Abstract Reasoning: The Complete Strategy Guide

Abstract reasoning is the wildcard of the AOSB psychometric tests. Unlike the numerical and verbal reasoning tests, you have probably never come across anything like it before. But the unfamiliarity cuts both ways. There is an advantage buried in it, as well as a risk.

The risk is that this is the test you are most likely to struggle to come to terms with. The advantage is that every candidate is in the same position, so structured preparation arguably moves your score further here than in any other psychometric test. It really is an area where practice pays dividends.

In this guide I will break the test down in full. The format, the strategy that works, and the six pattern dimensions every rule is built from, each with an interactive example you can try as you read.

What Is the Abstract Reasoning Test?

Abstract reasoning is one of the psychometric tests you sit at AOSB. It measures your ability to identify patterns and relationships between shapes, and to do it at speed.

You will sit it at both Briefing and Main Board. You are expected to answer 70 questions in 12 minutes, which gives you a limit of roughly 10 seconds per question.

The Format

Each problem presents two groups of shapes, Set A and Set B. Each group is shown as a panel of six boxes, and every box in a group obeys a single hidden rule. The two rules are independent of one another.

You are then shown five test figures, one at a time. For each, you must decide whether it belongs to Set A, Set B, or Neither. The test contains 14 of these puzzles at five questions each, giving 70 questions in total.

The example below shows what this looks like in practice. Try it for yourself.

The two sets remain unchanged for five questions at a time. You will almost always spend more time on the first question of each puzzle, because you must work out the two rules before you can classify anything. Once the rules are cracked, the remaining four questions go quickly.

The classification logic is strict. A figure belongs to Set A only if it clearly follows Set A's rule, and to Set B only if it clearly follows Set B's rule. In every other case, the answer is Neither.

The Skills It Draws On

Performing well draws on a handful of core skills:

• The ability to generate several possible explanations for a pattern, and to discard them quickly when they fail
• Skill in separating the features that carry the rule from the decoration that surrounds it
• Flexibility to switch between small details and the bigger picture as the puzzle demands
• The analytical discipline to verify a suspected rule against every box before trusting it
• Composure to keep producing ideas when a pattern refuses to reveal itself

None of those are things you either have or you don't. They are ways of working, and they get sharper every time you sit a test.

Why It Matters

No officer will ever be asked to sort shapes into sets, and the link to the job is not obvious. The shapes are a proxy.

The test measures a cluster of problem-solving skills that are difficult to assess directly. These include the ability to generate a range of ideas rather than fixating on the first one, to focus on relevant information while ignoring the clutter around it, and to switch between different types of information.

Add the discipline to test each idea properly, and you have precisely the skills an officer relies on when facing an unfamiliar problem with incomplete information. That is what the assessors are reading in your score.

The Strategy: Nine Rules for Test Day

This is a test you must walk into with habits already formed. Unlike verbal reasoning and numerical reasoning, which draw on skills built throughout your time at school, this test is likely to be completely new to you.

That makes familiarity the foundation. The more patterns and rules you expose yourself to before arriving at Westbury, the higher you are likely to score. But familiarity works best when it is built on sound habits, so follow these rules as you work through your practice tests.

1. Spend Your Time on the Rules

Seventy questions in 12 minutes sounds brutal, but the test is better understood as 14 puzzles than as 70 questions. Almost all of the work in each puzzle lies in cracking the two rules. Once you hold them, classifying a figure takes a few seconds at most.

This should shape how you spend your time. Of the roughly 50 seconds you have per puzzle, it is entirely reasonable to spend more than half working out the rules, leaving four or five seconds per figure.

It can be tempting to rush this stage, because 70 questions feels like a lot and staring at the example boxes feels unproductive. Resist that. A candidate who rushes the rules answers all five questions slowly and badly, while a candidate who nails the rules answers all five quickly and correctly.

