Two students can log identical revision hours and walk into their IB Math exam with very different scores. The gap between them isn’t discipline—it’s direction. Past paper practice without a diagnostic layer tends to reinforce what a student already handles well; the specific subtopics driving repeated mark loss sit untouched, session after session.
This compounds across a two-year course because the IB syllabus builds on itself. An unresolved gap in algebraic method will cost marks again when that skill reappears inside a calculus or probability question—the syllabus doesn’t forget. The fix isn’t more practice volume. It’s a short, repeating cycle: run a mixed diagnostic to expose the real gap, drill that specific subtopic, then re-test to confirm it’s closed.
Table of Contents
Step 1 — Run a Mixed Diagnostic Set
A diagnostic set is not a chapter test. It’s a short run of mixed-topic questions drawn from past papers or a tagged question bank—no topic prompt, no thematic cue—so retrieval has to work across the full syllabus. That’s closer to real exam conditions, which is exactly the point. The goal isn’t to confirm what you know; it’s to surface what breaks under pressure.
The output isn’t a score. It’s a logged list of every question where marks were dropped, recorded with enough detail to sort those drops by subtopic and error type. A score tells you where you landed; the log tells you why. One sitting produces a snapshot; running this check repeatedly across a revision period is what turns snapshots into a pattern you can act on.
- Setup (2 minutes) — Choose 15–25 mixed questions from a past paper set or the IB Math Questionbank. Open a blank error log.
- Work straight through under timed, exam-like conditions. Don’t pause to revise mid-set; the goal is to surface what fails under retrieval pressure, not to study along the way.
- Log every dropped-mark question (10 minutes): for each one, write the source and question ID, a surface topic tag (e.g. Functions, Calculus, Probability), one error type—recall/definition | method/setup | algebra/notation | interpretation/command term | calculator/technology | time/strategy—and one line naming the missing micro-step (e.g. “set domain restriction before solving,” “write +C and apply boundary condition”).
- Pick one drill target (3 minutes): the subtopic that appears at least twice in your log, OR caused the largest single mark loss, OR is a prerequisite for other questions you missed.
- Run a targeted drill block (30–45 minutes): pull 8–15 questions tagged to that subtopic from the IB Math Questionbank. Each time you make a mistake, add a one-line correction to your log that names the micro-step you should have used.
- Quick checkpoint (5 minutes): answer 2 fresh questions on the same subtopic without looking at solutions. If the same core steps still fail, repeat another short drill block on that subtopic; if the method now holds, schedule a mixed re-test within 24–72 hours rather than drifting into broad, untargeted revision.
Step 2 — Name the Subtopic, Not the Subject
Surface topic labels—”Functions,” “Statistics,” “Calculus”—are too broad to act on. The useful diagnostic output is specific: not “Calculus” but “integration by parts,” not “Probability” but “continuous random variable setup.” A student who logs only subject-level errors will design revision that’s still too diffuse to close the real gap. A 2025 systematic review in ZDM–Mathematics Education, drawing on 45 studies of formative assessment in mathematics, finds that diagnostic cycles of this kind—elicit evidence, interpret against a specific learning goal, decide the next step—tend to be more effective when content-specific and sustained regularly over time. Generic topic labels undercut the method at exactly the interpretation stage.
Cross-topic dependencies in IB Math AA HL make exact subtopic tagging especially consequential. Integration by parts from Calculus is required to handle continuous probability distribution questions in Statistics; Euler’s form from Number and Algebra feeds into trigonometric identity work in Geometry and Trigonometry. What looks like a statistics error may trace back to a calculus gap—and mislabeling it keeps revision pointed at the wrong target. For each dropped question, identify the specific operation or concept you failed to execute, and tag your log at that level before deciding what to drill next.
Step 3 — Drill the Subtopic, Then Re-Test with Mixed Questions
Once the diagnostic names a precise subtopic—proof by induction, oblique asymptote sketching—the next move is concentrated practice on that skill alone. Not a full chapter. Questions tagged to that subtopic, until the method is consistent. The IB Math Questionbank’s subtopic tagging makes this direct: pull a scoped set without encountering adjacent material. Speed matters less than reliable execution at this stage—the goal is a method that holds without prompting.
Re-testing is the step most students skip, and skipping it is where the cycle breaks. Confidence rises after drilling—but confidence built on isolated, labeled practice doesn’t prove the gap is closed under exam conditions. The re-test must use mixed questions drawn from the full paper range: identifying the problem type before solving it is what the exam actually demands, and targeted drilling alone can’t train that. If the previously weak subtopic holds under varied conditions, the loop closes. If marks still drop, the diagnostic begins again.
- Re-test setup: within 24–72 hours of drilling, run a mini mixed set of 8–12 questions covering major topics, including 2–3 where the target subtopic could plausibly appear. Select questions across different topics and sources—including from the IB Math Questionbank by topic tag—so you are identifying question type before solving, not repeating practiced patterns.
- Weekly cadence (10 minutes): scan your last 2–3 logs, pick the next target by highest repeat frequency or highest mark loss per appearance, and run the next diagnostic set.
Where to Point Your First Diagnostic
Not every AA HL subtopic carries equal exam weight—and the historical pattern is fairly legible. Practitioner evidence from Pietro Meloni Insights (March 2026) maps which AA HL subtopics have recurred as full long questions across multiple exam sessions; these are recurring past patterns, not a guaranteed forecast of future exam composition. Proof by induction has appeared as a full long Paper 1 question in almost every exam session. Function-behavior tasks on non-calculator Paper 1 are consistently underestimated despite being tested repeatedly. Three-dimensional vector geometry regularly anchors 15-plus-mark multi-part problems on Paper 2. In Statistics, continuous random variable questions sit at the intersection of Topics 4 and 5, meaning weak calculus will cost marks there directly. These areas are productive starting points for a first diagnostic given their historical frequency.
Those subtopics belong specifically to the AA HL syllabus. For AA SL and AI students, the same three-step cycle applies—run a mixed set, sort dropped marks by subtopic, drill the flagged skill, re-test with mixed questions—but the specific subtopic priorities differ by course and level. The practitioner evidence here is drawn from AA HL; let your own diagnostic output, rather than this list, determine where the cycle begins.
Turning the Diagnose–Drill Loop into a Habit
What accumulates from running this cycle—more than any individual score—is a precise map of where your preparation is genuinely solid and where you’ve been mistaking familiarity for competence. Each re-test either confirms the fix and moves you forward, or restarts the loop on a sharper target. Both outcomes are informative. Two students can log identical revision hours; the map is what produces the different scores.