Burn through your best 2025-spec papers in the first three weeks of revision and you’ll have a strong sense of how they felt—but almost no information about where you actually stand. The real constraint in IB Math AA SL preparation isn’t time or motivation; it’s the small pool of high-trust IB Math AA SL practice exams that match the current specification. Past papers and specimens are finite, unofficial prediction sets vary widely in quality, and when everything gets labeled a ‘mock,’ scarce materials disappear before your skills are developed enough to extract real diagnostic value from them.
The answer isn’t to search harder for more papers. It’s to assign each one a specific job. Three functions matter: building topic fluency through repeated retrieval, calibrating your readiness against grade boundaries, and running full-exam simulations under authentic conditions. Most preparation problems—random sequencing, shallow review, premature exhaustion of the best materials—trace back to treating all three as interchangeable. That taxonomy is the governing logic for the twelve weeks ahead. The unresolved practical question is which resources belong to which category—and how to tell before you’ve already used them up.
Four Resource Types and Their Purpose
The most reliable sorting criterion isn’t age or source—it’s whether a resource respects the AA SL exam’s structural split. Paper 1 is a 90-minute non-calculator assessment; Paper 2 is a 90-minute calculator paper. Any resource you treat as a full simulation needs to mirror that division exactly. Rehearsing non-calculator reasoning with a GDC quietly available isn’t practice for Paper 1—it’s training a different skill entirely.
From that structural constraint, four types take shape. IBO specimens and genuine 2025-specification papers earn the highest-trust simulation slot. Pre-2025 AA SL papers are useful for topic-level retrieval but only partially aligned with the current specification. Teacher-constructed mocks function as strong integration sets when they respect the current format. Third-party prediction papers vary enough in quality that automatic trust is a mistake.
Each type serves one of three functions. Full timed simulation is the rarest and should be reserved for the closest 2025-spec papers. Topic-fluency work is abundant—older papers filtered by topic and question banks can be reused many times without consuming a real mock slot. Calibration sits between the two: use scores from teacher mocks or structured question sets to benchmark readiness and decide what to address next.
To make that benchmark concrete, it helps to know how the overall AA SL mark is constructed. Paper 1 contributes 80 marks, Paper 2 contributes 80 marks, and the Exploration (IA) contributes 20 marks, giving a total of 100. Converting any mock score into a percentage against that total is how calibration becomes a planning tool rather than a feeling. A TestPrep Istanbul grade-boundary guide for the May 2025 session puts the overall bands (out of 100) this way: 7 = 80–100, 6 = 67–79, 5 = 53–66, 4 = 37–52, 3 = 26–36, 2 = 13–25, 1 = 0–12. Those bands are historical—future sessions set their own thresholds—but they give your mock scores a reference point. If your converted total sits in the mid-5 range, you know the gap to a 6 and can plan the remaining weeks around closing it.
A Twelve-Week Sequencing Framework
The twelve-week plan starts by locking a weekly rhythm, then placing specific papers into that structure rather than improvising week by week. Fixing the cadence first is the key move: spaced, distributed retrieval consistently outperforms concentrated effort. A 2024 STEM education study found that spaced retrieval practice produced a significant positive effect on test performance when results across nine introductory STEM courses were pooled, even though the effect was not uniformly strong in every course. For AA SL, that means prioritizing regular, repeated retrieval over a small number of isolated full-paper attempts, and building toward simulation rather than leading with it.
- Inventory your materials by function (10 minutes): label everything you have.
- Gold: the closest 2025-spec full Paper 1 and Paper 2 sets—few in number, highest trust.
- Silver: teacher mocks and strong third-party papers that pass your quality checks.
- Bronze: older papers by topic, question-bank sets, and short mixed-topic drills.
- Lock the weekly cadence first, then place papers into it:
- Two topic-fluency sessions (45–75 minutes each): targeted retrieval using Bronze materials, plus 10–15 minutes of correction.
- One mixed integration set (60–90 minutes): multi-topic, exam-style sequencing using Silver or Bronze.
- One timed component sitting, alternating formats—Week A as a Paper 1-style non-calculator block; Week B as a Paper 2-style calculator block.
- One review block within 24 hours of the timed sitting (45–90 minutes): error-log updates and markscheme learning.
- Reserve Gold sets for Weeks 10–12:
- Three Gold sets: run one full Paper 1 + Paper 2 simulation in each of Weeks 10, 11, and 12.
- Two Gold sets: run them in Weeks 10 and 12; use the best available Silver set in Week 11.
- One Gold set: run it in Week 12; use Silver sets or split-paper mocks in Weeks 10–11.
- Phase the emphasis as the weeks progress:
- Weeks 1–4: keep timed sittings as single-component blocks; prioritize Bronze by-topic retrieval and rapid correction.
- Weeks 5–9: make integration sets harder and more exam-like; increase Silver use; keep alternating timed Paper 1 and Paper 2 blocks.
- Weeks 10–12: replace the single-component timed block with a full Paper 1 + Paper 2 simulation on a fixed day; reserve the following day as a non-negotiable diagnostic review.
- Guardrails that prevent paper waste:
- Never schedule a Gold simulation without a dedicated review block the next day.
- Never run two full simulations on consecutive days without at least one retrieval-focused session between them.
