Radziwiłł's large-deviation conjecture for Selberg's central limit theorem

About 5 years old · traced to

Let τ\tau be uniformly distributed on [T,2T][T,2T], and set

σT2=12log⁡log⁡T.\sigma_T^2=\frac{1}{2}\log\log T.

For k>0k>0, consider the regime V∼k2log⁡log⁡TV\sim k\sqrt{2\log\log T}. Radziwiłł's conjecture. In this regime,

P(log⁡∣ζ(1/2+iτ)∣>σTV)∼Ck∫V∞e−x2/22π dx.\mathbf{P}\Big(\log |\zeta(1/2+{\rm i}\tau)|>\sigma_TV\Big)\sim C_k\int_V^\infty\frac{e^{-x^2/2}}{\sqrt{2\pi}}\,{\rm d}x.

This predicts that the constants from the moment conjecture correct the Gaussian tail in Selberg's central limit theorem at deviations of the order of the variance. The statement is supported by the moment conjecture and by the paper's numerical and theoretical evidence, but remains unproved in general.

References

Primary source

Eli Amzallag, Louis-Pierre Arguin, Emma Bailey, Kelvin Hui and Rajesh Rao, “Evidence of Random Matrix Corrections for the Large Deviations of Selberg's Central Limit Theorem”, arXiv:2104.07403 (2021).

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.