deficiencies in presentation.

60–69 The candidate has demonstrated a good or very good understanding of much of the material, and has completed most of the assignment satisfactorily, without showing the level of excellence expected of the above USM range.

50–59 The candidate has demonstrated an adequate understanding of the material and an adequate ability to apply their understanding, without showing the level of understanding expected of the above USM range.

40–49 The work submitted, while sufficient in quantity, suffers from sufficient defects to show a lack of adequate understanding or ability to apply results.

30–39 The candidate, while attempting a significant part of the mini-project, has displayed a very limited knowledge or understanding at the level required.

0–29   The candidate has either attempted only a fragment of the mini-project or has shown an   inadequate grasp of basic material.

Dissertations Qualitative Descriptors

Dissertations will be assessed with reference to the following qualitative descriptors.

90–100 Work of potentially publishable standard, as evidenced by originality or insight. The work should show depth and accuracy, and should have a clear focus. It is likely to go beyond the normal MSc level. The standard one sees in winners of one of the examination prizes.

80–89 Work in this range will be at the level of a strong candidate for a DPhil applicant. The project will be an easy choice as a winner of a college essay prize. It will have depth, accuracy and a clear focus. It will show a strong command of material at least at the MSc level. It is likely to contain original material, which may take the form of new mathematical propositions, new examples, or new calculations, for example.

70–79 The work submitted is of a generally high order, with depth, clarity and accuracy, but may have minor errors in content and/or deficiencies in presentation. It may contain original material, at least in the sense of new examples or calculations.

60–69 The candidate shows a good grasp of their subject, but without the command and clarity required for first class marks. Presentation, referencing and bibliography should be good, and the mathematics should have no more than minor errors.

50–59 The work shows an adequate grasp of the subject, but is likely to be marred by having material at too low a level, by serious or frequent errors, a high proportion of indiscriminate information, or poor presentation and references 

40–49 The candidate shows reasonable understanding of parts of the basic material, but reveals an inadequate competence with others. The material may be at too low a level. There are likely to be high levels of error or irrelevance, muddled or superficial ideas, or very poor writing style.

30–39 The candidate shows some limited grasp of at least part of the material.

0–29   Little evidence of understanding of the topic. The work is likely to show major misunderstanding and confusion.

13. Classification Criteria

To determine the final USM, F, the dissertation is given the weight of three written examinations or mini projects; the final USM, F, is calculated as

F = [(X + Y + A + B + C + 3D)/8]

where X, Y are the best two marks on Schedule II courses, A, B, C are the three highest other marks on written examinations or mini projects, and D is the dissertation mark.

Passes, merits and distinctions, are determined by the following rules:

(i) If any of X, Y, A, B, C, D is less than 50, the candidate is failed.

(ii) If either

(a)     D ≥ 70, and X, Y, A, B, C ≥ 70
or

(b)     D ≥ 80, X ≥ 70, Y ≥ 67 and (X + Y + A + B + C)/5 ≥ 70,

the candidate is awarded a distinction.

(iii) If either

(a)     D ≥ 65, and X, Y, A, B, C ≥ 65
or

(b)     D ≥ 75, X ≥ 65, Y ≥ 63 and (X + Y + A + B + C)/5 ≥ 65,

the candidate is awarded a merit.

(iv) In all other cases the candidate is awarded a pass.

[Note: Conditions (ii)(b) and (iii)(b) permit the examiners to interpret the requirement ‘excellence through­out the examination’ more broadly, to award a distinction or merit for particular excellence on the disser­tation where the mini projects are not uniformly of distinction or merit standard.]

14. 6. Consideration of Mitigating Circumstances

A candidate’s final outcome will first be considered using the classification rules/final outcome rules as described above in section 4. The exam board will then consider any further information they have on individual circumstances. 

Further information on the procedure is provided in the Examination and Assessment Framework, Annex E and information for students is provided at www.ox.ac.uk/students/academic/exams/problems-completing-your-assessment  

Further information on how to make an application for consideration of mitigating circumstances in an examination is available at http://www.ox.ac.uk/students/academic/exams/guidance.

15. Resits

A candidate who fails the MSc may be admitted to and examined on the course as offered in one of the two subsequent years. This resit attempt shall normally be taken at the next opportunity, but may be deferred once.

In such a case the examiners will specify at the time of failure which of the assessed components of the course may or must be redone. In such a case a student will not be eligible for a merit or distinction on the whole course. Where a resit is required after technical failure of an option (i.e. non-submission of the work), this module mark will be capped at 50.

No piece of written work shall be submitted for examination on more than one occasion. It is University policy that candidates who have initially failed an MSc are not normally eligible for the award of merit or distinction.