The nutrition math — what cooking does to the numbers¶

What this covers. A raw ingredient's nutrition label is a fact off a package; a cooked batch's nutrition is a computation — and cooking changes the numbers in ways that surprise people. Pasta gets less caloric per 100 grams when you boil it; a sauce gets more caloric when you reduce it; a fried potato picks up fat that was never "in" the potato; the salt you boiled the pasta in mostly goes down the drain. This chapter is the arithmetic of all of that: the one formula a batch's nutrition comes from, the two ways mass leaves a pot (only one of which takes nutrients with it), and worked examples — boiled spaghetti, rice simmered in chicken stock, drained brine, frying oil, pasta-water salt — showing exactly where each number lands.

1. One formula, two questions¶

Every computed nutrition figure for a cooked batch comes from a single fraction:

concentration  =  nutrients that ENTERED the food  ÷  weight of the FINISHED batch
(per 100 g)       (sum over the recipe's lines)       (the batch, weighed cooked)

The numerator and the denominator answer two different questions, and keeping them apart is the entire trick:

What entered the food? Only recipe lines can put nutrients in. Each line contributes its nutrient density (its profile per 100 grams) times the mass of it that actually ends up in the food.
How much food is there? The output weight — how much the finished batch weighs. A prep item's recipe already declares this (portions × portion size); it is the same single yield figure the batch's cost is divided by. There is no second notion of yield for nutrition.

Everything cooking does to the numbers — dilution, concentration, absorption, loss — is one of these two terms moving. The next three sections walk each case.

2. Water that was never an ingredient — dilution for free¶

Boil 400 g of dry spaghetti and you ladle out roughly 1 000 g of cooked spaghetti. The recipe has no "water" line — water costs nothing and carries no nutrients, so it has no business being an ingredient. And the math needs nothing from it:

line: dry spaghetti  400 g  @ 350 kcal/100g   →  1 400 kcal enter the pot
      (water: not a line — contributes nothing)
output, weighed:   1 000 g
─────────────────────────────────────────────────
1 400 kcal ÷ 1 000 g  =  140 kcal per 100 g

The 1 400 kcal that went in are now spread over 1 000 g instead of 400 g, so the concentration falls to two-fifths of the dry figure — same calories, more mass. The absorbed water never appears in the numerator (it brought no nutrients) and needs no line of its own; it shows up exactly once, in the weighed output. Un-listed water can only ever dilute.

3. Evaporation — concentration rises, and nothing needs recording¶

Run the same logic in reverse. Simmer a batch down — a stock reducing, a tomato sauce thickening — and water leaves as steam. Steam carries no calories. Every gram of protein, fat, carbohydrate, and sodium that entered the pot is still in it; there is simply less water spreading them out. The numerator is untouched; the denominator shrinks; concentration rises.

Worked example. Rice simmered in chicken stock, fully absorbed and reduced — no draining:
line: rice   400 g  @ 360 kcal/100g  →  1 440 kcal
line: stock 2 000 g @  15 kcal/100g  →    300 kcal   (all of it stays:
                                           evaporation takes water, not nutrients)
output, weighed:  1 600 g            →  1 740 kcal total
──────────────────────────────────────────────────────
1 740 ÷ 1 600 g  =  109 kcal per 100 g
2 400 g went into the pot; 1 600 g came out. The missing 800 g left as steam — and took nothing nutritional with it. Note the stock matters here precisely because it is an ingredient line (unlike water, it has a cost, sodium, and some calories of its own), so its full contribution rides into the numerator.

This is why evaporation never needs annotating: the weighed output already reflects the lost mass, and nutrients were never going to leave that way.

4. Draining and discarding — the only way nutrients leave¶

There is exactly one way for nutrients to leave a recipe's math: part of an ingredient line never ends up in the food. The liquid is poured off, the solid is fished out, or only a fraction was ever absorbed in the first place. For these lines — and only these — the recipe records the consumed mass: the grams of that line that actually enter the food, overriding the line quantity in the nutrition math. A line with no consumed mass recorded is simply counted in full, which is the right answer for the overwhelming majority of lines.

