REJECTION-ABLATION-2026-v1: A Post-Rejection Follow-up Sampling (PRFS) Corpus and Per-Filter Counterfactual Ablation for Solana Memecoin Trading (June 2026)

REJECTION-ABLATION-2026-v1 is a public, reproducibility-grade Post-Rejection Follow-up Sampling (PRFS) corpus and the per-filter counterfactual ablation derived from it. The corpus contains 496,848 post-rejection price-follow-up samples covering 3,125 unique rejected tokens across 8 rejection filters, collected from a live paper-trading system on the Solana / pump.fun bonding-curve marketplace. PRFS instruments a trading system so that every rejected candidate is followed up and its subsequent price path is sampled, making the counterfactual ("what would have happened had we entered?") observable. This release applies PRFS to per-filter ablation: for each rejection filter, the realized entry-to-exit outcome of the tokens it rejected. Measured per-token with entry-to-exit returns and survivorship correction, all 8 filters reject median losers (median PnL -49% to -98%): the rejection logic is validated. A naive reduction of the same corpus inverts the conclusion, which the accompanying paper decomposes into four measurement biases. Contents: the gzipped corpus (rejection_outcomes.jsonl.gz, ~46 MB), a dependency-free streaming analyzer (rejection_ablation.js) that reproduces every number in the paper, four per-filter result tables (main + three appendices), and three figures. Data is released under CC-BY-4.0; code under MIT. Methodology reference (PRFS): SSRN Working Paper 6607301. Companion paper: "Counterfactual Rejection Filter Ablation: An Empirical Application of PRFS to Solana Memecoin Trading" (arXiv + SSRN, in submission).