Year 6 · Western Australian selective entry · ACER ASET

Selective entry rewards
reasoning, not drilling.

Korva is at-home practice for the academic selective entry test, pitched at Year 6 level. It covers the four things the paper asks for — reading inference, multi-step quantitative reasoning, visual pattern and written expression — and after every answer it shows your child the reasoning, not just a tick.

No card required. See pricing · See a real question

Fo, the Korva tutor, beside a practice question showing a pattern grid with one square missing
384 questions, all pre-checked No card to start the trial Not affiliated with ACER Built in Perth since 2026
384 questions in the launch bank, across four sections
28 full-length reading passages at exam word counts
70 named skills, each tracked and reported separately
How Korva works

Four moves, in the same order, every session.

Nothing here is a metaphor for how it works. This is what the software does.

1

Your child chooses a section and a level.

Reading, quantitative, abstract or written expression — at easy, medium or hard. All three levels are pitched at Year 6 selective standard; easy is the floor of that standard, not below it. If you would rather they stopped choosing, you can lock the level from Settings and your lock overrides everything else.

2

Fo asks questions back. It does not hand over the answer.

Stuck children get a hint, and hints escalate rather than resolve — three per section, counted on the server, so a refresh does not buy a fourth. The full worked explanation is deliberately withheld until an answer is committed: handed over any earlier it is just an answer key.

Fo retells the explanation the question already shipped with, rather than inventing a new one, so the coach cannot contradict content that was checked before release.

3

Wrong answers come back tomorrow.

A missed question goes into the Mistake Bank and is due again the next day. It keeps coming back until it is answered correctly, which is the difference between having seen a question and having learned it.

4

Mastery updates per attempt, and you can read it.

Every attempt updates a per-skill accuracy figure, so a weak area shows up as a number long before it shows up as a bad result. Once there are ten attempts to work from, the dashboard also computes a single test-ready score out of 100 — and says “not yet” until then rather than showing a figure built on three questions.

This is what your child sees

Nothing on this page was written for the website.

Four unedited items from the bank your child practises against — the passage, the figure, the options and the explanation, exactly as they are served.

Reading Prose fiction Year 6 370 words

The Cartographer’s Daughter

Mei Lin knelt beside the wooden chest in the attic, her fingers tracing the brass hinges her grandfather had polished every spring. Inside lay dozens of maps, each one hand-drawn on yellowing paper, their edges curled like ancient scrolls. She lifted the topmost sheet carefully, recognising the familiar coastline of their province before the new dam had swallowed the valley whole.

Mei Lin had learned cartography at university, where professors taught her to trust satellite imagery and digital coordinates, to value precision over sentiment. She had rolled her eyes at her grandfather’s methods, his insistence on visiting each location personally, his habit of noting details no surveyor would consider relevant: which hillside grew the sweetest persimmons, where migrating cranes rested each October. Yet now, holding these fragile documents, she felt something shift within her chest, a recognition she couldn’t quite name.

Then, deliberately, she reached for her own surveying tools. Not the laptop or the GPS unit, but the leather-bound journal she’d received at graduation and never opened. She turned to the first blank page, testing her pen against the paper. Tomorrow she would walk the western hills herself, before the machines arrived. She would map what mattered, not just what measured. Her grandfather would have understood the distinction. Perhaps, she thought, brushing dust from an ornate border he’d drawn around a forgotten shrine, she finally did too.

Paragraphs 1, 3 and 5 of five. The full passage runs 370 words in the app, numbered by line, and carries five questions — vocabulary, detail, main idea, inference and tone.

Inference Reading · Year 6

In the final paragraph (lines 26-33), Mei Lin’s decision to use the leather-bound journal rather than her laptop suggests she is

Pick an option. Every answer comes back with the line it turns on, and why each wrong option was tempting.

Abstract Hard · 3×3 matrix

Two shapes that matter and one that never does

?
A
B
C
D

The figure is drawn, not described.

Abstract reasoning on the real paper is entirely visual, so a shape explained in words would be a different task — an easier one. Korva ships its own renderer: the question is stored as a structured specification and turned into the picture above at serve time.

The grid, the four options and the position of the answer on this page are not a mock-up. They came out of the same function the practice page calls, which is why this section can be regenerated from the bank rather than maintained by hand.

Source: questions[35] in the launch seed — the skill Korva names Row-and-Column Shape Pairs, at the hard tier. One of 98 abstract items.

The rule: Every cell holds exactly two shapes in a little four-square grid. The ROW decides one of the two shapes and the COLUMN decides the other, so every cell is its row’s shape plus its column’s shape. Where the shapes sit inside the four squares means nothing.

Why C: Look at what is in each little grid, not where it sits — the squares things land in change every time and mean nothing. The pale shape is the same in every single box, so ignore it — only the two dark ones follow the row and the column. So the answer is C.

Quantitative Shared stimulus

Isla has 3 times as many coins as Ravi. Together they have 44 coins.

