Spend the final ten to twelve weeks of AA HL on familiar topic sets or premature full-paper runs, and you’ve burned the one window where targeted work actually moves the needle. The problem rarely is effort—most students in the middle band are working hard. It’s sequencing. Capable, broadly syllabus-complete students misallocate this phase constantly, treating it like mid-course revision when it calls for something structurally different. In late DP2, how you allocate each study block matters more than how many blocks you log.
The exam itself has shifted the terms. Multi-part calculus and vector questions in Section B demand that you carry a single context across several sub-parts, reusing earlier results accurately—a demand that short drills never replicate. Paper 3 pushes further into unfamiliar application settings and sustained reasoning across five or six linked steps. Moderate weaknesses in calculus, complex numbers, and differential equations that felt survivable mid-course become disproportionately expensive when they sit at the center of these higher-mark structures.
Running a Rapid Skills Audit
Untargeted drilling is the most expensive mistake you can make at this stage—it generates volume, feels like forward motion, and does very little about the specific gaps that will cost you marks. The IB Math AA HL question bank’s real value right now isn’t as a topic refresher; it’s as a diagnostic instrument. Filter to medium–high difficulty across four clusters—calculus applications, vector geometry, complex numbers, and probability distributions—and attempt a representative spread without notes, untimed. For each question, tag the primary issue as a procedural slip, a concept gap, or an unfamiliar application, so you can separate what’s remediable in a few targeted sessions from what needs a full re-teach block. Those are very different time commitments, and confusing them is how Phase 1 expands indefinitely.
- Timebox: 2 blocks of about 45 minutes (or 3 blocks of 30), attempting questions without notes.
- Sample: pick 8–10 medium–high difficulty questions per cluster (around 35–40 total across calculus applications, vector geometry, complex numbers, and probability distributions).
- Score each question quickly for triage: 2 = correct with a clear, coherent method; 1 = some method marks likely but with a break in logic, communication, or accuracy; 0 = no viable method or a wrong setup.
- Tag each miss with a single label: procedural slip, concept gap, unfamiliar application, or communication/method-mark loss.
- Rank gap candidates: Fix-first gaps are those that appear repeatedly, look procedurally or application-based (rather than deeply conceptual), and are causing many 0s or 1s; rare, foundational concept gaps belong in a separate re-teach block if you have time, not in your main drilling list.
- Lock Phase 1 scope by choosing only 2–3 fix-first gaps in total, across all clusters. Everything else is Phase 2 exposure, not something you will grind in the early weeks.
- Stop auditing once you have sampled all four clusters and the same 2–3 gap themes keep appearing in your last ten or so questions; at that point, the audit has done its job and you should move into focused practice.
Once the audit is complete, treat its output as the basis for spaced, retrieval-based practice rather than topic cramming. A 2025 ERIC meta-analysis on mathematics learning found a small-to-medium benefit for spaced over massed practice (g = 0.28) and a positive but less robust effect for testing versus restudy (g = 0.18)—which supports distributing targeted question-bank sessions across weeks rather than bunching them by topic. That still leaves one harder call: not every gap the audit surfaces will repay the same investment, and the prioritization logic isn’t self-evident from the data alone.
Targeted Gap Closure
The audit output is a short list, not a curriculum. For most AA HL students in the middle band, calculus applications and complex numbers deserve first attention when they show up as remediable gaps—they sit inside longer, higher-mark questions more reliably than the other clusters. Vector geometry and probability distributions matter too, but if your main losses trace back to calculus or complex-number work, those clusters offer the best return per revision hour. Start there.
Once your 2–3 fix-first gaps are locked in, shift into disciplined drilling. Work medium–high difficulty questions in focused blocks of 30–40 minutes per cluster, spread across non-consecutive days, and mark immediately against markschemes. Use the same error tags from the audit—procedural, concept, unfamiliar, communication—to watch whether mistakes are converging into minor slips or persisting as genuine gaps. Revisit the same micro-skill or question type three to five days later, without checking your notes first, to confirm the gain holds. And if a cluster didn’t surface in the audit as a clear weakness, keep it out of Phase 1 entirely. The time you protect there is exactly what funds the extended-response work still ahead.
