TMUA Preparation: Join the Next Batch

The TMUA is now the test that decides a large part of your maths, computer science and economics application to Cambridge, Imperial, LSE, UCL, Warwick and Durham. Oxford has also retired the MAT in its favour.

Our TMUA course opens for the next cycle shortly, and this is what it is built around.

The part almost nobody prepares for

The test is two papers, each 20 multiple-choice questions in 75 minutes, with no calculator and no dictionary.

Paper 1 applies AS-level pure mathematics to unfamiliar problems. Candidates prepare for this instinctively, and most arrive competent.

Paper 2 adds elementary logic and proof, and this is where results are decided. Necessary and sufficient conditions, a statement versus its converse, negating a quantified claim, disproof by counterexample, and the question that hands you a plausible argument and asks which step fails.

No school syllabus teaches this. It is also a small body of content, learnable in a few weeks, which makes it the highest return per hour anywhere in the test. That is the asymmetry the course is built on.

Two facts that change how you sit it

There is no pass mark. LSE states this outright. Any service quoting you a guaranteed cut-off is inventing it. Your score is weighed with your grades, statement and interview.

There is no negative marking. A blank can only cost you. So in the last four minutes, guessing three unattempted questions beats re-checking one you have already answered. Most candidates instinctively do the opposite.

Why multiple choice changes the maths

Every question carries its options, and those are information. The full algebraic derivation is frequently the slowest route available.

We teach substitution of convenient values, boundary and degenerate cases, elimination by sign or magnitude, and working backwards from the answers. Candidates who solve everything properly tend to stall around question fifteen, with correct work on the page and five questions untouched.

Which sitting you must take

Sittings fall in October and January, and the choice is dictated by your universities rather than your readiness.

Cambridge and Oxford applicants must sit in October — the mid-October UCAS deadline means a January score lands too late. Warwick, Durham, LSE and Imperial generally accept January, which buys a longer run-up.

For 2027 entry the October window was 12–16 October 2026; the January window is 4–8 January 2027, with booking opening 26 October 2026. Confirm every date on the official UAT-UK site, because they move each cycle.

How the course runs

Sixteen weeks at about four hours a week, online, in groups capped at eight — the largest size at which every student's timed paper can actually be reviewed individually.

The first third closes Paper 1 gaps found by diagnostic rather than assumption. The middle third is logic and proof. The final third is almost entirely full timed papers, reviewed by error type: careless, conceptual or pacing, because the remedy for each is completely different and an unclassified error repeats for months.

What the logic questions actually look like

Abstract descriptions of "logic and proof" help nobody, so here is the shape of it.

A typical Paper 2 item gives you a statement such as "if n squared is even then n is even" and asks which of five related claims must also be true. The work is recognising that the converse, the inverse and the contrapositive are three different statements, only one of which is logically equivalent to the original.

Another standard type presents a four-line proof and asks which line first fails. The error is rarely arithmetic. It is usually a step that assumes what was meant to be proved, divides by something that could be zero, or quietly swaps a sufficient condition for a necessary one.

A third type asks you to disprove a general claim. Candidates write paragraphs; the answer is one well-chosen number. Finding that number quickly is a trained reflex, and it is worth full marks.

None of this requires mathematics beyond AS level. It requires precision about what a statement claims, which is exactly why a few focused weeks move the needle so far.

The mistakes that cost the most marks

Treating every question as worth the same time. They score the same but cost wildly different amounts. The discipline is abandoning a question that is not yielding within about ninety seconds and returning only if time allows.

Deriving instead of eliminating. Options are information. Often two can be dismissed by sign or magnitude in seconds, turning a hard question into a coin flip at worst.

Leaving blanks. With no negative marking this is pure loss, yet candidates routinely run out of time with unanswered questions they never even guessed at.

Preparing only Paper 1 content. The most expensive mistake of all, because it feels like productive work right up until the paper that decides your score.

Who this course suits, and who it does not

It suits a student with AS-level pure mathematics reasonably secure who is applying to a university that requires or recommends the TMUA, and who wants the logic content taught properly rather than mentioned.

It does not suit a student whose algebra and trigonometry are still shaky. The test applies that material to unfamiliar problems, so gaps underneath get exposed rather than hidden, and the right first step is closing them. We will say so at the diagnostic rather than taking the enrolment.

Joining the next batch

Enrolment for the January sitting cohort is open now, and the October 2027 cohort opens in spring. Batches are capped at eight, so they close in order of enquiry.

If you want evidence before committing, we will mark one diagnostic paper and talk through where your marks actually went. That conversation tells you more than the score does.