Comparing two foods without being misled: the method

Four traps before any comparison

The FSVO database describes 1,216 generic foods across 19 families, with up to 22 components each, always per 100 g. That shared basis is what makes a comparison possible, under four conditions.

• The preparation state. Raw, dried, cooked, drained: each is its own row. 100 g of raw polished rice holds 352 kcal and 12.3 g of water; boiled, the same rice drops to 117 kcal with 71.2 g of water.
• The reading basis. A gap is read per 100 g; the portion selector only recalculates the same value for another weight.
• The “n/a” cells. Value not available in the source, never a zero.
• The origin of the figure. A, E or C applies per cell and can change from one row to the next.

In the list of foods, the state appears in the label of every entry.

The method, applied to salmon and tuna

Both entries sit in the same category, “Sea fish”, and in the same state, raw: the comparison is between two foods, not between two cooking methods.

Raw tuna shows more protein; raw farmed salmon shows more fat, and the widest gap is in omega-3 fatty acids: 2.1 g against 0.14 g. Energy follows the fat. None of these four rows summarises the other three. Origin is worth a look too: on the salmon side, protein is measured (A) and fat estimated (E); on the tuna side, both values are measured. The ranking of foods highest in omega-3 places these contents within the whole database.

One food, two rows: dried lentils, cooked lentils

Whole dried lentils show 313 kcal, 24.4 g of protein, 23.2 g of dietary fibre and 8 mg of iron per 100 g. Whole cooked lentils, without added fat or salt, come to 111 kcal, 8.8 g of protein, 8.5 g of fibre and 2.5 mg of iron.

Nothing vanished: water came in. The dried row holds 9.7 g of water per 100 g, the cooked row 68.7 g. At equal weight, far less dry matter is left. A second, less visible difference: in the dried row, protein, fibre and iron are measured (A); in the cooked row they are calculated (C). Comparing dried with cooked stays legitimate, provided both states are named. The ranking of foods highest in protein contains both rows: the label says which.

“n/a” and A/E/C: what the cell actually says

“n/a” means the value is not available in the source: the FSVO has not published it for that food, and this site replaces it with no in-house estimate. An “n/a” cell is therefore not a zero. If either food carries that mark on a row, the row drops out of the comparison: it reads neither as an absence nor as an advantage.

The origin indicator travels with every single value: A for a value from analytical measurement, E for a value estimated from reliable sources, C for a value calculated using recognised rules, for instance from the ingredients of a recipe. It applies per cell, not per food. When two values sit close together, knowing that one is measured and the other estimated changes how much the gap is worth. All three indicators are set out in the methodology.

What the comparison tool shows, and what it does not

The food comparison tool places two entries side by side, per 100 g, across every available row: energy, protein, fat, carbohydrates, fibre, minerals, vitamins. For each row the higher of the two contents is flagged and the gap is computed as the page is displayed. “n/a” stays visible as “n/a”. The category and state of both foods are shown alongside, so comparability can be checked at a glance.

What it does not do: add those rows up into a single grade. No composite score is computed and no “best food” is named. An entry with more protein and more fat than another remains exactly that: two facts, not a verdict. The rankings follow the same rule, with one nutrient at a time and the sorting method published.