The January sitting of the TMUA runs from 4 to 8 January 2027, and booking closes on 21 December. That leaves about seven weeks. Most candidates will spend them working through past papers and wondering why their scores are not moving.
Who the January sitting is for
There are only two sittings for 2027 entry, and the first has gone. January suits three groups: applicants who did not book in October and have since decided on a course that requires the test; applicants adding a university such as Imperial, LSE or Warwick to a list that did not need it; and gap-year or late-deciding candidates.
It is not available to everyone. A candidate may sit the test once per admissions cycle, so anyone who took it in October cannot take it again. Standard applicants to Oxford and Cambridge are required to sit in October, which means January is not an option for them either.
Which courses require the test has changed considerably for 2027 entry, with both Oxford and Cambridge extending it to more of their mathematics and computer science degrees. Check the course page for each university your child is applying to rather than relying on what was true last year.
What the test is
Two papers, seventy-five minutes each, sat back to back with no break. Twenty multiple-choice questions on each paper, all carrying equal weight. No calculator, no formula booklet, and no negative marking, so there is no reason to leave a question blank.
Paper 1 is Applications of Mathematical Knowledge. Paper 2 is Mathematical Reasoning. The two are scaled together into a single score from 1.0 to 9.0.
Why strong A level candidates underperform
The content is not advanced. Almost all of it sits within AS-level pure mathematics, and a surprising amount of Paper 2 reaches back to GCSE number work: primes, factors, highest common factors and lowest common multiples. A pupil predicted A*A*A has been taught every topic on the paper.
That is precisely why the test catches them out. An A level question names its method. It tells you to differentiate, or to use the binomial expansion, and the work is in executing a procedure you have been drilled in. A TMUA question tells you what is true and leaves you to find the route. The mathematics is familiar; the demand is not.
So the pupil who has never been genuinely stuck in a maths lesson sits down, meets a question that does not announce itself, and discovers that the skill being tested is one nobody taught them.
Paper 1: the clock and the missing calculator
Seventy-five minutes for twenty questions is under four minutes each, and some of those questions take a minute while others take six. Without a calculator, every awkward surd, every fraction, every piece of arithmetic has to be done by hand and done quickly.
Candidates who are accurate but slow lose marks they could have had. The fix is not more problems; it is fluency so automatic that the arithmetic costs no attention at all, leaving the whole of it for the question.
Paper 2: the half nobody has studied
Roughly half of Paper 2 is logic and proof, and this is the part of the TMUA that is genuinely unfamiliar. It asks candidates to negate a statement containing “for all” or “there exists”, to tell a necessary condition from a sufficient one, to distinguish a converse from a contrapositive, to find a counterexample, and to follow or complete a proof by contradiction.
A level mentions proof. It does not teach this, and it does not test it this way. Most candidates meet it for the first time in a past paper, decide it is strange, and go back to the algebra they are comfortable with.
That is a mistake, because it is the most learnable part of the entire test. The material is finite, it takes a couple of weeks of focused work, and the gains are immediate. A candidate who has done that work is answering ten questions that their rivals are guessing at.
What the score means
There is no pass mark and no grade boundaries. The score is scaled against the cohort, with the median set at 4.5 and the ninetieth percentile at 7.0. Universities read it alongside the rest of the application rather than as a threshold, and published figures are scarce: Warwick has said that most of its 2025-entry mathematics offers went to applicants scoring 5.0 or above, while noting that some below that still received offers.
What this means in practice is that the test is designed to spread out candidates who all have A* predictions. It is not trying to find out whether your child can do A level mathematics. It already assumes they can.
What seven weeks should look like
The first fortnight: logic and proof. Learn it properly as a topic, not as something encountered in passing while doing past papers. This is where the quickest gains are.
The middle weeks: timed sections, not whole papers. Ten questions in thirty-seven minutes, marked and reviewed properly. Full papers are exhausting and teach less, because the review is what produces improvement and nobody reviews forty questions carefully.
The last fortnight: whole papers to time. By then the point is stamina and decision-making under pressure: knowing when to abandon a question, when to eliminate options and guess, and how to pace two papers sat back to back.
And book before 21 December. Registration is the one part of this that cannot be fixed afterwards.
Where we can help
Mr Armitage led admissions test preparation as Head of Mathematics at University College School, and our tutors prepare pupils for the TMUA, ESAT, STEP and TARA. We build programmes around the time available and the courses a pupil is applying to.