Hadwiger's Conjecture (1943) is one of the most important open problems in graph theory: every graph with chromatic number at least t contains the complete graph K_t as a minor. It generalizes the four-color theorem. It has resisted proof for over eighty years.
In 1993, Gerards and Seymour proposed the Odd Hadwiger Conjecture — a strengthening. Replace “minor” with “odd minor.” An odd minor requires not just connected subgraphs (branch sets) for each vertex of K_t, but connected subgraphs equipped with proper 2-colorings, where edges between branch sets connect vertices of the same color. The parity constraint looks like an additional demand. It should make the conjecture harder to satisfy and therefore harder to disprove.
Kühn, Sauermann, Steiner, and Wigderson (arXiv:2512.20392, December 2025) disproved the Odd Hadwiger Conjecture. They constructed graphs with chromatic number at least (3/2 − o(1))t that contain no K_t as an odd minor. The construction uses a double-blowup of sparse random triangle-free graphs, exploiting recent advances in the triangle-free process to achieve edge density above the classical barrier while maintaining pseudorandomness.
The construction cannot disprove Hadwiger's Conjecture. The authors state this explicitly. The parity constraint — the feature that made the odd version a “strengthening” — is precisely what the counterexample attacks. Random graphs disrupt the structured 2-colorings that odd minors require. Standard minors impose no parity condition, so the same disruption has nothing to target.
The through-claim: the condition that looks like a strengthening is actually a vulnerability. The Odd Hadwiger Conjecture demands more structure from the minor — proper 2-colorings, parity-respecting edges. That additional structure is precisely what a random counterexample can destroy. The standard conjecture demands less structure from the minor, and because it demands less, there is less to attack. Weakness in specification is strength in resistance.
This inverts the intuition that stronger conjectures are harder to disprove. In mathematics, “stronger” typically means “claims more,” and claims that claim more should fail more easily. But here, the “more” is structural — the odd version doesn't claim a larger quantity but a more specific geometry. And specificity gives the adversary a target. A conjecture that says “something connected exists” is harder to disprove than one that says “something connected with a proper 2-coloring exists,” because the second conjecture has committed to a structural feature that randomness can destroy.
The general pattern: adding constraints to a claim can make the claim more vulnerable, not less, when the constraints create structure that counterexamples can exploit. The four-color theorem is hard to attack partly because graph planarity is a flexible, topological condition — there's no rigid algebraic structure for a counterexample to break. The Odd Hadwiger Conjecture introduced algebraic rigidity (parity) into a topological problem, and the rigidity became the point of failure. The lesson isn't that strong conjectures are easy to disprove. The lesson is that structural specificity in a conjecture's hypotheses is a surface area for attack, and the weakest version of a claim — the one that commits to the least structure — is often the hardest to refute.