Studio Matrx Monthly · Volume 1 · Issue 2 · July 2026
Amogh N P
 In loving memory of Amogh N P — Architect · Designer · Visionary 
NATA 2026 / Module 5Volume 1 · Issue 2 · July 2026
Part B · on screen5.3 · Logical Derivation

Probably true is a wrong answer

All the coloured drawings are on A3. Some A3 sheets are unfinished. Does it follow that some coloured drawings are unfinished? It feels like it should. It is very probably true in any real portfolio. And it does not follow — the unfinished ones might all be the sheets without colour. That gap, between what is likely and what is forced, is the entire subject of this lesson and the place the marks live.

ByAmogh N P· Architect & interior designer7 min read · verified 2026-07-16
A single set of footprints crossing wet sand into the distance in low raking light, with nobody in sight

Necessarily means necessarily

The word in the stem is doing enormous work and candidates read past it. Which conclusion follows? means: is there any arrangement of the facts, however contrived, in which the premises are true and the conclusion is false? If there is even one, the conclusion does not follow. Full stop.

So the test is not does this seem right? It is can I break it? Try to build a counterexample. If you can, delete the option — and you can usually do it in about five seconds, which is faster than reasoning forwards.

Back to the portfolio. Coloured drawings are a subset of A3 sheets. Some A3 sheets are unfinished. Can I arrange it so no coloured drawing is unfinished? Yes, trivially: put all the unfinished ones among the non-coloured A3 sheets. Counterexample found, option dead, five seconds.

The habit to build is attack the option rather than derive the answer. Deriving is slow and produces one candidate. Attacking is fast and kills three.

A counterexample drawn with circles: the unfinished sheets can all sit outside the coloured set "ALL COLOURED DRAWINGS ARE ON A3. SOME A3 SHEETS ARE UNFINISHED." DOES IT FOLLOW THAT SOME COLOURED DRAWINGS ARE UNFINISHED? A3 SHEETS COLOURED UNFINISHED the unfinished ones can ALL sit OUTSIDE the coloured set → COUNTEREXAMPLE FOUND → IT DOES NOT FOLLOW five seconds. no forward reasoning. "VERY LIKELY IN ANY REAL PORTFOLIO" IS EXACTLY WHAT THE DISTRACTOR IS MADE OF. THE GRAMMAR: "some" = at least one, POSSIBLY ALL · "all A are B" does NOT reverse · "no A are B" DOES reverse ATTACK THE OPTION, DO NOT DERIVE THE ANSWER. DRAW THE DIAGRAM THAT BREAKS IT.
Necessarily means necessarily

Some means at least one, and possibly all

Ordinary English and logical English disagree here, and the disagreement is deliberately exploited.

In conversation, some of the drawings are finished implies that some are not. In logic it implies no such thing. Some means at least one, and is perfectly satisfied if all of them are. When someone says some, you may not infer that the rest are excluded — the everyday implication is a courtesy of conversation, not a fact about the world.

Similarly: all A are B tells you nothing whatsoever about whether all B are A. All architects have drawn a plan; that does not make everyone who has drawn a plan an architect. Reversing a universal is probably the single commonest logical error there is, and the options will offer you the reversal every time.

And no A are B does reverse safely — if no cats are dogs, no dogs are cats. That one is symmetric. Knowing which relations reverse and which do not is most of this subject: all does not reverse, no does, some does.

Draw the circles

The same instruction as the last lesson, arriving for the same reason: get it out of your head.

Overlapping circles — one for each category — turn these questions from verbal reasoning into looking at a picture, and looking at a picture is something you are extremely good at. Draw the A3 sheets as a big circle. Draw coloured drawings as a smaller circle entirely inside it. Now shade some of the A3 region as unfinished, and notice you can put that shading outside the coloured circle. The counterexample is now visible rather than imagined.

This is also why the counterexample method is fast. You are not searching an abstract space; you are asking whether a shape can be drawn. If it can, the conclusion does not follow.

A small extra: draw the diagram that makes the conclusion false, not the one that makes it true. You are trying to break the option, so build the arrangement that breaks it. If you cannot draw a counterexample after two attempts, the conclusion probably does follow — and you have your answer from the other direction.

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.

OfficialNATA 2026 Information Brochure V2.0 · §4.0

Part B examines six named areas: Visual Reasoning, Logical Derivation, General Knowledge/Architecture and Design, Language Interpretation, Design Sensitivity and Thinking, and Numerical Ability.

