Counted from 339 marked wrong answers in the free bank · Maths, English, Science and Chinese · read 18 October 2026
A mark comes home, and the child makes the same mistake again next week. These are the mistakes our own marked answers show most often — counted, not guessed.
Every name below is a thinking pattern we record against a real wrong answer, with how many records it carries. The line after each one is the correction we give the child.
The other 97 named records spread across 35 further patterns, and 46 distinct patterns are in use in all. No pattern in the list is decorative: each one names a signal a human can point at in the answer itself.
Read the list as a parent and one thing stands out: almost none of it is a method problem. A child who knows the method still loses marks on what the words and the question actually mean.
The same count, split by subject, so you know which one to look at first.
In every one of the four subjects the commonest leak is a meaning, not a method. That is the one thing worth telling a child before the next paper.
This is what we store against a wrong answer — the thinking behind the choice, not a tick or a cross.
‘Conference’ means a formal meeting. The detective is deciding who the thief is, not holding a meeting. From our own bank: the stored explanation behind one wrong choice in a P6 vocabulary question (question e3, option a), read from the served file on 15 October 2026. The reason is about the child's thinking, and it never says the child was careless.
A tick tells a child nothing. A reason like this one tells them the exact step to change, which is why every wrong option in the free drills carries one.
So the numbers can be checked rather than trusted.
The free practice page loads one question file. On 18 October 2026 it held 113 questions with three wrong options each — 339 wrong answers. Every wrong option carries a written explanation, and 239 of the 339 also carry a named thinking pattern; the other 100 are the 96 written for Secondary 1 and Secondary 2 English, which name the error in words without a pattern tag yet, plus four deliberately left unnamed, and the notes say why for each one. Those four are counted in the total and excluded from the pattern counts, which is why the list above adds to 239.
The tagging rule we hold ourselves to: a pattern is only ever named on a signal a human can point at. The tag is rejected unless the phrase it quotes appears word for word in that answer's own explanation, so a name here is evidence, not a label somebody guessed at.
For anyone who wants to re-run it: the file is
/psle/psle-all.js, 133,919 bytes, and the first twelve characters of its SHA-256 fingerprint are
f5c68ac80ffe.
Stated plainly, because a number like this is easy to over-read.
In our own bank the most common named mistake is picking a word that means something else (25 records), then turning a meaning the wrong way round (19), then giving a body part or an object the job of another part (15). Eleven patterns carry nine or more records each - 142 of the 239 named mistakes between them.
Every wrong option in the bank our free practice page loads carries a written explanation of the thinking behind it. On 18 October 2026 that file was read and its 339 explanations were counted: 113 questions, three wrong options each, 239 of them carrying a named pattern (the 96 written for Secondary 1 and Secondary 2 name the thinking error in words, without a pattern tag yet). The count is of our own bank and of nothing else.
The bank records the thinking behind a wrong choice, not the marks each wrong choice costs, so it cannot rank mistakes by marks. What it can say honestly is which kinds of mis-thinking show up most often in the answers children actually give.
Yes. Picking a word that means something else is the most common pattern in both - 25 records overall, 8 of them in English and 17 in Chinese, and it is the same mis-thinking in two languages: the choice matched the sound or the familiar phrase instead of the meaning the sentence needed.
No. This is a count of records in our bank, not a forecast and not a national sample. It says which mistakes are worth knowing about; it cannot say what any one child will do. The free diagnostic is what shows which topics are weak for your child.
No. This page counts the mistakes the bank records. Nothing here reads your child's writing.
Yes. The drills are free for your first month, from the day you register, with no signup and no card. The 3-minute diagnostic is free as well.
It counts mistakes. It does not know your child.
Nothing here reads your child's answers, and none of it is a prediction. The page's only job is to name the mistakes that keep turning up, so a parent can recognise one when it does. Which of them your child actually makes is a question only practice can answer — the free 3-minute diagnostic gives you a topic-by-topic gap map for Mathematics, with no card and no signup, and every wrong answer in the free drills comes back with a written reason rather than a mark.
Everything on this page is free — no signup, no card. Send it to one parent on WhatsApp; it takes two seconds and it is the only way other families hear about us.