Normal-distribution and variance conjecture for modular symbols

About 7 years old · traced to

Let EE be a semistable elliptic curve over Q\mathbb{Q} with squarefree conductor NN, and define

CE=6π2∏ℓ∣N(1+ℓ−1)−1L(Sym⁡2(E),1).\mathcal{C}_E=\frac{6}{\pi^2}\prod_{\ell\mid N}(1+\ell^{-1})^{-1}L(\operatorname{Sym}^2(E),1).

For m≥1m\geq 1, define

Var⁡(m)=1φ(m)∑a∈(Z/mZ)×([a/m]+)2.\operatorname{Var}(m)=\frac{1}{\varphi(m)}\sum_{a\in(\mathbb{Z}/m\mathbb{Z})^\times}([a/m]^+)^2.

Modular-symbol distribution conjecture. As mm tends to infinity, the distribution of the values {[a/m]+/log⁡(m):a∈(Z/mZ)×}\{[a/m]^+/\sqrt{\log(m)}:a\in(\mathbb{Z}/m\mathbb{Z})^\times\} converges to a normal distribution with mean zero and variance CE\mathcal{C}_E. Moreover, for every divisor κ\kappa of NN, there is a constant DE,κ∈R\mathcal{D}_{E,\kappa}\in\mathbb{R} such that

lim⁡m→∞(m,N)=κ(Var⁡(m)−CElog⁡(m))=DE,κ.\lim_{\substack{m\to\infty\\(m,N)=\kappa}}\bigl(\operatorname{Var}(m)-\mathcal{C}_E\log(m)\bigr)=\mathcal{D}_{E,\kappa}.

Petridis and Risager proved an averaged distribution theorem over m≤Xm\leq X; this conjecture concerns each modulus tending to infinity and includes a refined second-order variance statement.

References

Primary source

Barry Mazur and Karl Rubin, “Arithmetic conjectures suggested by the statistical behavior of modular symbols”, arXiv:1910.12798 (2020).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.