How many coins does Ravi have?

A44
B12
C11
D33

Isla has 3 times what Ravi has, so together that is 3 + 1 = 4 equal shares. So 4 shares = 44, which gives Ravi = 44 ÷ 4 = 11. That’s C. Option A uses the grand total, not one person. Option B is one off. Option D gives the one with the bigger share.

Note that every wrong option is a named mistake, not noise. That is how the real paper builds distractors, and it is why guessing scores badly.

Written expression Statement prompt

Fair does not always mean equal.

Use this idea as the basis for a piece of writing. You may write in any style.

Fo opens with Fair and equal are two different words here. What is the first situation you can think of where treating everyone the same still was not fair?

Marked on ideas, organisation and expression — the three criteria the exam scores. A piece of writing has no single correct answer, so Korva never converts it into a mastery percentage, and it is left out of the test-ready score for the same reason.

Three questions, one sitting.

The same three items above, in the order a session actually runs them. Sections and levels are mixed by the child, not by us.

1
InferenceReading · medium

“Mei Lin’s decision to use the leather-bound journal rather than her laptop suggests she is…”

Answered → the line it turns on, and why each wrong option was tempting.

2
Row-and-Column Shape PairsAbstract · hard

“Two shapes that matter and one that never does” — a 3×3 matrix with one cell missing.

Missed → into the Mistake Bank, due again tomorrow, and again until it is right.

3
Word Problems with AlgebraQuantitative · medium

“Isla has 3 times as many coins as Ravi. Together they have 44 coins. How many coins does Ravi have?”

Either way → that skill’s accuracy moves on your dashboard the same minute.

Fo, when your child is stuck

It asks a better question. It never answers the one on screen.

A hint escalates instead of resolving, and there are only 3 of them per question. The count is kept on the server, against that question, so clearing cookies or reloading does not buy a fourth, and asking for the last rung first still gets the first one. After the third, the only thing left is to be shown the answer — and that is recorded as a question your child did not get.

The three rungs below are Korva’s own hint text, printed from the constant the app serves them from. A live hint is written for the question in front of your child and is held to the same rule: rung three may hand over the method, never the answer.

1
Orient

Start with what the question is actually asking for — say it back in your own words. What is it after?

2
Narrow

Now find the one thing that changes between the parts. What moves, and what stays put?

3
Method, never the answer

You have the pieces. Work out that middle step on its own first, then see which option it points at.

Test-ready score
68/100
Example view

Inference82%
Word Problems with Algebra74%
Working Back from Area61%
Row-and-Column Shape Pairs44%
Communicating Ideas4 pieces

Marked by rubric, not scored — so it is counted, never converted into a percentage.

Illustrative figures, to show the shape of the view. Not any child’s results. The skill names are real: these are five of the 70 Korva tracks, spelled as your dashboard spells them.

Parent transparency

You see the reasoning, not just the score.

Open any session your child has done and you get the question as they saw it — passage, table or figure included — the option they chose, the option that was right, and the full explanation they were shown. There is no summarised version that hides the detail.

  • Every question attempted, with your child’s answer beside the correct one
  • The full explanation your child was given, word for word
  • Per-skill mastery across all four sections, updated on every attempt
  • A test-ready score out of 100, withheld until there are ten attempts behind it
  • A Sunday-to-Saturday activity chart, with empty days shown as real zeroes
  • A difficulty lock, if you would rather choose the level yourself
Methodology

How a question earns its way in front of a child.

Korva uses AI to write practice material. The interesting part is not that it does; it is what has to happen before anything it writes is allowed to be shown.

1 · A written specification, not a vibe

Each section has its own long-form spec encoding the conventions the real paper uses — passage types, stem phrasings, option vocabulary, how distractors are built. Questions are generated against that document, and a subtype is pre-picked at random so the easiest format is not chosen every time.

2 · Abstract items are solved again, independently

The most damaging failure a language model has is marking a wrong answer correct and then justifying it fluently. So every abstract puzzle is handed to a second, stronger model that solves it from scratch. If that solver disagrees with the marked answer, the question is thrown away and rebuilt.

3 · The hard tier fails closed

Anything claiming the hard tier is scored by a judge that has to show its working against the material it is anchored to — it cannot just assert a number. Missing anchors, an API error, a reply it cannot parse — every one of those means do not ship, rather than shipping by default. What each section is anchored to differs, and card 6 says which.

4 · Figures are rendered, never narrated

Abstract reasoning is a visual task, so Korva ships its own renderer and stores each figure as a structured specification. The item shown further up this page came out of that renderer, not out of a design tool.

5 · The bank is fixed before launch day

What your child meets comes from a reviewed bank of 384 questions that ships with the app, not from something improvised while they wait. Fo’s hints and written-expression feedback are produced live; the questions are not.

