You spent weeks building a practice history, and it gave you a rough expectation. Then the certificate arrived and the expectation did not survive it. The result sits noticeably below what your papers had been producing, or — less discussed but just as disorienting — comfortably above. Either way the first instinct is the same: decide which of the two numbers was the real one, and treat the other as a mistake. That instinct is the problem. Neither number is fake, and neither is a fraud exposed by the other. They are different measurements, taken under different conditions, on different scales, and the gap between them is worth understanding before it drives a decision you cannot easily undo.
A practice score and a dMAT score are not the same measurement
Start with the arithmetic, because a surprising share of the gap lives there rather than in your ability.
The official result is not a count of your correct answers. As our guide to the dMAT score, percentile rank and certificate sets out, the reported figures are standardised, and the percentile rank places you against the candidates who actually sat the test. A practice platform's number is almost never that. It is usually raw accuracy — how many you got right — or a percentile against whoever happened to use that platform. Those are not the same quantity, and subtracting one from the other produces a difference that looks meaningful and is not.
The cohort matters as much as the scale. People taking practice tests on a preparation site are a self-selected group: they are mid-preparation, sitting at different stages, with different amounts of work behind them. The exam cohort is everyone who registered and turned up. A percentile computed against the first group tells you very little about where you stand in the second, which is exactly why a responsible platform labels its own percentile as an internal estimate rather than a forecast of your certificate.
So before reading anything into the gap, ask what the two numbers were even counting. If your practice history was raw percentages and your certificate reports a standardised score and a percentile rank, a large part of the discrepancy is a units problem, not a performance problem.
Conditions explain most of the rest
Whatever remains after the scales are accounted for is usually about circumstance, not capacity.
Practice scores are generated in the conditions you allowed. A paper sat at home, on a familiar interface, with a pause for tea, a phone within reach, rough paper you would not have on the day, and no travel beforehand, is a materially easier paper than the one you sat in the exam room — even with identical questions. Our note on why mock scores keep swinging makes the point for practice-to-practice variation, and it applies with more force here: if your practice was looser than the exam, the official number is the more honest measurement of the two, and the practice average was quietly flattering you.
The comparison runs the other way too. If you rehearsed under genuinely strict conditions — full length, no pauses, on a screen — and the official result came in higher, that is not a fluke to be embarrassed about. It is what disciplined rehearsal is supposed to do on the day.
Sample size deserves one sentence of its own. A handful of papers is a small sample, and small samples have wide spreads. If you had been quietly treating your best paper as your true level, you were comparing one exam against the top of a band rather than against its middle. The width of that band was always the more informative number.
Which direction the gap runs changes what to do
The reasoning is different depending on which way you were surprised.
If the result came in below your practice history, resist the leap straight to a retake. Do the diagnosis first. Look at the band rather than your best paper, ask honestly whether your practice conditions resembled the exam, and read the certificate's own structure before anything else: how to read your two dMAT module scores explains what a lopsided result points at, and a clear split is far more actionable than a total that merely disappointed you. Your error log and a careful pass over your last few papers will usually tell you more than the certificate does. Only then is the retake question a real question, and how to decide whether to retake the dMAT works through it properly — including the point that the certificate does not expire, so nothing forces the decision this week.
If the result came in above your practice history, enjoy it, then be careful with the conclusion you draw. The temptation is to decide retrospectively that the preparation was excessive. It is equally likely that it worked. If you sit again in a later cycle for any reason, or if a programme asks you to demonstrate something else, the habits are the asset — not the assumption that you can reproduce the result on less.
In both directions, one practical fact settles the hierarchy between the two numbers. The official result is the one that travels: it is what you download, what you submit, and what recipients read, and the official information states that universities decide individually how they use dMAT results. Your practice scores go nowhere. They were never evidence for anyone else — they were instrumentation for you, and instrumentation is still useful after the measurement it was calibrating has been taken. If you are rebuilding a practice baseline under honest conditions, the free dMAT diagnostic is a reasonable place to start it.
The gap between your mocks and your certificate is not a verdict on which number was real. It is information about the distance between the conditions you practised in and the conditions you were tested in — and that distance is the most fixable thing in the whole comparison. Use the official result for decisions, use the practice history for diagnosis, and check anything time-sensitive against the official sources for your cycle rather than against a number you half-remember from a good afternoon of practice.
