Policies and support

The rules of the course, most of which are rules about numbers, and where to go when something is in the way.

Most of these are the ordinary rules of any course. Two of them are peculiar to this one, and they come first.

Do not play the specimen before week ten

You get the rules of Close Quarters in week one. The game itself is linked from the home page from the first day, on purpose, and you are asked not to open it before the week ten tutorial.

Nobody can enforce this and nobody will try. It is here because the course only works once: if you play first, you spend nine weeks defending your first impression, and the week ten tutorial has nothing left to show you. If you have already played it, say so on your first quiz. That is useful data and it costs you nothing.

Every number carries its papers

A number in this course is reported with its unit, the number of runs behind it, the seeds or the command that reproduces it, and the machine and program it ran in, because the same seeds can end differently in Node and in a browser, or on another computer (week 2). The working run count is 51. This applies to assessed work, to the quizzes, and to what you say out loud in a tutorial.

The course’s seeds are 97, 194, 291 and on, each 97 past the last. Tutorials count them, so “seeds 1 to 51” means 97 to 4947, and week 2’s seed 97 is seed 1. “Ordinary card choices” means the player takes the leftmost of the three cards it is offered.

Three things are treated as not having been submitted at all:

This is the only rule in the course that is about integrity and method at the same time.

Late work and extensions

Extensions are easy to get and are arranged before the deadline, by email to Tamsin. You do not need to explain much. After the deadline there is nothing to extend.

Tutorial quizzes are sat in the room at the end of the tutorial they belong to. There is no make-up quiz; the two lowest of the twelve are dropped instead, so two missed weeks cost nothing.

Your own work, and help from others

Look at each other’s screens. Argue about each other’s numbers. That is what the tutorials are for. Your fork, your quiz answers and your defence are your own.

You may use an AI coding tool to write policy functions and measurement scripts, the same as any other tool, and you should say that you did. Two limits: you must be able to explain every line of your simulated player when asked, and a number is never evidence because a tool said it. It is evidence because the command that produces it runs on someone else’s machine.

Inventing a measurement, or editing one to make a change look better, is academic misconduct and will be treated as such.

If something about the course is a barrier

Tell Tamsin early, in whatever detail you are comfortable with.

Two things are already in place. The final defence is a recording, so it can be made in as many takes as you need. And the week ten tutorial has an observer’s version: if ten minutes of fast play is not possible or not pleasant for you, you sit beside someone who is playing and write the list from what you see. It produces the same kind of findings and is marked the same way.

Who to ask

The person next to you in the tutorial, then Wei, then Tamsin. For anything that should not wait, email either of them and put the course code in the subject line.