Strong geography restriction for rational homology spheres
For every rational homology -sphere , let the nonzero -torsion summands of , viewed as an -module, have lengths , where a summand of length is isomorphic to . If at least two such summands occur, then .
References
- U torsion structure in Heegaard Floer homology — arXiv — Yuhui Chen, Shunyu Wan
Source: arXiv
Progress summary
A 2026 preprint claims an infinite family of counterexamples, but the disproof has not been independently checked.
The problem asks whether every rational homology sphere satisfies the strong Heegaard Floer geography restriction formulated in the 2024 paper by Isakova, Lidman, and Zemke. That paper left the question open while proving it for important surgery classes.
Known results
- Knot surgeries, including nontrivial rational surgeries, satisfy the restriction (Isakova, Lidman, and Zemke, 2024).
- Sufficiently large surgeries on links in satisfy it (Isakova, Lidman, and Zemke, 2024).
- The result extends to nullhomologous knots and links in any -space (Isakova, Lidman, and Zemke, 2024).
- Rustamov’s examples , , are not large surgeries on links in ; this is not a counterexample to the geography restriction.
October 2026 claimed counterexample
On October 20, 2026, a report described Yuhui Chen and Shunyu Wan’s preprint as constructing an infinite family of explicit integer homology spheres violating the strong restriction and lying outside several surgery descriptions. This would disprove the conjecture, but the retrieved evidence contains no independent mathematical verification.
Current status (as of October 2026): The restriction is proved for several surgery classes, while Chen and Wan claim an infinite-family disproof for general rational homology spheres; that claim remains unverified.
Solutions 0
No solutions have been posted yet.