A Different Mode — Section B and Paper 3 Fluency
Extended questions in Section B and Paper 3 demand skills that topic drills barely touch. You have to sustain a single context across five or six sub-parts, carry earlier results accurately into later reasoning, and hold mathematical communication together throughout. Drilling short, one- or two-mark items builds topic familiarity. It doesn’t build the extended-response machinery that Section B and Paper 3 actually test—and by late DP2, that gap is already visible in timed conditions.
In Phase 2, build regular sessions around full extended-response items: choose one question from the AA HL bank, work it through completely with clear reasoning and written justifications, then compare your solution to the markscheme to see exactly where method and communication marks are awarded and where they aren’t.
Phase 2 sessions should also train you to find entry points and maintain stamina on long Section B questions: take a multi-part question, start from a later sub-part you can interpret, earn whatever method marks you can, then reconstruct the chain linking back to earlier parts—because later sub-parts can still be accessible even when part (a) stalls you. Mix that with occasional full timed sections to rehearse both the micro-skill of entering a problem midstream and the stamina to stay focused across a long, connected task. Both phases have clear jobs. The harder question is how to sequence them within ten weeks without letting one crowd out the other.
The Ten-Week Schedule
The sequence is a dependency, not a preference. Phase 2 extended-response work requires that your remediable gaps are already largely closed—otherwise you’re practicing extended reasoning on a foundation with known holes. That’s why the audit comes first, Phase 1 closes before Phase 2 opens, and the transition between them happens at a fixed point rather than whenever Phase 1 feels comfortable. In weeks 1–2, complete the skills audit across all four clusters and lock your Phase 1 target list. Run the no-notes, medium–high difficulty sampling across all four clusters, and keep going until the same two or three gap themes keep reappearing in your final set of questions. At that point, the audit has converged. Stop and move in.
Weeks 3–7 are for Phase 1 gap closure using spaced retrieval. Run two or three targeted sessions per week on your chosen gaps, each 30–40 minutes, and use your log and its decision rules to judge progress week by week. Don’t let Phase 1 run past week 7. Even if minor issues linger, the extended-response runway from week 8 onward isn’t negotiable—late discoveries matter far less than having enough time to build fluency in long, connected work.
- For each question you attempt, keep a one-line log: date, cluster, question ID, score (2/1/0), error tag (procedural, concept, unfamiliar, or communication), a short note on the fix, and a planned reattempt date.
- When you schedule practice, build in reattempts of the same micro-skill or a closely parallel question 3–5 days after the first attempt, and do the reattempt without checking notes beforehand.
- If a particular gap still has fewer than about half of its questions scoring 2 after two full reattempt cycles, either allocate one focused re-teach block followed by a final test, or consciously drop it as a Phase 1 target and move that time into Phase 2 extended-response and exam-simulation work.
In weeks 8–10, shift most of your time into the Section B and Paper 3 mode you built in Phase 2 prep: full extended-response sessions and a small number of full timed simulations. Preserve three or four authentic past papers for weeks 9–10 only, under strict timed conditions, and continue using the question bank for additional Section B and Paper 3 items so those papers stay as clean benchmarks. In the final days, reduce new-question volume and focus on markschemes, your error log, and the communication standards that turn sound methods into actual marks.
Locking In the Audit-First Sequence
The audit-first sequence isn’t a scheduling preference—it’s structurally necessary. You can’t identify which gaps are costing you marks without focused diagnostic work, can’t close them efficiently without spaced retrieval, and can’t build extended-response fluency on gaps that haven’t been addressed. Students who skip the audit and go straight to mock papers aren’t working smarter; they’re rehearsing their current weaknesses at full scale. The next move is concrete: open the IB Math AA HL question bank, filter to medium–high difficulty in calculus and complex numbers, and treat that first session as a diagnostic pass—no notes, quick scoring, honest tagging of where your methods actually break. Ten weeks is enough time to do this properly. It’s not enough time to do it twice.