2. Treat Set A and Set B as Separate Puzzles

Set A and Set B are independent puzzles with independent rules, and you should crack them separately. Set B is not the opposite of Set A, and it is not a variation on Set A's theme.

Be careful here, because the two sets will often look related. They have to. Every test figure must plausibly belong to either set, which forces the two sets to share a visual vocabulary.

That surface similarity is exactly what tempts candidates into assuming the sets are opposites, finding one rule, inverting it, and moving on. Occasionally the two rules really do sit at opposite ends of one dimension, so an opposite is a hypothesis you are allowed to test. It is never one you are allowed to assume.

The safe habit is to crack Set A, then start Set B from scratch.

3. Find What All Six Boxes Share

The real rule is the one feature shared by all six boxes in a set. Anything that changes from box to box is noise.

The method is to hunt for what is common. The discipline is to verify a suspected rule against every box before you trust it, because a rule that survives three boxes and fails on the fourth was never the rule at all.

You also need to beware of distractors. The tests deliberately fill the boxes with features that have nothing to do with the rule, such as stray arrows, odd letters, jumbled sizes and strange blobby outlines. On harder puzzles the noise is often louder than the signal.

There is a sting in the tail too. A feature that is noise in one puzzle can be the rule in the next, so no dimension can ever be permanently ignored.

Finally, watch for the classic bait among the test figures. This is a figure that exaggerates the noise, for example one containing far more shapes than any example box showed. If shape count is not part of the rule, the figure still belongs.

Judge every figure against the rule you have worked out, never against how the example boxes happen to look.

4. Take Neither Seriously

Neither is a full answer in its own right, not a fallback for when you are unsure. A figure that fits neither rule is Neither. A figure that only partly satisfies a rule is also Neither.

Across a full test, the three answers appear in roughly similar proportions, and this is deliberate. If one answer were rare, a candidate who simply defaulted to the other two would gain an advantage, and the tests are built to prevent exactly that.

So if you have answered ten questions and not once said Neither, the problem is likely your classifying, not the test.

5. Watch the Clock

You have roughly 10 seconds per question, but the puzzle is what really matters, so it is more useful to think in terms of 50 seconds per puzzle. Working to that pace completes the test at 11:40, leaving just enough time to click an answer for anything unfinished.

Some puzzles will crack in 20 seconds and others will need more, so don't obsess over any single one. It does help to have a few yardsticks memorised before the test, and the table below sets them out.

If you complete puzzle 7 in under 5:50 and puzzle 11 in under 9:10, you can relax in the knowledge that you are working at the right speed.

Like the other tests, this one ramps up in difficulty as it goes, with the later puzzles carrying the subtler rules. If you are comfortably ahead of time during the first half, you are buying yourself additional time for the puzzles that will need it.

6. Answer Every Question

There is no negative marking, so you will not lose marks for an incorrect guess. You are literally leaving marks on the table if you don't answer every question.

Unlike numerical reasoning, there is no such thing as an educated guess here. You cannot roughly work out a shape, so a guess is a pure 1 in 3 however long you stare at it.

This simplifies your endgame considerably. Where numerical reasoning rewards you for reserving two minutes to make estimates, here your buffer only needs to be long enough to click through whatever remains.

When you reach your final minute, take stock of how many puzzles are still unanswered. Be ready to start clicking through, so that nothing is left blank at the end.

7. Move On If a Rule Won't Crack

At some point a rule will simply refuse to come. When that happens, give yourself a firm cut-off, somewhere around 60 to 70 seconds on a single puzzle, then answer all five questions as best you can and move on.

The temptation to keep digging is stronger here than in the other tests, because a cracked rule pays off across five questions rather than one. Resist it. The extra time is unlikely to solve the rule, and every second sunk into one stubborn puzzle is stolen from all the ones that follow.

Before you write a puzzle off completely, remember that cracking just one of the two rules is worth far more than nothing.