- If Paper 1 non-calculator errors are rising across two consecutive weeks, make the next timed component a Paper 1-style block—even if the alternation would have suggested Paper 2.
The phases are sequenced this way because different kinds of practice extract different value, and the order isn’t arbitrary. Weeks 1–4 use frequent retrieval across the whole syllabus to make every topic accessible on demand, supported by topic-filtered question banks without consuming any limited full papers. Platforms designed for IB exam preparation, such as Revision Village, are especially useful here precisely because they deliver repeated retrieval without spending simulation-grade resources. Weeks 5–9 extend that fluency into connected thinking, pushing algebra, functions, calculus, and statistics together inside longer mixed questions where timing and sequencing strategy develop under realistic pressure. By Weeks 10–12, you’re sitting full-length papers with skills sharp enough that every mark lost becomes a precise, actionable signal—but only if you have a system for reading what those papers are actually telling you.
Turning Every Completed Paper Into a Diagnostic Tool
A completed paper is only as useful as what you do with it in the 24 hours that follow. Start by separating error types: a conceptual gap, where the underlying idea didn’t land, requires different repair than an execution slip, where an algebra error or misread command term dropped marks you otherwise knew how to earn. Then maintain a rolling log that groups misses by topic and question type, because patterns only surface across a couple of weeks, not from a single sitting read in isolation. Study the markscheme for each miss not just to find the right answer but to understand what earns marks: method marks, explicit justification, correct rounding, valid calculator setup. From there, calibration becomes active. Compare each paper’s percentage to the boundary bands you’re using as a benchmark, then adjust—more topic repair, more timed practice, or another integration set, depending on what the log shows.
Setup (once): keep a running error log with six fields per miss—Topic; Question type; Error type (concept, execution, or strategy); Trigger (for example: algebra slip, misread command term, calculator setup); What the markscheme rewarded; Next drill to prevent a repeat.
After every timed sitting (within 24 hours): rework all misses without notes, then check the markscheme and update the log so you capture the scoring standard, not just the correct answer.
Weekly review (about 15 minutes, same day each week): scan only the last two weeks, tally the top two topics and the single most common error type, and decide how to adjust the upcoming sessions.
Decision rules: if the same concept gap appears at least twice in two weeks, swap one mixed integration set for a targeted topic-fluency session on that area; if execution errors dominate, convert one session into short timed drills with a forced line-by-line checking routine; if you’re repeatedly missing method marks, spend a short session rewriting three past solutions in the exact structure the markscheme rewards.
Calibration tie-in (optional): note your paper percentage next to each week number so you can track whether your error profile is improving even when the raw score is temporarily flat.
All of this diagnostic work depends on one thing the log can’t control: the quality of the papers being fed into it.
Evaluating Third-Party Papers and Preparing for 2025 Variant Differences
Not every third-party paper earns a full sitting. Before investing two or three hours under timed conditions, a 30-minute audit tells you whether a paper belongs in a simulation slot or should be downgraded to topic practice. Check first whether it clearly separates Paper 1 (non-calculator) from Paper 2 (calculator) and whether both components match the expected length and structure. Sample questions from early, mid, and late in each paper: at least one should require multiple steps and draw on more than one topic, and the command terms should read like real IB phrasing rather than generic instructions. Skim the markscheme for method marks and explicit reasoning expectations—a markscheme that only records final answers can’t show you how the marks are actually allocated.
The calibration question is ultimately simple: does this paper produce scores that reflect something true about your readiness? Strong markers include command terms such as ‘find,’ ‘show that,’ or ‘hence’; questions that link across the syllabus rather than testing single skills in isolation; calculator assumptions that respect the non-calculator/calculator divide already built into your weekly cadence; and a difficulty spread that places most questions in the plausible IB range rather than at the extremes. If a paper fails two or more of these, downgrade it. Use its stronger questions as drills, but don’t base a calibration decision on its scores.
The same logic applies to 2025 variant uncertainty. You can’t know in advance exactly how a specific function, sequence, or statistics concept will be framed in a given session—but that uncertainty shrinks considerably when you’ve met each idea across many different question formats, spaced over weeks, and progressively mixed with other topics. That’s precisely what the Bronze-to-Silver progression in the twelve-week framework is designed to produce: Bronze topic retrieval across Weeks 1–9 builds underlying fluency, and Silver integration sets force those concepts into multi-topic combinations before any Gold simulation is attempted. By the time an unfamiliar question framing appears, the underlying structure is already well-rehearsed. When classification, sequencing, diagnostics, and quality screening are all working together, the fixed pool of papers you have becomes something more than a resource constraint—it becomes a readiness signal you can trust.
Turning Scarce Papers Into a Structured Advantage
Scarcity of authentic 2025-spec AA SL papers isn’t a solvable problem. It’s a fixed condition. The students who benefit most from that constraint aren’t the ones who track down more papers—they’re the ones who stop treating preparation as a volume question.
Assign each resource a clear function. Protect your highest-trust papers until your skills are developed enough to actually learn from them. Log errors with enough precision to adjust the following week, and screen third-party materials before trusting them with a Gold simulation slot.
Done right, twelve weeks of deliberate sequencing will extract more diagnostic value from three well-timed Gold papers than twelve weeks of unstructured sitting will extract from thirty. The papers within your reach are enough. The question is whether you’ve given each one a job worth doing.
