Elite Mathematics 60 — Methodology v1.0
The frozen annual audit procedure · 2026-27
World's Top Schools Mathematics Ranking Methodology v1.0. The ranking identifies the pre-tertiary schools that provide the world's strongest institutional environment for exceptional mathematics. It is deliberately not a ranking of schools attended by the world's most gifted individual mathematicians. A single extraordinary IMO gold medallist can improve a school's evidence, but cannot by themselves make that school one of the world's best mathematics schools.
1. What the ranking is trying to measure
The ranking asks six questions: · Does the school repeatedly produce exceptional competitive mathematicians? — 35% · How advanced is mathematics actually taught? — 20% · How deeply does mathematical excellence penetrate the school? — 15% · Do pupils undertake proof, research and independent mathematics? — 10% · How strong are the faculty and university mathematics links? — 10% · Is there a genuine institutional mathematics culture? — 10% Total: 100 points. The external-results metric is a rolling ten-season ledger. From 2027 onwards, use 2018–2027 inclusive and roll forward one year each edition. (The 2026 draft described 2016–2026 as ten seasons; 2016–2026 inclusive is eleven — this correction locks the window to ten completed seasons going forward.)
2. Eligibility
A school qualifies only if it is a genuine pre-tertiary secondary-school institution providing general education alongside its mathematics provision. University-affiliated high schools qualify if pupils remain secondary-school students. Specialist mathematics schools, science schools, selective magnets, gymnasiums, lyceums, boarding schools, state schools and independents are all eligible. Early-entry university programmes are excluded where pupils cease being secondary-school students and primarily undertake tertiary study. A standalone Olympiad academy, private coaching centre, after-school mathematics programme or national training camp does not qualify. School networks must be ranked at campus/school level, not aggregated across an entire group. Where an institution has separate domestic and international divisions, results must be attributed to the specific division that generated them. An international division cannot inherit the domestic division's Olympiad record.
3. International & National Mathematical Olympiad Achievement — 35 points
The dominant outcome metric. Build a rolling ten-season school-attributed ledger from official competition and national-selection records. A. IMO achievement — 18 points. Ledger values per pupil: Gold 6 · Silver 3.5 · Bronze 2 · Honourable Mention 0.75 · National-team participation without HM/medal 0.25. These are ledger values, not direct ranking points. Repeated appearances by the same pupil remain valid, but breadth counts each pupil only once. Assign the /18 score by strength of the ten-season ledger: · Sustained multi-student world-leading medal engine — 16–18 · Repeated IMO medals/team places across many years — 13–15.5 · Strong recurring representation and medals — 10–12.5 · Several team appearances/medals — 7–9.5 · Occasional exceptional representation — 3–6.5 · Little or no verified IMO evidence — 0–2.5 B. National-team penetration and breadth — 6 points. The anti-prodigy measure. Calculate national IMO team places supplied by the school ÷ total available national IMO places over ten seasons. Record unique pupils and years represented. · Dominant national pipeline across many years — 5.5–6 · Repeatedly supplies a substantial share of team — 4–5 · Regular representation — 2.5–3.5 · Occasional representation — 1–2 · Isolated individual — 0–0.5 This is the measure that distinguishes Seoul Science High's institutional record from Tonbridge's exceptional Alex Chui era. C. Other international competitions — 4 points. Use officially recognised competitions including EGMO, APMO, Romanian Master of Mathematics, Balkan Mathematical Olympiad, MEMO and comparable regional/international Olympiads. Ordinary commercial mathematics competitions are not equivalent. Highest scores require repeated multi-pupil success. D. National mathematical Olympiad strength — 4 points. Audit the country's principal national selection competition (national Olympiad finals, CMO-style structures, USAMO, BMO selection, equivalent official pathways). Measure national champions; top finalists; training-team selection; density of qualifiers; and repeated school representation. Essential because IMO teams are tiny — a school might produce 20 outstanding national finalists but only one eventual IMO contestant. E. Consistency and institutional breadth — 3 points. Rewards number of different years with high-level results; unique successful pupils; multiple age cohorts; and evidence that the pipeline survives graduation of star pupils. A ten-year school producing ten different international-level mathematicians should beat an otherwise equivalent school whose record is largely one pupil.
4. Advanced Mathematics Curriculum Depth — 20 points
What pupils can formally study at the school, not what a privately tutored pupil learns elsewhere. Standard A Level Mathematics, IB Mathematics AA HL, AP Calculus or equivalent is a starting point, not evidence of world-leading provision. At the upper end look for formal access to multivariable calculus, linear algebra, real analysis, abstract algebra, number theory, combinatorics, differential equations, probability, mathematical modelling, topology/differential geometry, proof-based discrete mathematics and advanced mathematical seminars. · Extraordinary secondary curriculum approaching undergraduate mathematical breadth — 18–20 · Multiple formal post-school-standard/proof-based courses — 15–17.5 · Very strong advanced programme beyond normal school qualification — 12–14.5 · Strong national/AP/IB/A Level provision plus substantial extension — 8–11.5 · Predominantly standard advanced-secondary curriculum — 4–7.5 · Limited evidence — 0–3.5 Formal courses carry more weight than clubs. Externally taken university courses can contribute, but should not receive the same institutional credit unless the school has deliberately integrated them into its programme.
