The examiner never sees your model
This is the fact that reorganises everything else about A3, and it is stated plainly in the official manual: your finished composition is placed in a shooting kit and photographed in three views, and those photographs are what gets uploaded. You will never meet the person marking it. They will never meet your model. Everything you built has to survive the trip through a camera.
Three views, and none of them is behind you
The views are front, oblique-left and oblique-right, with the barcode visible. That is the whole record of your work.
Start with what that removes. Nothing is photographed from behind, so a form hidden at the back may as well not exist. Nothing is photographed from above, so a composition whose logic is a plan — a clever arrangement you would only read looking down — will not read at all. There is no fourth angle to rescue you.
And start with what it adds. You know the angles in advance. That is unusual and it is a gift. You cannot know the forms, but you can know exactly where the camera will stand, which means you can build for it. A candidate who has practised photographing their attempts from those three positions is preparing for the actual assessment. A candidate who has only ever looked at their models on the desk is preparing for an assessment that does not happen.
What a camera takes away
A photograph flattens. That is the whole problem, and it is worth being precise about what gets lost.
Depth collapses. Two forms 6 cm apart, one behind the other, can photograph as though they are touching. Overlaps that read clearly in the hand become ambiguous tangles head-on. This is why the frontal view is the one that kills compositions: it is the angle with the least depth information.
Silhouette becomes everything. When depth flattens, the outline does the work. A composition with a strong, legible silhouette photographs well from any angle. One that relies on interior relationships you can only appreciate by moving your head does not survive.
Small gaps vanish. A 3mm slot you were proud of will not be visible. If a void is part of your composition — and it should be — it has to be big enough to read at photographic scale.
The fix for all three is the same and it is unglamorous: check the flat view. Close one eye, or better, actually take the photograph. If it reads there, it reads everywhere.
Build for the obliques, survive the front
Two of your three views are oblique, which tells you where the composition should live.
The obliques are generous. They restore depth, reveal overlaps, and show the composition as an object in space — this is where a well-composed model looks best. So the obliques are where you should be *ambitious*: the diagonal, the layered depth, the interesting relationship between a near mass and a far one.
The front is the test. It is the flattest, meanest angle, and it is where a composition either holds or falls apart. So the front is where you should be *defensive*: check that the silhouette reads, that nothing important is hiding directly behind something else, and that the whole thing does not collapse into a single confused blob.
The working sequence, then: compose for the obliques, then step round to the front and ask whether it survives. If it does not, the usual fix is small — rotate the whole composition a few degrees so that what was hidden becomes visible, and what was a tangle becomes an overlap you can read.
The rules behind this
Sourced to the official brochure rather than restated here, so there is one place to correct when the Council revises it.
The A3 foam kit is supplied at the centre. You do not bring one, and you cannot know its contents in advance.
The question paper carries a KIT CODE (D1, D2, D3...) and the invigilator issues the matching kit: a set of 3D forms made of foam plus a 15cm x 20cm foam base sheet. Glue is provided. The finished model is photographed at the centre in three views and uploaded.
Read this carefully: Consequence worth stating plainly: A3 cannot be rehearsed as an artefact, only as a capability. Anyone selling 'the NATA foam kit' is selling a proxy.
Source · verified 2026-07-16
Part A is three questions: A1 Composition and Colour (25), A2 Sketching and Composition in Black and White (25), A3 3D Composition (30).
Eighty marks across three questions means each is worth roughly 13% of the whole exam. There is no salvaging a bad question by volume, the way a 50-question paper allows.
Source · verified 2026-07-16
What almost everyone believes
“I should build the best model I can and it will be judged on how good it is.”
It will be judged on how good three photographs of it are — front, oblique-left, oblique-right. Those are not the same thing.
A photograph flattens depth, collapses overlaps, hides anything at the back and erases small gaps. Compositions that reward walking around them — the ones that feel best in the hand — routinely collapse in a flat frontal view, and the frontal view is one of the three the examiner gets. The candidates who lose marks here are frequently the ones whose models are genuinely good, which is what makes it worth knowing in advance. You cannot know the kit, but you know exactly where the camera stands.
Depending on how long you have
Foundation
Understand the skill. Months out, or starting from zero.
Get into the habit of photographing everything you build, from the beginning. Not to keep it — to see it. A photograph is a brutally honest second opinion, and learning to predict what it will show you is a skill that takes months and pays off in one morning.
Drill
The practice protocol. What to repeat, how often, how to score it.
Every practice attempt gets three photographs from the exam angles, and you judge from the photographs only — never from the object on the desk. Mark yourself on the frontal view specifically, because that is where the failures hide. The gap between how good your model feels and how good it photographs is the number you are trying to close.
Exam-Day
What to actually do under the constraint — 108 seconds, no instruments, one pass.
Before you finish, step round to the front of the base and look at it flat, from eye level. Ask one question: does the silhouette read, or is this a blob? If something important is hidden, rotate the composition a few degrees rather than rebuilding — it is usually enough, and you do not have time for anything else.
Try it
Twenty minutes, and you will need a phone. The point is the gap between the object and its record.
- 01Build any composition on your 15 x 20 base. Take as long as you like — quality is not the variable here.
- 02Look at it on the desk and rate it out of ten. Write the number down before you go further.
- 03Photograph it from directly in front at eye level, then 45 degrees left, then 45 degrees right.
- 04Look only at the three photographs and rate it again. Write that number down too.
- 05The gap between the two numbers is what you are training away. If the frontal view is much worse than the others, that is the specific thing to fix — usually by rotating a few degrees, not rebuilding.
The short version
You are marked on three photographs, not on a model: front, oblique-left, oblique-right, and nothing from behind or above. The camera flattens depth, kills small gaps and promotes silhouette. Compose ambitiously for the obliques, then check defensively against the front — that flat, mean view is where compositions fail, and a few degrees of rotation usually rescues it. You cannot know your kit; you know exactly where the camera will stand.
Next: doing all of this in thirty minutes, which is the constraint that makes the rest of the module real.
Questions people actually ask
- How is the NATA A3 model assessed?
- The finished model is placed in a shooting kit at the centre and photographed in three views — front, oblique-left and oblique-right — with the barcode visible, and the photographs are uploaded. You do not photograph it yourself and you do not carry it out. The photographs are the entire record of your work.
- Which view of my A3 composition matters most?
- The front is the one to defend, because it carries the least depth information and is where compositions collapse into an unreadable blob. The obliques are more forgiving and show the composition at its best. Compose for the obliques, then check the front survives.
- Does anything get photographed from above or behind?
- No. The three views are front and the two obliques. A form hidden at the back contributes nothing, and a composition whose logic only reads in plan will not read at all.