8. Expect the Later Puzzles to Combine Rules

In the early puzzles, each set is usually governed by a single obvious feature. In the later ones, the rules become subtler and some combine two or occasionally three dimensions at once. A set might require that every filled shape is large and has four sides, so that colour, size and side count all matter simultaneously.

There is a practical tell for these. If a single feature seems to explain part of what the boxes share but not all of it, the rule is often a combination rather than a single dimension.

Say you notice that every box contains a six-sided shape, but that on its own does not feel like enough. Look again and you might find a further condition sitting on top of it, such as that six-sided shape always being shaded. The feature you spotted is usually one half of the rule, so build on it rather than throwing it out.

9. When Nothing Jumps Out

Most rules announce themselves if you look at the set naturally for a few seconds, and that natural look should always be your first move. When it produces nothing, don't simply stare harder.

Fall back on working through the pattern dimensions one by one, in order, using the SPONCS checklist introduced in the next section. A deliberate sweep through the dimensions takes a handful of seconds and will surface the rule far more reliably than hoping it appears.

These guidelines will only get you so far, because they are no substitute for practice. Every puzzle you work through adds patterns to your mental library, and at 10 seconds per question, recognition beats reasoning every time.

Do enough tests and most rules stop needing to be worked out at all, because you have simply seen their kind before. That is the state you want to arrive at Westbury in.

The Question Types: SPONCS

Every rule in the test is built from six pattern dimensions, and they are captured by the acronym SPONCS. Once you have a strategy in place, these dimensions are what you should familiarise yourself with next.

• S for Shape. The type of shape, or its number of sides
• P for Position. Where shapes sit in the box, and what is touching what
• O for Orientation. Direction, rotation and angle
• N for Number. The count of shapes, sides, dots or intersections
• C for Colour. Shading, black or white, hatching and inversion
• S for Size. Relative size, always judged within a single box

You may see the P expanded elsewhere as Pattern. I use Position, because it is concrete and far easier to train against.

The acronym covers shape puzzles, which make up most of the test. The occasional puzzle built from letters or numbers carries its own tag, Alphanumeric, and is covered at the end of this section.

A word on how to use SPONCS, because it is a training tool first and a test tool second. Early in your preparation, run the checklist deliberately, working through the six dimensions in order against each set. By test day, that checking should have become automatic, so that you see fills, counts and positions without reciting anything.

The acronym does remain available as a fallback in the real test, and if a rule refuses to surface it is exactly what you should reach for. But if you find yourself reciting it on every puzzle at Westbury, your preparation has fallen short. The aim is to internalise it until you no longer notice you are using it.

Each dimension below is illustrated with an interactive example for you to try. These examples are deliberately easy, because their job is to demonstrate the dimension cleanly. In the real tests, the rules are subtler, noisier and harder to find.

Shape

The rule turns on the identity or class of shape present. In its easy form, a particular shape appears in every box, such as a triangle.

In its harder forms, the rule is about a class rather than a named shape, for example that every shape is curved or that every shape is straight-edged. The tests disguise the class behind irregular, blobby outlines that don't look like textbook geometry. A blob is still curved, however strange it looks.

Position

The rule turns on where things sit. In its easy form, something always occupies the same region of the box, such as a shape in the top-left.

In its harder forms, the rule is relational rather than absolute. A dot always sits inside another shape, two shapes always touch, or no two shapes ever touch. Relational rules read instantly once seen, but they are rarely the first thing candidates think to check, which is what makes them effective.

Orientation

The rule turns on which way things face. Arrows are the test's signature device here, and they deserve a special warning. Arrows scattered across a puzzle are usually decoration, but every so often their direction is the rule, which is exactly why they make such effective distractors.

In its easy form, all arrows point the same way in every box. In its harder forms, the direction is consistent within each box but differs between boxes, punishing the candidate who checks one box and generalises.

Number

The rule turns on a count. In its easy form, every box contains the same number of shapes, which is close to instant to spot.

