Policies and support

No late work, and AI use declared rather than hidden. The course-wide rules are stated here once, so an assessment page can point at one of them rather than repeat it.

What each piece of assessed work asks for, and how it is marked, stays on its own page. What is on this page applies across all of them.

Late work and extensions

No late work is accepted. The only route past a deadline is an extension, applied for through the University’s documented process before that deadline passes.

There is no per-day deduction and no grace period, because there is no late submission for one to apply to. That reads as severe until you look at what is being handed in. Only two of the five assessed items have an ordinary deadline that a later date would move, and here is what the rule means for each.

The labs. A lab is prepared before the session and finished inside it, so once the session ends there is no version of it left to hand in. The remedy is built into the mark instead: the best ten of your eleven lab marks count, and the one that falls away is there so that a single missed week costs you nothing. It is not a lab you are invited to skip: the dropped mark covers one missed week, and a second one costs you.

The race. An event, in a room, on a draw the whole cohort shares — the stochastic events drawn once before anybody races. There is no artefact to hand in late, so an extension here means an alternate arrangement, not a later date.

The instrument. Its cutoff is the race rather than a penalty scale: the race runs the instrument as it was submitted, unchanged after its deadline. An extension that pushed the instrument past the race would leave the race with nothing to mark.

Assignment 1. It is submitted in two stages and they behave differently. Stage one — the page fixing what evidence would make you retain the bought recommendation unchanged and what would make you reject it outright — has no late window at all, because the race data unlocks on submission: being late is not possible, and being unsubmitted is. Stage two, the audit itself, has an ordinary deadline, and an extension applies to it in the ordinary way.

The defence. One document, written after the result is known and marked against the calls it defends. This is the item where an ordinary extension is straightforward.

Extensions are granted through the University’s process, on the grounds that process recognises. The course does not operate a second one beside it.

Academic integrity and AI use

AI use is permitted. Undeclared AI use is the breach.

That is this course’s own doctrine turned around and pointed at the student. An instrument — your own race simulator — is submitted with a declared domain of validity, a document naming the conditions under which it should not be believed. A submission is handed in with a statement of what wrote it. Both arrive carrying an account of their own limits, and it is the same move twice.

A declaration is a few lines submitted with the work: what you used, where you used it, and what you did with what came back. It costs you nothing. No mark in this course is reduced because a declaration names a model, and there is no version of your submission that scores higher for having said nothing.

The breach is submitting someone else’s work as your own. A model’s counts and so does a classmate’s, and the declaration is what separates work you can account for from work you cannot. Suspected breaches are handled under the University’s academic integrity rules, which apply here as they do anywhere else. The course does not run a procedure of its own alongside them.

Collaboration

Discuss freely. Submit your own work.

The labs are individual work done in a shared room, which is the arrangement most often misread. Working through how tyre wear behaves with the person beside you is the room doing what it is for. Handing them your implementation is not. Name who you worked with, and cite code you did not write — including code a model wrote for you, which is the declaration above doing its job in a smaller place.

The race is individual too, and the draw belongs to the cohort rather than to you. The safety car — the period where the field is neutralised behind a lead car and nobody may overtake — comes out on the same lap for every student, and the rain arrives at the same minute. There is nothing in the draw to leak and nothing to be gained by comparing draws, because the only thing that differs between two students is what they did with the same events.

Support, adjustments and the escalation path

Ask the tutor leading that week first. All three tutors staff every lab, and each leads a block: Tobias Wren weeks 1 to 4, Naila Sorge weeks 5 to 9, and Kwame Ferriter weeks 10 to 12.

Course-level questions, and anything about assessment, go to the convenor, Rafaela Dunne. If her answer does not resolve it, the matter leaves the course: it goes to the School of Invented Disciplines, and past that through the University’s own complaints and appeals process. The course has no further stage of its own.

Adjustments, against an environment that does not vary. The pit wall room is deliberately identical for every student — the same screens, the same seat, the same setup — and that is what makes two students’ calls comparable at all. A student with an adjustment in place will ask what that means for them, and the answer is a distinction rather than an exception. The environment is held constant. Access to it is not.

Adjustments are arranged through the University’s process, and this course accommodates what that process grants without varying what is being assessed: the same race, the same draw, and the same question about whether each call agreed with the instrument that recommended it. The convenor arranges what an adjustment requires in the room.

Code of conduct

The room. The pit wall is shared equipment and its environment is not to be modified. A machine left changed is a change to the next student’s assessment.

The race session. You do not interfere with another student’s run. The car ahead is another student’s and they are deciding about you while you are deciding about them — that is the race. Reaching past your own screens is not.

Discussion. The course’s doctrine is that disagreement is a bug report, and it holds between people as it does between a student and a model: a claim gets tested, not its author.

The University’s own code of conduct covers everything general, and it applies in this room as it does in any other.