6 · We say which sections we can measure

Our abstract and quantitative material is calibrated against genuine Year-6 selective items. For reading we hold no such paper, so that material is judged against written criteria instead — and we would rather tell you that than let you assume otherwise.

Independent preparation. Korva Academy is not affiliated with, endorsed by, or connected to ACER or the WA Department of Education. We prepare students for the skills the ASET assesses; we do not reproduce its questions, and no material on this site is taken from a live or past paper.
Who builds this

A small team in Perth, since 2026.

Korva is not a marketplace of freelance tutors and it is not a repackaged overseas workbook. It is a piece of software, built here, for one exam.

Founder

Ashlin

Engineering at Curtin University, based in Perth. Wrote the question engines, the checking gates and the parent dashboard. Korva has been in development since 2026.

Not published yet
Co-founder

Profile to come

We have not written this one up yet, and we would rather leave a visible gap than publish a biography we cannot stand behind. It goes up when it is real.

Questions parents ask

The things worth knowing before you start.

What test is this for, exactly?

The academic selective entry test used in Western Australia — the ACER ASET, sat in Year 6 for Year 7 entry to Perth Modern School and the other academically selective programs. Korva is pitched at Year 6 level throughout; there is no lower year tier, because the exam does not have one.

How does the free trial work, and do I need a card?

Seven days free, and no card is required to start. During the trial you can set up one child. If you do nothing at the end of it, the account simply stops giving access to practice — nothing is charged and nothing is deleted.

What does it cost after that, and can I cancel?

$7 a week, which covers up to 2 children on the one account. Cancelling is self-service through the billing portal, which also lets you update your card and pull up past invoices. Full detail is on the pricing page.

We are also considering a coaching centre. How is this different?

We can only tell you what Korva is, and you should compare it against whatever else you are looking at yourself. Korva is software, not a class: there is no timetable, no travel, and no fixed pace. Your child practises when it suits, one question at a time, and each question comes back with the reasoning attached. You read the same detail they saw, from your own account, whenever you want. It is $7 a week for the household, and it is not a substitute for a teacher who knows your child — it is a way to get a lot of well-checked practice done at home.

Do I need to sit with my child while they practise?

No. Practice is self-paced and the explanations are written for the child to read on their own. Your job is the review afterwards, which is what the dashboard is for. If you would rather control the difficulty than let them pick, there is a lock in Settings that overrides their choice.

Can my child just look up the answer?

The worked explanation is withheld until an answer is committed — the request is refused outright before that, because an explanation handed over early is an answer key. Hints are capped at three per section and the count is kept on the server, so clearing cookies or refreshing does not reset it. Hints also escalate rather than resolve: Fo asks what the first clue is, it does not say what the answer is.

How much do you know about my child?

A nickname your child picks, and a year level. We never ask you for your child's name — there is no field for it. Your child builds a two-word nickname from lists we supply, so what we hold is “Swift Fox”, not a real name. No surname, no date of birth, no school, no address, no photograph. Profiles made before we changed this still hold the first name that was typed then, and we prompt you to swap it for a nickname. Written-expression responses are a special case: the raw text your child wrote is deleted automatically 24 hours after it is marked, and the score, the word count and Fo’s feedback are kept — that feedback quotes short extracts of their writing, so a phrase or a sentence can outlive the erasure.

You use AI. Why should I trust the questions?

Because we do not trust them either, and the system is built on that assumption. Every abstract puzzle is re-solved by a second, independent model and discarded if the two disagree. Anything claiming the hard tier is scored by a judge that refuses to pass content when it cannot get a clear answer — for abstract and quantitative that judge compares against genuine Year-6 selective items, and for reading, where we hold no such paper, against written criteria instead. Questions are checked and banked before launch rather than improvised while your child waits.

Is Korva connected to ACER or the Department of Education?

No. Korva Academy is not affiliated with, endorsed by, or connected to ACER or the WA Department of Education. We prepare students for the skills the test assesses; we do not reproduce its questions, and nothing on this site is taken from a live or past paper.

What does my child need to use it?

A web browser. There is no app to install. It works on a laptop, a tablet or a phone, though the reading passages and the abstract figures are easier on a bigger screen.

What we can actually promise

Built for a child to use under your supervision.

You own the account

Parents register and configure everything. Children practise inside the space you set up, and cannot change it.

Minimal data

A nickname your child picks, and a year level. No surname, no date of birth, no school, no photograph.

Checked before it ships

Abstract puzzles are re-solved by a second, independent model before they can be served. Hard-tier content is scored by a judge that fails closed.

Deletion on request

Remove a child’s data at any time. Written responses are purged automatically 24 hours after they are marked.

No advertising

No ads, no third-party trackers, nothing sold to anyone. The subscription is the whole business model.

Made in Perth

Built in Western Australia, for the Western Australian selective entry test, and for no other exam.

Start with one question.

Seven days free, no card required. Long enough to see what your child can actually do, and what they cannot do yet.