The thinking problem, and sitting the paper

The thinking move: the paper pays for shapes, not for knowledge

Three hours, one hundred marks, every one of them written prose. Four definitions at seven marks each, six short answers at ten, and one applied thinking problem at twelve. Eighty-eight of those hundred marks are long-form. The consequence is uncomfortable and worth accepting early: your knowledge is not what is being marked — your ability to deploy it in a particular shape, under time pressure, is. Two candidates who know exactly the same amount will differ by twenty marks on the strength of shape alone.

So start with the clock. One hundred marks in one hundred and eighty minutes is a little under two minutes a mark, and you must reserve fifteen to twenty minutes for reading the paper and checking at the end. That gives roughly eleven minutes for a seven-mark definition, sixteen for a ten-mark short answer, and twenty for the thinking problem. Those numbers are not advisory. The commonest way to lose a grade band on this paper is to write a beautiful fourteen-minute answer to a seven-mark question and then have eleven minutes for a ten-mark one.

The three shapes

A seven-mark definition has five beats. Name the category — what kind of thing is this, a measure, an activity, a notation, an umbrella activity? Then the definition with its load-bearing qualifiers intact. Then place it: what it relates to, what it is part of, what it is assessed by. Then elaborate the internal structure — the categories, levels, steps or components. Then a concrete example, worked far enough that the concept is visibly doing something. The exam description asks for examples in as many words, so this is not optional: a definition with no example caps around five out of seven however fluent it is.

A ten-mark short answer is almost always a compare, and it has four beats. Define both terms crisply — that is about four marks, two each. Say how they are similar, and mean it; the paper's own sample question asks this explicitly and it is the beat candidates skip. Say how they are fundamentally different, on named axes rather than on size or sophistication. Then give one example where they diverge — a case in which one holds and the other does not. The dominant failure is answering half the question: defining both and never comparing, or comparing and never defining.

A twelve-mark thinking problem has four parts, and they are unlike anything else on the paper. About two marks for framing the project — naming the attributes that make the decision what it is. About four for applying named course concepts precisely, with their definitions carried in passing. About four for the argument: why this, why not the alternatives, and what it costs you. And about two for the concession — the residual risk, the limitation, or the condition under which you would decide differently. The failure mode is a chapter summary with the project's name sprinkled through it. The marks are for decisions traceable to project attributes.

The two sentences almost nobody writes

The first is the concession. "This leaves X uncovered, and here is what I would do about it" costs one sentence and reliably carries the last two marks of a twelve-mark answer, because it is the sentence that proves you were making a decision rather than reciting a recommendation. Write it before you run out of time, not after.

The second is the divergence example in a compare question: the case where one concept holds and the other fails. It is what turns "these are different" into evidence that you understand how they are different. On this paper it is worth one to two marks every single time.

Worked example — the same fact, worth three marks or eight

Here is one BorrowBox fact: we built a locker simulator in the first increment. Written three ways.

Version one, no marks. "We built a simulator early because we thought the vendor might be late." True, and worth nothing. It names no concept, no artefact, no number.

Version two, about two marks. "The locker SDK was a technical risk, so we mitigated it by building a simulator." Better: two named concepts, correctly used. But it still describes what happened rather than why it was the right call, and any project could have said it.

Version three, about five marks and rising. "The vendor SDK was rated 60 per cent probability and impact 1 in week one, because it sits on the critical path — the driver, wire-up and field-trial chain has zero slack — giving a risk exposure of about 22,800 dollars. Mitigation was two developer-weeks building a simulator against the published wire protocol, which is cheap against that exposure, so we built it; monitoring was the vendor's forum cadence, reviewed every second stand-up; management was a pre-agreed trigger at the end of week 6 to ship with a manual librarian handover. What it leaves uncovered is the value proposition rather than the date: a launch without lockers is a worse product, so the trigger buys us a decision, not an outcome."

The third version is not longer because it contains more facts. It is longer because every clause is doing one of the four jobs: framing, applying a named concept, arguing from the project's own structure, and conceding a limit. That is the whole technique, and you can practise it on any fact you already know.