The template

Set up one sheet (or one section) per subject, with these six columns:

DateQuestion / sourceYour answer → correct answerWhy it went wrongTypeResolved?
14 Jul 2023 HSC Adv. Maths, Q11(b) — photo saved x = 3 only → x = 3, x = −5 Divided both sides by x instead of factorising; lost the second root Application ☐ (re-test 21 Jul)
15 Jul Chem qbank set 4, Q9 Chose B → D Didn't know Le Chatelier applies to pressure change here — genuine gap Knowledge
15 Jul Physics past paper, Q3 9.8 → −9.8 m s⁻² Knew it; dropped the sign rushing the last line Execution ☑ 22 Jul

To use it: copy the header row into Google Sheets or Excel, or rule six columns across a double page. That's the whole setup. The value is in the columns you might be tempted to skip, so here's why each one exists.

Every column, justified

  • Date — cheap to write, and it's what makes re-testing schedulable ("everything unresolved older than a week"). It also shows you later whether a mistake type is fading or persisting across the term.
  • Question / source — enough to re-attempt the question cold: paper, year, question number, or a photo/screenshot filename. Do not copy the question out by hand. Transcription is the number-one killer of error logs; a photo takes four seconds and preserves the diagram too.
  • Your answer → correct answer — both, not just the right one. The gap between them is often the diagnosis: seeing you wrote sin where cos belonged says more than the mark scheme alone.
  • Why it went wrong — one honest sentence, written immediately while you still remember your reasoning. "Bad at trig" is a feeling; "used the sine rule with the non-included angle" is a fix waiting to happen. This column is the one that turns the log from a record into a plan.
  • TypeKnowledge (didn't know it → revise that point), Application (knew it, misapplied it → drill that question type), or Execution (knew it, rushed it → fix checking routine). Only the first type is helped by more revision; the category column stops you treating all three the same way. The full argument for these categories is in the complete error log method.
  • Resolved? — the column that separates a working log from a scrapbook of regrets. It only gets ticked after you re-attempt the question later, cold, and get it right without help. Reading your old mistake and nodding does not count.

The routine that makes it work

A template without a review routine becomes a graveyard of screenshots within a month. The minimum viable routine:

  • Log immediately — the moment the mistake happens, under a minute per entry. Batching for "Sunday" fails because the why-it-went-wrong context is gone by Sunday.
  • Review weekly, by re-attempting — 20–40 minutes, every unresolved entry, no notes open. Right → tick. Wrong again → it stays, now flagged as a genuine weak spot deserving study time. Re-attempting cold is the step that actually builds recall — the reason is in why you blank on questions you've revised.
  • Before an exam, revise from the log first — every entry is a question that took marks off you specifically. No revision guide is that targeted.
Reality check. Expect each manual entry to cost 2–4 minutes and the weekly review to cost half an hour. That's the honest price of the manual version — and it's worth paying. If you stop paying it mid-term, don't abandon the method; automate it instead.

Or skip the template entirely

Everything above — the capture, the photo, the why-it-went-wrong diagnosis, the category, the re-testing — is exactly the clerical loop errorboard runs for you. Make a mistake while studying and it's captured with a screenshot, analysed, categorised and filed into a per-topic error log automatically; the quiz then brings your old mistakes back until you stop getting them wrong. It works on Windows, on iPad with a pencil, and with a photo of your paper working from any phone — and it's free to start.

Template or app, the method is the same. Six columns, one honest sentence per mistake, and a weekly re-test. Start today; by your next exam the log will know your weaknesses better than you do.