Statistical-noise reduction in correlation analysis of high-energy nuclear collisions with event-mixing

TL;DR

This paper extends and verifies a statistical-noise reduction method for correlation analysis in high-energy nuclear collisions, demonstrating its effectiveness across various particle-pair mixing algorithms through Monte Carlo simulations.

Contribution

The paper generalizes the Reid-Trainor noise reduction method to include arbitrary event subsets and applies it to common mixing algorithms, confirming its broad applicability.

Findings

Statistical-noise reduction occurs in multiple mixing algorithms.

Final errors depend on particle-pair counts, not single-particle counts.

Monte Carlo simulations verify the predicted noise reduction.

Abstract

The error propagation and statistical-noise reduction method of Reid and Trainor for two-point correlation applications in high-energy collisions is extended to include particle-pair references constructed by mixing two particles from all event-pair combinations within event subsets of arbitrary size. The Reid-Trainor method is also applied to other particle-pair mixing algorithms commonly used in correlation analysis of particle production from high-energy nuclear collisions. The statistical-noise reduction, inherent in the Reid-Trainor event-mixing procedure, is shown to occur for these other event-mixing algorithms as well. Monte Carlo simulation results are presented which verify the predicted degree of noise reduction. In each case the final errors are determined by the bin-wise particle-pair number, rather than by the bin-wise single-particle count.