30 · Six Hour Ambush

Putnam preparation — the actual blueprint. The target is not merely ambitious; it is stronger than winning the competition. In the 2025 Putnam the highest score was 110 out of 120, reached by one participant. Out of 4,335 contestants, 2,618 scored zero. The top-100 cutoff was around 51. This is not an exam where finishing the textbook and behaving responsibly produces a pleasant A*. It is a six-hour ambush conducted with integers. A score of 120 on half of your blind mocks is a sensible ultimate training benchmark, but nobody can honestly provide a finite checklist that guarantees 120 on an unseen paper. Creative problem solving has no deterministic syllabus. Anyone promising certainty is selling motivational wallpaper. What you can build is a closed-loop system that makes your preparation brutally complete.

First: check whether you can officially sit it. The official Putnam is open only to undergraduates enrolled at institutions in the United States, Canada, or Mexico. A UK university doesn't qualify. For 2026 the exam is scheduled for Saturday, December 5th, and the format is changing from two three-hour sittings to four 90-minute sessions with designated breaks. Train using that structure. Even if you can't officially sit it, Putnam preparation is extremely valuable for Oxford-style interviews, the IMC, proof-writing maturity, quant interviews, and serious preparation for advanced mathematics.

What the Putnam actually tests: not primarily advanced theorems. It rewards detecting the right elementary structure before wasting ninety minutes attacking the wrong structure elegantly. You need mastery across proof methods — contradiction, induction, extremal arguments, invariants, bijections, pigeonhole; algebra — polynomial identities, roots, substitutions, recurrences, functional equations; number theory — congruences, divisibility, valuations, Diophantine equations, parity, prime factorisation; combinatorics — counting, double counting, graph arguments, generating-function intuition; analysis — inequalities, sequences, limits, continuity, integration tricks, approximation; linear algebra — rank, eigenvalues, determinants, clever dimension arguments; probability — symmetry, expectation, conditioning, independence; and geometry — coordinate methods, transformations, inequalities, occasional synthetic insight. The MIT Putnam seminar organises training around techniques — hidden independence, sums and integrals, recurrences, inequalities, probability, congruences — combined with mixed supplementary sets. That's the right model: technique-specific drilling followed by random retrieval under pressure.

The minimal resource stack — do NOT collect twenty-seven books and build a Notion museum to your future greatness. Tier one, non-negotiable: the MAA Putnam Archive — your primary dataset. Preserve recent papers for blind mocks; don't casually read solutions from the last decade over breakfast. Use 1938–2009 for topic drilling, 2010–2015 for intermediate mocks, 2016–2025 as the protected blind exam bank. Putnam and Beyond by Andreescu and Gelca — the core structured textbook, over a thousand problems organised by topic with full solutions; build the pattern library before attacking random full papers. MIT OpenCourseWare 18.A34, Mathematical Problem Solving — the Putnam seminar, weekly topic sets; each week MIT students work a long set and submit their six best solutions. Copy the mechanism. Tier two, only after the core stack is active: the Kedlaya–Poonen–Vakil volume for 1985–2000, used for post-mortem commentary, not passive reading; Zeitz's The Art and Craft of Problem Solving for when your weakness is tactical invention — working backwards, invariants, small cases, symmetry, changing representation; and the older Putnam collections as reserve ammunition.