# Australian PLR 2024–25 grouped payment calculations Python 3.10 or newer, standard library only. In a clean folder containing `reproduce.py` and `selected-source-payment-bands.csv`, run: ``` python reproduce.py --input-dir INPUT_FOLDER --output-dir OUTPUT_FOLDER ``` Use separate input and output folders. Nothing is downloaded, and no source input is changed. The script writes five derived files: all 45 band distributions and cumulative grouped points; three reported-class threshold comparisons; a summary; conditional top-tail identification intervals; and synthetic mathematical formula-test results. Exact fractions are preserved. Display percentages are rounded to two decimal places using decimal half-up arithmetic. The outcome population is the official source's **Authors-column paid claimants** in Australian PLR 2024–25, not all writers, eligible/registered nonpaid people or publisher claimants. Publisher and combined source columns are retained to verify the factual extraction only. The source tables are selected factual inputs, not individual claimant records or the original PDF. Read `methods.md`, `rights.md` and `sources.json` for definitions, assumptions and attribution. A deterministic conditional identification interval is not a sampling confidence interval. No Gini, midpoint payment records, causal cap comparison or income advice is produced.