Corrections

This page is the public log of substantive corrections. When a nutrient number, verdict, or claim is shown to be wrong, the change is made and noted here with the date — the site does not quietly rewrite history.

2026-09-04 — Site launch. No corrections to date. The log starts here so that it starts as a habit: the first correction will appear below this line.

2026-09-15 — Verdict colours corrected in two articles (four rows). A rule-set audit of published pages found four rows whose colour did not follow the thresholds we publish. In Potatoes and Blood Pressure: home-prepared hash browns (533.5 mg sodium, 898.6 mg potassium per cup) were shown as green on the reasoning that the potassium column lifted the verdict — that was wrong, because 533.5 mg is above the 340 mg red line and the potassium pairing rule cannot lower a red verdict; the row is now red. Two mashed-potato rows in the same article (665.7 mg with milk and butter, 634.2 mg with milk alone) were labelled amber and are also above the 340 mg line; both are now red, and the surrounding text was corrected to match. In the canned soup case file, the low-sodium can (423.9 mg prepared) was labelled amber and is likewise above 340 mg; it is now red. No nutrient number was changed. Every figure was checked against the USDA-derived database — the numbers were right; the colour, and in one case the reasoning offered for it, was not.

2026-09-15 — Rule library v1.35: the rule set clarified, not changed in substance. The published methodology now states three things it previously left implicit or approximate. First, a verdict is produced only by sodium, potassium and the four veto triggers — saturated fat, sugar, fiber and calorie columns are context and do not move a verdict. Second, the potassium pairing threshold is exactly 350 mg per serving, replacing the earlier "roughly 400 mg+". Third, the pairing rule is explicitly one-directional: it can lift an amber food to green and can never rescue a red one. No data was edited in this change. Where a published page disagreed with the clarified rule set, the page was corrected — see the entry above.

2026-09-15 — Rule library v1.36, and three seaweed rows re-rated. A row-by-row audit of every published table turned up one gap in our own published rule set and one page that had been coloured on the wrong basis. The gap: the methodology page listed four veto categories, but ultra-processed carriers — foods whose sodium arrives in the formulation rather than the ingredient list — have always been vetoed in the rule library. The published list now matches the library and says five. The pickle veto was also stated too broadly: it now applies at or above 140 mg per serving, which is what our dill-pickle page already assumed when it rated a 6.3 mg reduced-sodium pickle green. The page error: the seaweed verdict rated three rows against the per-100 g value while printing the per-serving figure beside it. Raw kelp moves from amber to green (23.3 mg per tablespoon), raw wakame from red to green (87.2 mg), and dry tsunomata from red to amber (216.6 mg per cup). The article's conclusion is unchanged — the dried concentrates are where the sodium sits, fresh and rehydrated forms are not — and it now rests on the rows that carry it rather than on the labelling. The same pass removed a potassium-lift tag printed beside a row (dried agar, 86.7 mg) that was green on sodium alone, and corrected the Labor Day game plan, where sweet pickle relish (121.7 mg per tablespoon) was shown red while the dill-pickle page showed the same figure as amber. It is green on its number, and both pages now agree. No nutrient number was changed.