Randomized investigation reveals no arithmetic frequencies in long-range deviations of Riemann zeros, suggesting further research needed.
This preprint investigates whether arithmetic structure associated with prime numbers can be detected in the long-range deviations of the Δ₃ statistic from Random Matrix Theory predictions for the nontrivial zeros of the Riemann zeta function. Using 60 evaluation points derived from zeros at height T ∼ 10⁸–10⁹, we compare five statistical models via residual sum of squares and the Bayesian Information Criterion. To account for correlations in the data, we apply block bootstrap resampling. As a control against overfitting, we include a model with freely optimized frequencies (Model 1b). While a statistically significant periodic structure exists in the residuals, none of the arithmetic models (based on frequencies ln p, ln(pq), or their combinations) provides a competitive explanation. The free-frequency model significantly outperforms all arithmetic constructions (ΔBIC = 21.86). Even the pure Random Matrix Theory baseline attains a better BIC value than the arithmetic models. These results indicate that, at the examined height, the observed long-range deviations of Δ₃(L) cannot be attributed to the simplest arithmetic frequencies in an additive form. The work provides a rigorous statistical test of concrete arithmetic hypotheses and outlines directions for future research at greater heights.
No takes yet. Share an insight, caveat, or question.
Serhii Kanivets (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: