On the typical rank of elliptic curves over \(\mathbb{Q}(T)\) (Q2674251): Difference between revisions

For elliptic curves over \(\mathbb Q\), ordered by height, there are heuristics that the average rank ought to be \(1/2\). The present article concerns elliptic curves over \(\mathbb Q(t)\). Here the Mordell-Weil group comprising all rational points over \(\mathbb Q(t)\) (or even over \(\bar{\mathbb Q}(t)\)) is again finitely generated unless the curve is constant and the underlying Néron-Kodaira model is a product (but isotrivial fibrations are allowed). It was recently conjectured by \textit{A. Cowan} in [``Conjecture: 100\% of elliptic surfaces over \(\mathbb Q\) have rank zero'', Preprint, \url{arXiv: 2009.08622}] that the average rank over all admissible elliptic curves over \(\mathbb Q(t)\), ordered by the height of the coefficients of the Weierstrass form, should be zero. The authors verify this for special families where the polynomials involved have degree at most two, using the Turàn sieve. They also obtain quantitative bounds which could possibly be further improved using the results of \textit{A. Castillo} and \textit{R. Dietmann} [Acta Arith. 180, No. 1, 1--14 (2017; Zbl 1427.11061)]. In the meantime the results have been superseded and generalized greatly in a preprint by \textit{R. Kloostermann} [``The average Mordell-Weil rank of elliptic surfaces over number fields'', Preprint, \url{arXiv: 2204.12102}].