Visual Reasoning — understanding and reconstructing 2D and 3D composition. Logical Derivation — decoding a situation or context and drawing conclusions. General Knowledge, Architecture and Design — current issues, important buildings, historical progression, innovation in materials and construction. Language Interpretation — meaning of words and sentences, English grammar. Design Sensitivity and Thinking — observing and analysing people, space, product, environment; semantics, metaphor, problem identification. Numerical Ability — basic mathematics and its association with creative thinking; unfolding space using geometry.

Source · verified 2026-07-16

OfficialNATA 2026 Information Brochure V2.0 · §4.0

Part B allows 108 seconds per question, presented one after another, on an adaptive engine.

90 minutes across 50 questions. The adaptive structure dates to NATA 2025 per the President's foreword in V2.0, which states that NATA 2026 continues it.

Source · verified 2026-07-16

What almost everyone believes

If a conclusion is very likely given the premises, it follows.

It follows only if it cannot be false while the premises are true. Very likely is a wrong answer, and it is the answer the distractor is built from.

Ordinary reasoning is probabilistic and this question type is not, so the intuition that serves you everywhere else actively misleads here. Setters exploit exactly that: the tempting option is the one that is nearly always true in practice and not forced by the logic. All coloured drawings are A3, some A3 are unfinished — 'some coloured drawings are unfinished' feels compelling and can be broken in five seconds by putting the unfinished ones among the non-coloured sheets. The defence is to attack rather than derive: try to build the counterexample first.

Depending on how long you have

Foundation

Understand the skill. Months out, or starting from zero.

Learn the small grammar cold: some means at least one and possibly all; all does not reverse; no does reverse. Then practise with overlapping circles until they are automatic. It is a tiny body of knowledge and it covers almost all of this area, which makes it very good value.

Drill

The practice protocol. What to repeat, how often, how to score it.

Practise attacking rather than deriving. For each option, spend five seconds trying to draw an arrangement where the premises hold and the conclusion fails. Track how often the option that felt obviously right died to a counterexample — for most people that number is uncomfortable, and it is the whole lesson.

Exam-Day

What to actually do under the constraint — 108 seconds, no instruments, one pass.

Read the stem for the word: does it say follows necessarily, or could be true? They are different questions. Draw circles. Try to break each option rather than prove one. And distrust the option that feels obviously right — in this area that feeling is manufactured, and it is what the distractor is for.

Try it

Fifteen minutes, with a pen. Draw every one — do not do these in your head.

  1. 01All the studios have windows. Some rooms with windows are cold. Does it follow that some studios are cold? Draw it. (No — put the cold windowed rooms outside the studio circle.)
  2. 02No architects are unregistered. All unregistered designers are uninsured. Does anything follow about architects and insurance? Draw it.
  3. 03Some drawings are in colour. All colour drawings are on A3. Does it follow that some drawings are on A3? Draw it. (Yes.)
  4. 04For each, try FIRST to draw the arrangement that makes the conclusion false. Only if you cannot, accept it.
  5. 05Note how often your first instinct was the probable answer rather than the necessary one. That gap is what you are training away.

The short version

The question asks what follows necessarily, and very likely is a wrong answer — it is what the distractor is made of, because ordinary reasoning is probabilistic and this is not. Attack rather than derive: spend five seconds trying to build an arrangement where the premises hold and the conclusion fails, which kills options faster than reasoning forwards produces them. Learn the small grammar: some means at least one and possibly all; all does not reverse; no does. And draw the circles — draw the one that breaks the option.

Next: the traps, and the difference between what a passage assumes and what it implies.

Questions people actually ask

What does 'follows necessarily' mean in a NATA logic question?
It means there is no possible arrangement in which the premises are true and the conclusion is false. If you can construct even one counterexample, however contrived, the conclusion does not follow. Very likely is not enough, and the tempting distractor is usually exactly the very-likely option.
Does 'some' mean 'not all' in logic questions?
No. Some means at least one, and is perfectly satisfied if all of them are. The everyday implication that some excludes all is a courtesy of conversation, not a fact — and it is deliberately exploited. Similarly, 'all A are B' tells you nothing about whether all B are A.
What is the fastest way to test a logical conclusion?
Attack it rather than derive it. Draw overlapping circles and try to build the arrangement where the premises hold and the conclusion fails. That takes about five seconds and kills options, whereas reasoning forwards is slow and produces only one candidate. If you cannot break it after two attempts, it probably follows.