TMUA Score Conversion: Raw Marks to the 1.0-9.0 Scale

This guide does one job: it explains how raw TMUA marks become a scaled score, with worked examples and an honest conversion estimate. For the marking rules themselves, read TMUA scoring explained; for why no official boundary table exists, read TMUA grade boundaries.

You sit two papers, you get two whole numbers out of 20, and a few weeks later a screen shows you something like 6.3. The step in the middle, raw marks turning into a scaled score, is the part nobody explains properly, partly because the exam board keeps the exact mapping to itself. Here is what the conversion actually does, what we genuinely know about it, and how to turn a practice-paper raw mark into a realistic score estimate right now.

The raw marks: what gets counted

The input to the conversion is simple. Each paper has 20 multiple-choice questions, each worth exactly one mark. A correct answer scores $+1$; a wrong or blank answer scores $0$. There is no negative marking and no partial credit, so your raw result is just two whole numbers: a mark out of 20 for Paper 1 and a mark out of 20 for Paper 2, a maximum of 40 across the test.

Those two numbers are what the conversion machinery receives. Everything else about your sitting, which questions you got right, how long you took, how elegant your working was on the rough paper, is invisible to it. Two candidates with the same raw mark on the same paper get the same scaled score, always.

The output: a 1.0 to 9.0 score per paper

Each paper's raw mark is converted separately onto a scale running from 1.0 to 9.0 in steps of 0.1, and your result is reported as two scaled scores, one per paper. The scale is calibrated so that a given score means the same standard of performance whichever sitting produced it: in a harder year, fewer raw marks reach any given score, and in an easier year, more. That calibration is the entire reason the conversion is not a fixed lookup table, and it is re-tuned for every sitting.

Two consequences follow immediately. First, there is no universal "14 out of 20 equals 6.0" fact, because the mapping moves with the paper. Second, any site offering you a precise official conversion chart for a modern sitting is offering something UAT-UK has never published.

What we actually know about the mapping

An estimate is only as good as its anchors, so here are the facts the public record supports:

• The typical candidate scores around 4.5. UAT-UK's own explanation of results for the October 2024 and October 2025 sittings puts the median near the middle-lower part of the scale, at about 4.5.
• About 10% of candidates score above 7.0. The same source puts 7.0+ in roughly the top tenth of the field.
• The endpoints are fixed. A blank paper maps to 1.0 and a perfect 20 out of 20 maps to 9.0, since the scale has to span the full raw range.
• The marking is +1 with no deductions, so raw marks only ever count correct answers.

That is enough to pin the shape of the curve at the bottom, the middle, the upper tier and the top. What it cannot give you is the exact score for every raw mark in between, which is precisely the part UAT-UK keeps internal.

Our estimated conversion anchors

CrackTMUA's free score calculator is built on the anchors below. They are our estimates, constructed to honour the public facts above: 10 out of 20 sits at the 4.5 median, 16 out of 20 sits near the 7.0 top-decile line, and the endpoints are fixed. Between anchors the calculator interpolates linearly and reports a range of about half a point either side, because pretending to more precision than the anchors support would be dishonest.

Read the left column as the average of your two paper marks. The real test reports each paper separately, but for a quick "where am I?" estimate a blended figure is the useful one, and it is what the calculator returns along with its range.

Worked examples

Three examples show how to use the table, and what the numbers feel like in practice.

The typical candidate. You score 10 on Paper 1 and 10 on Paper 2. Your average raw mark is 10, which sits exactly on the 4.5 anchor. That is the median TMUA performance: half the field does better, half does worse, and it is a long way from failure, whatever it feels like next to the scores people post online.

A solid, above-average sitting. You score 14 on Paper 1 and 12 on Paper 2. The average is 13, halfway between the 12 anchor (5.2) and the 14 anchor (6.0), so the estimate is about 5.6, with an honest range of roughly 5.1 to 6.1. That is comfortably above the typical candidate and inside the range several courses regard as competitive.

The top decile. You score 16 on both papers. The average of 16 sits on the 7.0 anchor, which is roughly the top 10% of the whole field. Notice what that implies: even four wrong answers per paper still leaves you among the strongest candidates in the country. Nobody needs a perfect paper.

There is one more lesson hiding in the anchor spacing. Between the 10 and 16 anchors, one extra average raw mark is worth about 0.4 to 0.5 of a band: climbing from 12 to 14 average raw lifts the estimate from 5.2 to 6.0. Since your average moves by half a mark for each extra correct answer on one paper, every two extra questions you convert across the test buys you roughly half a band through the middle of the scale. The conversion is steepest exactly where most applicants sit, which is why small accuracy gains show up so visibly in scaled scores.

The kind of mark that moves your scaled score

Those two extra questions are rarely the monsters at the end of the paper. They are the mid-difficulty questions in the middle third, the ones the typical candidate loses to a slip or a slow method and the 6.5+ candidate banks in under four minutes. Here is a real one, rated 6.0 on our model, exactly the level where scaled scores are won and lost. Attempt it before you reveal the solution:

If that took you six minutes or a lucky guess, that is not a talent problem; it is a fluency problem, and fluency is the trainable part. Converting two or three questions like this from "lost" to "banked" is the single most efficient scaled-score improvement available to most applicants.

Why the exact 2024+ mapping is not published

Before 2024, the paper-based TMUA era under Cambridge Assessment came with conversion charts for its sittings, which is why you can still find raw-to-scale tables attached to the old papers. Two things ended that. First, UAT-UK took over from the October 2024 sitting, moved the test on-screen at Pearson VUE, and has not published a conversion for any on-screen sitting. Second, the reporting scale itself was rescaled in 2024, with the same performance now landing roughly 1 to 1.5 points lower than it would have on the old papers, so the old charts stopped describing the current test anyway.

The practical rule: if you mark an old paper against its original conversion chart, treat the resulting score as a museum piece. It tells you how that number would have read years ago, not what the same raw performance means today. For a present-day read, put the raw marks through the anchors above instead.

Get your estimate, then improve it

The fastest way to use everything above is the free TMUA score calculator: enter your two raw marks from any timed paper and it returns the estimated band with its honest range, using exactly the anchors in this guide. It takes about ten seconds and gives you a defensible answer to "what would this paper have scored?"

Then remember what the conversion rewards. Because the middle of the curve is steep, the highest-value training is converting mid-difficulty questions reliably under time pressure, not occasionally cracking the hardest ones. That is a volume game: CrackTMUA has 640+ questions, every official past-paper question plus 280+ originals, each with a worked solution that names the trap and the fastest method, filterable by difficulty so you can live exactly in the band that moves your score. The free tier gives you 10 questions a day; Premium unlocks everything, including 24+ timed mocks and a predicted band fitted to your real attempts, for £49 one-time with 12 months of access. Start practising free, and let the raw marks take care of the conversion.

Frequently asked questions