5. Cohort Penetration & Institutional Priority — 15 points
Prevents the ranking becoming 'which school happens to contain the world's best six mathematicians?'. Instead measures how much advanced mathematics shapes the institution. Use actual participation percentages wherever available. · Mathematics is central to the school; specialist/accelerated mathematics reaches a large proportion of pupils — 13–15 · Major specialist stream involving roughly 25–50% of relevant cohort — 10–12.5 · Substantial programme involving approximately 10–25% — 7–9.5 · Established programme but relatively small participation — 4–6.5 · Small Olympiad club or isolated pupils — 1–3.5 · No institutional evidence — 0–0.5 Do not pretend to know a percentage where the school does not publish one. Where exact participation is unknown, score conservatively within a range and reduce the confidence grade. This metric particularly benefits specialist mathematical gymnasiums and mathematics schools where mathematical depth is built into the educational model.
6. Mathematical Research, Proof & Independent Work — 10 points
Distinguishes advanced teaching from mathematical creation. Look for formal proof seminars, independent investigations, research projects, mathematical modelling, extended mathematical theses, externally mentored research, student mathematical journals, conference presentation and publishable work. · Formal, substantial and widely embedded mathematical research/proof culture — 9–10 · Established research programme with external validation — 7–8.5 · Significant independent mathematical work but limited penetration — 4–6.5 · Occasional enrichment projects — 1–3.5 · No meaningful evidence — 0–0.5 An IB Extended Essay in Mathematics does not by itself demonstrate a world-leading research programme. Likewise, ordinary classroom investigations should not be relabelled as mathematical research.
7. Faculty & University Mathematics Integration — 10 points
The mathematical expertise and institutional connections available directly to pupils. Strongest evidence: active mathematicians teaching regularly; university mathematics faculty delivering courses; sustained mentoring; formal partnerships with mathematics departments or institutes; access to undergraduate/graduate seminars; embedded university coursework. · Exceptional, systematic university/professional mathematics integration — 9–10 · Deep sustained partnerships and specialist faculty — 7–8.5 · Strong specialist teaching plus regular external mentorship — 5–6.5 · Periodic university interaction — 2–4.5 · Minimal evidence — 0–1.5 Prestigious proximity does not count. A school located beside MIT, Cambridge or Moscow State University receives no credit unless its pupils demonstrably benefit from the relationship.
8. Mathematical Culture, Communication & Ecosystem — 10 points
Whether mathematics visibly lives throughout the institution. Evidence: sustained math circles; internal Olympiads; pupil-led problem-solving; peer coaching; seminars; mathematical publications; visiting mathematicians; school-hosted tournaments; public lectures; national problem-setting contributions; mathematical camps; older pupils mentoring younger mathematicians. · Mathematics is a defining institutional culture — 9–10 · Extremely strong, multi-layered ecosystem — 7–8.5 · Strong enrichment and competition culture — 5–6.5 · Good conventional extracurricular provision — 2–4.5 · Little beyond lessons — 0–1.5 Hosting an IMO or major tournament does not itself create a high score. The school must have an underlying mathematical culture.
9. Evidence hierarchy
Evidence must be weighted by quality. Tier 1 evidence should carry the audit wherever possible: IMO and official Olympiad records, national mathematics organisations, ministries of education, official school curricula, formal university documentation and published competition results. Tier 2: school annual reports, school news releases, recognised academic organisations and reputable national media reporting specific verified results. Tier 3: specialist education press, alumni materials and reliable secondary databases. Social media, student claims, LinkedIn profiles and promotional marketing should normally be used only as leads or corroboration. A result should never be attributed to a school simply because an unsourced website says the pupil attended it. For an Olympiad result, attribute the pupil to the school attended at the time of the competition.
10. Avoiding double counting
Several safeguards are mandatory. An IMO result can count in the IMO ledger and contribute to breadth/consistency analysis, because those measure different things. It should not be separately counted as another generic international competition achievement. A pupil winning gold at IMO in three consecutive years receives all three medal results. However, they count as one unique pupil in the breadth calculation. A school cannot receive curriculum points for mathematics learned only through a national Olympiad camp or private tutor. National team training belongs primarily to the national system. Give the school credit for producing the pupil, but do not describe national coaching as school provision unless there is evidence that it actually is.
11. Confidence grades
Confidence must remain completely separate from rank and score. · A — Ten-season outcome ledger substantially complete; institutional evidence unusually comprehensive. · A− — Very strong evidence with only minor gaps. · B+ — Strong evidence, but one meaningful component remains incomplete. · B — Inclusion is convincing; exact position remains materially uncertain. · B− — Credible Top-60 candidate but significant evidence gaps make position volatile. · C+ or below — Keep on challenge/audit list rather than present as a secure final-ranked institution. A school must never gain ranking points because it publishes more data. Likewise, difficult-to-research Chinese, Iranian, Russian or Vietnamese schools must not automatically lose points. Missing evidence affects confidence, unless there is affirmative evidence that the provision itself is absent.
