TMUA PAST PAPERS

Download the specimen papers and every past sitting from the UAT-UK preparation pages. They are free, they include the official answer keys, and they are the only questions written by the people who write the test.

Do not substitute A-level papers, MAT questions or generic problem sets for them at the start. The TMUA has a distinctive question style, particularly on Paper 2, and mixing sources early makes it harder to learn that style.

Third party question banks have a place later, once the official supply is running low and you need volume on a specific weakness. They are a supplement, never the spine.

The Galleria holds the written work behind a 9.0, given away in full. Nothing in it is gated, and no email address is required to open anything.

Model answers: Specimen Paper 1, Specimen Paper 2, 2017 Paper 1, 2017 Paper 2, 2022 Paper 1 and 2022 Paper 2, each worked in the intended route rather than the first route that occurs.

Live solved sittings: 2017 Paper 1, 2018 Paper 1, 2018 Paper 2 and 2020 Paper 1, recorded as the paper was actually worked, including the abandoned lines. These are the documents to read if you want to see what a fast solution looks like before it is tidied up.

Written analyses: Specimen Paper 1 and 2017 Paper 1, question by question, classifying what each question is really testing and where the marks leak.

The Blueprint: the method itself, how a paper is sat, marked, classified and converted into rules. The Recipe and Espresso documents compress the same material for the last fortnight.

Sit it in one block, 75 minutes, no calculator, phone in another room, and write the clock time next to each question number as you finish it. The timing record is half the value of the sitting.

Mark it blind against the official key. Do not read a single solution yet.

Classify every wrong or slow question as knowledge, method, execution or time. Knowledge means the fact was missing. Method means the route was longer than the intended one. Execution means a slip. Time means it was never reached.

Only now read the model answer, and read it for the route rather than the answer. Write one line for each question in the form when I see X, do Y.

File those lines in a single running document. Before the next paper, read the document. This is the entire mechanism, and it is what separates a candidate who has done ten papers from a candidate who has learned from ten papers.

Weeks twelve to nine: one paper as a cold diagnostic, then content repair on whatever the diagnostic exposed. No further full papers in this block.

Weeks eight to four: one timed paper per week, fully classified, alternating Paper 1 and Paper 2 so the weaker paper gets the majority of attention.

Weeks three to one: full pairs, both papers back to back at the time of day your centre will sit you, with a fifteen minute gap.

Final week: no new papers. Re-read your own error log and re-do only the questions you got wrong the first time.

If you run out of official papers before week three, revisit the earliest ones. A paper sat twelve weeks earlier is close to cold again, and the second attempt tells you whether the rules you wrote actually stuck.

Reading a paper before sitting it. Once seen, it is a worked example rather than a test, and you have lost your only honest measurement of that paper.

Sitting papers untimed. Timing is not a detail on a test where the intended solutions are short and the obvious solutions are not.

Marking and moving on. A score with no classification produces the feeling of progress and none of the substance.

Neglecting Paper 2 because Paper 1 is more satisfying to practise. The overall score is the average of the two, so a point gained on the weaker paper is worth half a point overall, and it is far cheaper to find.