Sixty questions are printed. You may attempt twenty. Each is worth 4.75 marks with no negative marking, and scores only if you select exactly the right set of options.
Two facts follow immediately, and most preparation ignores both.
Fact 1: a half-solved question is worth nothing
There is no partial credit. If three options are correct and you select two of them, you score zero — the same as if you had left it blank, and the same as if you had selected all four at random.
So the unit of progress in Part C is not "questions I have some idea about". It is questions I can close. When you practise, mark each attempt as closed or not closed, and count only the first kind.
Fact 2: no negative marking makes selection cheap, not free
Since a wrong set costs nothing, there is no reason to leave an attempt unused. But your attempts are capped at twenty, and — critically — if you answer more than twenty, only the first twenty in paper order count. An unused attempt is a wasted 4.75 marks; a badly spent one is the same.
A method for the three hours
First pass — read all sixty, answer nothing. Mark each question as closable now, maybe, or no. This costs perhaps twenty minutes and it is the highest-value twenty minutes in the paper, because it stops you from spending forty minutes on question 7 while question 52 sits unread.
Second pass — close the certain ones. Work through the closable now set. For each, verify all four options, not just the ones you think are right. Part C punishes the habit of stopping once you have found a correct option.
Third pass — convert the maybes. With the remaining time, take the maybe questions in order of how close they feel. Stop when you reach twenty closed.
If you reach twenty before time runs out, spend what is left re-verifying the ones you closed fastest. An arithmetic slip in a question you were sure about costs the same 4.75 marks as a question you never read.
How to verify all four options quickly
Most Part C options in this exam fail for a nameable reason: a converse that does not hold, a hypothesis quietly dropped, a boundary case, an unbounded example where boundedness was needed. Learning those failure patterns is faster than solving each option from scratch.
That is exactly why our previous-year questions are tagged by trap type rather than only by topic — you can drill "converse fails" or "closed and bounded is not compact here" as a category and start recognising them on sight. See the tagged PYQ bank, and the counterexample bank for the standard "this looks true but is not" examples.
Practise with the real constraint
A practice set of twenty Part C questions does not train this skill, because there is nothing to choose. Our full papers put all sixty in front of you and cap you at twenty — which is uncomfortable the first time, and that discomfort is the point.