Corrections
Every correction to a figure or a claim on this site is logged here: what was wrong, when it was published, and when it was fixed. Corrections are made on the page itself as well — a note on this page does not correct the page a reader is actually looking at.
What was wrong. Each seller row counted the figures that seller publishes and printed the total as “N readings” — so 81 of the 106 rows told a reader this site had read that seller two or more times, and 40 claimed four or more. Every seller in this index has been read exactly once.
What it says now. The row states how many figures the seller publishes and the date they were read, in the summary a reader sees without opening the row.
What was wrong. The one-line introduction promised “the date every figure was read and what it changed from”. This site holds no price-change record; nothing in it dates a price change.
What it says now. The introduction states the reading date and says outright that no price change can be shown yet.
What was wrong. Both pages stated that sellers listed here pay this site a commission. No seller in the index pays this site anything, and /disclosure’s own computed figures said so six lines below the sentence.
What it says now. One plain sentence, in the footer of every page: the site is reader-supported and may earn an affiliate commission. The count of sellers that actually pay is computed on /disclosure.
What was wrong. Three phrases were presented as direct quotations from the NIH Office of Dietary Supplements niacin fact sheet without being in this site’s source ledger. They are not alleged to be inaccurate — they were unverifiable here, on the one cited host that cannot be re-fetched from this desk.
What it says now. The three quotations are withdrawn and the withdrawal is stated on the page. The claim about supplemental versus dietary niacin is kept as this site’s own sentence; the quotation that is in the ledger is kept verbatim.
What was wrong. Under the heading “No paid links on this page”, the page said this tool page “carries a paid link, a seller button or a tracked outbound hop of any kind” — a dropped negation. The page carries none of those, which the build check confirmed in the same build.
What it says now. The sentence states the negative, and the negation now sits outside the singular/plural branch so it cannot be dropped from one of them again.
What was wrong. “Routes covered” counted 9 by including “unspecified”, which is the absence of a stated route rather than a route. /disclosure listed the eight route names on the same data.
What it says now. Both surfaces exclude “unspecified”, and the count is 8.
What was wrong. Both pages said a row with a single reading “reads ‘first reading’”. No row on this site printed that label; rows print the month they were read.
What it says now. Both pages describe the label a row actually prints.
What was wrong. Every board told readers that payment cannot move a row because “a build check recomputes every board with the payment flag switched on for every seller”. The claim was true; the build check did not exist.
What it says now. The build check exists, runs on every build, and fails it if marking all 106 sellers as paying changes any board’s membership, any board’s order, or any seller’s score.
Report something
contact@nadprices.com. A screenshot of the seller’s page with the date visible is the most useful thing you can send.