Not by paper. By what keeps going wrong. That sentence is the subtitle of the clinic inside the Maths Challenge room, and it is the single decision the whole room is built around. When your child gets a question wrong, the wrong answer does not go into a pile called "JMC 2019". It goes into an entry with a name — the mistake it points to — and that entry carries a count.

It is a small-sounding change of filing. In practice it changes what an evening of revision is for.

The problem with filing by paper

Sit four past papers and you have four piles of wrong answers. Each pile is organised by question number, which is an accident of how the paper was printed. Nothing connects question 14 of one year to question 9 of another, even when the same misreading produced both.

So the honest experience of paper-by-paper revision is: your child goes over a paper, understands each wrong answer at the moment it is explained, feels reasonably good, and then makes the same error a fortnight later in a different disguise. Nobody was ever shown that it was the same error. The filing hid it.

Filed by paper Filed by mistake
"Six wrong on the 2019 paper" "This one, four times"
Same error appears as four separate items Four items collapse into one, with a count of four
You revise the paper you happened to sit You work the thing that keeps recurring
Finished when the paper is finished Finished when the mistake stops recurring

The right-hand column is what the clinic does. The count is the whole argument: a ×4 next to a mistake is not a score, it is an instruction about where the next twenty minutes should go.

The four states a mistake can be in

Open Things to fix and each entry is in one of four states.

  • Active — it is live. It has been happening and nothing has yet shown that it has stopped.
  • Watching — there is evidence it may be settling, but not enough to call it.
  • Quiet — it has not turned up for a while.
  • No diagnosis — a real wrong answer that does not point at a particular mistake. More on these below.

The states matter because they answer the question a child actually has, which is not "how many did I get wrong" but "what should I do now". An active entry with a high count is the obvious place to start. A quiet one can be left alone.

And there is a fifth thing the clinic will tell you, which is arguably the most useful of all: "this one came back". That tag appears on a mistake that had gone quiet and then reappeared. A first occurrence might be tiredness. A recurrence after a gap is a genuine gap in understanding, and it is worth more of your child's attention than three fresh wrong answers.

No diagnosis, and why it is there

Some wrong answers are filed under No diagnosis, described in the product as "real wrong answers with no story — worth a look, nothing to fix by name."

The reasoning is stated plainly: some wrong answers are just wrong. UKMT lays out plausible numbers on purpose, so an option a child picked may not point at any particular misconception — they may simply have run out of time, or guessed, or slipped.

This category exists because the alternative is worse. A clinic that insisted on naming a cause for every single wrong answer would invent causes, and an invented diagnosis sends a child off to fix something that was never broken. Leaving a bucket honestly labelled "no story here" is what keeps the named entries trustworthy. We go into this further in when a wrong answer means nothing.

Our reading, not UKMT's

The clinic carries a line on the page itself: "Which mistake an option points to is our reading, not UKMT's."

That is there deliberately. UKMT does not publish, for each wrong option, the misconception it was designed to catch. What the clinic offers is a considered reading of why a particular wrong option is tempting — good enough to organise practice around, not an official ruling, and we would rather say so than let it be mistaken for one. If your child disagrees with a naming, they are not wrong to; the useful part is still the count and the pattern.

The shared trap

Some questions have one trap that several wrong options lead to. Those are tagged shared trap, with the note: "This question had a trap several answers lead to — here's the one you picked."

It is a small honesty, and a helpful one. It tells your child that two classmates who picked different letters may have made the same mistake, and that the letter itself was never the interesting part.

Hiding one, and why that is allowed

A child can hide a mistake. The product's own words: "You'll still meet it in a paper — we just won't keep bringing it up." Hidden ones can be shown again later.

This looks like a strange thing to build until you picture a list of eleven named failings staring at a twelve-year-old. A clinic that a child avoids opening does nothing. Letting them park one — without pretending it has been fixed, and without removing it from future papers — is what keeps the door openable.

What actually moves a mistake

Reading about a mistake does not shift it. From an entry your child can take it to Professor Pi, who walks the hint ladder to a worked twin — a fully worked similar problem — and then asks them to redo their own. Pi never gives the answer to their own question. There is a card that treats it, for spaced practice. And there is the mechanism the product names as the one that counts: a re-trap, booked about a week later, described in what a re-trap is.

The fastest way to fill the clinic in the first place, if your child has already sat a paper at school, is to type in the 25 letters.

FAQ

How is this different from just going through a past paper's answers?

A paper's answers are filed by question number, so the same underlying error shows up as four unrelated items across four papers. The clinic files by the mistake instead, with a count — so four scattered wrong answers become one entry marked with a four, and your child can see that it is one thing, not four.

Is the named mistake the official UKMT explanation?

No, and the product says so on the page: which mistake an option points to is our reading, not UKMT's. It is a considered reading of why that particular wrong option is tempting, offered as a starting point for practice — not an official ruling.

What happens if a wrong answer does not fit any named mistake?

It is filed under No diagnosis, with the explanation that some wrong answers are just wrong. UKMT lays out plausible numbers on purpose, so an option may not point at any particular mistake. Those are worth a look, but there is nothing to fix by name.

Duke Harewood built aitutors.me for his own KS3-aged daughter. The clinic exists because she kept making one particular slip across three different papers, and nothing in the papers themselves would ever have told her that. Updated 10 September 2026.