Birch’s theorem and the variable-bound problem for the smooth Hasse principle

For each integer d≥5d\geq 5, determine the least integer B(d)B(d) such that the following holds: for every homogeneous form F∈Q[x1,…,xn]F\in\mathbb{Q}[x_1,\ldots,x_n] of degree dd, let XF⊂Pn−1X_F\subset\mathbb{P}^{n-1} be the hypersurface F=0F=0 and let σ(F)=dim⁡Sing⁡(XF)\sigma(F)=\dim\operatorname{Sing}(X_F). If n−σ(F)≥B(d)n-\sigma(F)\geq B(d) and XFX_F has a nonsingular point over every completion Qv\mathbb{Q}_v of Q\mathbb{Q}, then XFX_F has a nonsingular point over Q\mathbb{Q}.

References

Primary source

arXiv

Progress summary

Refreshed
Claimed progress

A September 2026 preprint claims a better variable threshold for Birch’s theorem, but its exact bound and correctness have not been independently verified.

The problem concerns lowering the number of variables needed for the smooth Hasse principle for degree-dd forms, particularly when d≥5d\ge 5. The retrieved sources record substantial earlier improvements but do not identify a final optimal threshold.

Known results

  • An earlier theorem establishes the smooth Hasse principle when n−σ≥(d−12d)2dn-\sigma\geq\left(d-\frac12\sqrt d\right)2^d.
  • For d∈{3,4,5,6,7,8,9}d\in\{3,4,5,6,7,8,9\}, a stronger condition is n−σ>34d2d−2dn-\sigma>\frac34d2^d-2d.
  • These bounds include nonsingular cubic forms in at least 1313 variables and save 1818 variables over Birch’s bound for quintic forms.

September 2026 claimed improvement

A September 2026 preprint, Improved bounds in Birch's theorem for forms in many variables, reports an improvement to the best known variable bound for degrees d≥5d\ge 5. The numerical bound is not given in the retrieved abstract, and the claim is unrefereed and therefore unverified.

Current status (as of September 2026): Earlier explicit bounds are established, while a newer improvement for d≥5d\ge 5 is claimed but unverified; the optimal variable-bound problem remains open.

Sources

Solutions 0

No solutions have been posted yet.