In its harder forms, the thing being counted is not visually grouped, such as the total number of sides in the box or the number of intersections between lines. These are difficult precisely because you must think to count something the eye doesn't naturally gather. The hardest number rules relate two counts to each other, for example the number of dots matching the number of straight-edged shapes.

Colour

The rule turns on fill and shading, whether black, white, grey, hatched or half-filled. Colour is the most salient dimension of all, and the first place nearly every candidate looks. That is exactly why the tests so often use it as noise rather than signal.

In its easy form, a particular fill is always present or always absent. In its harder forms, the rule is about proportion rather than presence, for example that each box contains more black shapes than white ones. That defeats the candidate who spots that colour matters but stops at the first formulation.

Size

The rule turns on relative size, and it is always judged within a single box. You will never need to remember an absolute size from one box to the next, and no rule will hinge on two shapes being exactly the same size, because neither can be judged by eye.

What you will meet instead is the largest or smallest shape in a box carrying some property, or the shapes in a box following a clear size order. When size is the rule, the differences are wide enough to see at a glance. If you find yourself squinting to compare two shapes, size is almost certainly not the rule.

Alphanumeric

Occasionally a puzzle uses letters or numbers instead of shapes. The rules behave the same way, with one hidden rule per set, but they draw on different material.

They might be sequences that ascend in fixed steps, vowels against consonants, curved letterforms against straight ones, or the number of enclosed regions in a letter. You will know an alphanumeric puzzle the moment you see it, and no acronym is needed to spot one.

Finally, be aware that these dimensions do not always appear alone. As covered in the strategy section, the harder puzzles combine two or more of them in a single rule.

Target Your Weak Spots

One wrong answer tells you very little. Everyone misreads a pattern occasionally under the clock. But a weakness that appears across several tests is different. That is a genuine area for improvement, and it should be targeted.

This is why every question in the Officer Ready practice bank is tagged in line with the SPONCS headings above. Here is what that looks like at the end of a full 70-question test.

A solid score, but look down the breakdown. The position category was 8 from 14, well below everything around it. That said, one test proves nothing on its own, which is where the cross-test view comes in.

Across every test you sit, the picture builds. Each dimension shows your accuracy over every question you have attempted, with a marker for the average candidate on the same questions. Below is how this is displayed inside the Officer Ready platform.

Now the story is unmistakable. This candidate sits above average on most dimensions, but Position has settled at 58% across 49 questions. That is not bad luck. That is a weakness that needs to be targeted, and awareness is the first step.

It is also a fixable shortcoming. Relational position rules, such as the dot inside the shape and the figures that touch, are rarely the first thing candidates think to check. Drawing awareness to this issue before targeted practice will do more for this candidate's score than another full test sat in hope.

How to Prepare for Abstract Reasoning

Preparation here is a short loop, repeated. Sit a test, find the dimension that cost you marks, then test again.

• Learn SPONCS first. Six dimensions, known before you sit anything, so that you have some strategies in place before you begin practising.
• Sit a full test, timed. Seventy questions in 12 minutes.
• Review by tag. Go through every wrong answer and note if there were any patterns.
• Target the Weak Areas. Go into your next test knowing which dimensions to look out for.
• Keep practising. Sit enough tests that you start to see patterns and rules automatically.

The Officer Ready tests are built for this. They mirror the real assessment at 70 questions in 12 minutes, with the same two-set structure. Every question is tagged so the review step takes care of itself.

Common Mistakes

Almost every mark dropped in this test comes from a handful of repeat offenders. Avoid these pitfalls:

Frequently Asked Questions

Final Thoughts

Abstract reasoning starts as the least familiar of the psychometric tests and ends as one of the most predictable. The format never changes, the rules are built from six knowable dimensions, and the strategy is learnable. What remains is familiarity, and familiarity is simply banked practice.

Spend your time on the rules, and let your wrong answers tell you which dimension to drill next. In this test, more than either of the others, your score is a function of the patterns you have already seen. Build the mental library now, and by the time you arrive at Westbury you should be able to recognise most patterns and rules automatically.