ATP 2021–2025 Cohort-100 Tournament Portfolio Optimization Dataset

Description

This repository contains the model-ready data, frozen analysis outputs, validation files and reproduction code supporting a robust access-contingent ATP tournament-portfolio optimization study. The release covers a fixed cohort of 100 players drawn from the observed 2025 rank 1–300 universe, historical estimation and calibration periods from 2021 to 2024, and chronological evaluation over the complete 2025 season. The package includes tournament opportunities, weekly player states, entry-feasibility estimates, ranking-point schedules, travel arcs, optimized policies, statistical comparisons, sensitivity results and complete provenance documentation. Raw third-party match files and row-level reproductions of ATP result pages are not redistributed.

Steps to reproduce

Extract the archive while preserving its directory structure. Run python code/validate_release.py to verify the released data and headline counts. Run python code/create_rally_figures.py to reproduce the manuscript figures from the frozen analysis outputs. Full model regeneration requires the source snapshots identified in data/provenance.csv; directory configuration is documented in README.md.

Institutions

• Istanbul Health and Technology University

  Istanbul, Istanbul

Categories

Licence