12. Scoring precision
For the annual audit sheet, score each component in 0.5-point increments at most. Example: · Olympiad achievement — 32.5 / 35 · Curriculum — 18.0 / 20 · Cohort penetration — 14.0 / 15 · Research/proof — 8.5 / 10 · Faculty/university — 9.0 / 10 · Mathematical culture — 9.5 / 10 · Total — 91.5 / 100 Do not publish a score such as 91.37. And do not publish even 91.5 unless the six component scores have actually been completed and preserved in the audit sheet. This was the principal weakness of the earlier draft rankings: the totals looked more precise than the underlying component audit demonstrated.
13. Ranking and tie-break rules
Rank primarily by total score. Where two schools finish within 0.5 points and the evidence cannot credibly distinguish them, permit a tie. Where a tie genuinely needs breaking, use this order: 1. International & National Olympiad Achievement 2. Cohort Penetration & Institutional Priority 3. Advanced Curriculum Depth 4. Number of unique internationally successful pupils 5. Ten-season consistency Confidence is never a tie-breaker. An A-confidence school is not automatically better than a B-confidence school.
14. Annual rerun procedure
For each year's ranking, follow the same sequence: 1. Freeze a publication cutoff date. Ideally after that year's IMO and other major annual competitions. 2. Advance the ten-season window by one year. For the 2027 edition, use 2018–2027. Drop 2017 completely. 3. Rebuild the external-results ledger first. Add the latest IMO, EGMO/APMO/RMM/regional and national Olympiad results. 4. Refresh the candidate universe. Begin with the previous year's Top 60, then add every school revealed by the latest international medalists, national-team selections and major national finalists. Search in native languages. 5. Audit new challengers before preserving incumbents. Previous rank provides no scoring advantage. 6. Refresh the five institutional categories. Curriculum/faculty structures need not be rebuilt from zero if unchanged, but they must be checked for changes. 7. Score all six categories independently. 8. Assign confidence only after scoring. 9. Run an eligibility scrub. Confirm that every institution remains pre-tertiary and that the achievements belong to the ranked school/division. 10. Apply tie-break rules and freeze the ranking. 11. Maintain a challenge list outside the published ranking. Do not force 60 schools if the evidence supports fewer. 12. Archive the audit sheet and source ledger so the following year's changes can be reproduced rather than researched from scratch.
15. What does not count
University destinations should not directly score in this ranking. Neither should school prestige, fees, endowment, alumni fame, university brand affiliation, selectivity by itself, general examination results, parent wealth, private tutoring access or the reputation of a national mathematics system. A school must earn its ranking through mathematics delivered and outcomes generated while pupils are at the school. Likewise, admissions accessibility should remain a separate reader-information field. Whether an international child can apply is extremely useful to parents, but it should not change a school's mathematical score.
16. The core philosophical safeguard
The ranking should always distinguish three different things: · Peak talent — how extraordinary are the best pupils? · Pipeline strength — does the school repeatedly produce them? · Institutional mathematics — what mathematical education and culture does the school itself provide? The world's #1 Mathematics school should be exceptional on all three. That is why our methodology can place a school with several current IMO gold medallists below an institution with a slightly lower peak but a far deeper ten-year pipeline and much broader mathematical culture.
Our view
This is now frozen as World's Top Schools Mathematics Ranking Methodology v1.0 and will be used unchanged for the 2027 rerun unless we discover a genuine methodological flaw. The most important improvement is defining the rolling window correctly: ten completed competition seasons, so 2018–2027 for next year's ranking. The 35-point Olympiad metric gives us unusually hard evidence, while the remaining 65 points stop the ranking becoming a league table of teenage prodigies. Cohort penetration is particularly important. It rewards schools where exceptional mathematics is genuinely institutional. If we preserve the underlying pupil-by-pupil Olympiad ledger and six component scores this year, next year's ranking becomes an update and audit exercise rather than a complete rebuild.
Confidence grades
- A
Ten-season outcome ledger substantially complete; institutional evidence unusually comprehensive.
- A-
Very strong evidence with only minor gaps.
- B+
Strong evidence, but one meaningful component remains incomplete.
- B
Inclusion is convincing; exact position remains materially uncertain.
- B-
Credible Top-60 candidate but significant evidence gaps make position volatile.
International Access flag
- International Access
The school provides a credible application route for a pupil currently based overseas. This does not mean admission is easy — schools such as NUS High, KSA, Tonbridge, St Paul's, Kaisei, La Salle, Phillips Exeter, Harker and ASMS remain extremely selective.
- Conditional Access
Overseas nationality is not the whole story. Access typically depends on local eligibility (MOE/AEIS in Singapore, right to UK state education, Russian residency, Taiwanese Ministry of Education participation) or on the family relocating with the child. Not open international admission, but not closed either.
- No International Access
An overseas family cannot simply apply internationally to the ranked programme — the pathway is domestic, catchment-restricted, or reserved for citizens/residents. TJHSST, Stuyvesant, Montgomery Blair, IMSA and Philippine Science are the clearest examples.