The recurring cases:

Case	Line says	Actually enters	Consumed mass
Drained brine or marinade	200 g	poured away	0 g
Aromatics cooked in, then removed (bay leaf, onion half)	50 g	flavor only	0 g
Frying oil	500 g in the fryer	only the absorbed share	≈ 15–75 g
Breading station	300 g of flour set out	what sticks	the pickup
Rice boiled in stock, excess poured off	2 000 g of stock	the absorbed share	≈ 600 g
Salt in the pasta water, water drained	5 g	the absorbed fraction	≈ 1 g

The last row is the everyday trap. Salt is a real ingredient line (it has a cost), but most of it leaves with the drained water — count it in full and the dish's sodium is badly overstated, on exactly the nutrient where overstating matters. The same drain that takes the water takes the salt; the consumed mass is what stayed with the pasta.

Frying oil runs the other direction and is the reason the concept is "consumed mass" and not "waste": the line puts 500 g in the fryer, and the question is not how much was thrown away but how much rode into the food — the absorbed 15 g for a light sauté, far more for breaded deep-fried items. Those absorbed grams carry oil's full ~885 kcal/100g density, which is why getting this one number roughly right moves a fried dish's figure more than any other input.

the two exits from a pot, side by side:
  EVAPORATION   mass ↓ as steam   nutrients stay    nothing to record
  DRAIN/DISCARD mass ↓ as liquid  nutrients LEAVE   consumed mass on that line

5. One batch, end to end¶

A 10-portion batch of fried rice, every rule above in one place:

line                          qty      consumed   kcal/100g   kcal in
rice (dry)                    1 000 g  all        360         3 600
chicken stock                 2 500 g  all        15            375   (reduces in)
frying oil                      400 g  60 g       885           531   (absorbed share)
soy-cured marinade, drained     300 g  0 g        90              0   (poured off)
scallions, ginger knob out      80 g   60 g       32             19   (knob removed)
─────────────────────────────────────────────────────────────────────
entered the food                                              4 525 kcal
output, weighed cooked:  3 200 g
concentration:  4 525 ÷ 3 200 g           =  141 kcal per 100 g
one portion (batch ÷ 10):                 =  453 kcal per serving

Reading the table: the stock counts in full (it reduced — evaporation), the marinade counts at zero (it drained), the oil counts at its absorbed share, and the un-listed boiling water from the rice never appears at all — it lives only inside the 3 200 g the batch weighed when finished. Sodium follows the same columns with its own densities, which is how the drained marinade's salt correctly stays out of the dish.

6. Three yields, kept straight¶

Three different "yield" ideas touch a recipe, and they answer different questions — collapsing them is the classic way kitchen math goes wrong:

Cost yield (EP/AP) is about value: peeling a 124 g carrot down to 100 g concentrates the carrot's cost onto the grams that survive. It belongs to costing and never touches a nutrient figure (see Recipes & production).
Output weight is about the whole batch's mass: water cooked in or out. It is the denominator here — and the same number cost-per-output divides by.
Consumed mass is about one line's mass entering the food: the drained brine, the absorbed oil.

The honest limits of the computed figure are part of the math, too — and the rule is to never paper over them. A computed profile is only as good as its inputs, and there are two different kinds of gap, handled differently:

A single missing nutrient on an ingredient that is otherwise fully accounted for — say the flour's sodium was never recorded — makes only the batch's sodium incomplete. Every other number is still exact, and the computation says "sodium unknown" rather than quietly filling in a zero.
A missing weight is worse, because it loses a whole ingredient. A batch counted in pieces — dumplings, patties, portioned sauces — needs a declared output weight before a per-100-gram figure means anything; the same holds for an ingredient measured by the piece or the milliliter with no known gram-equivalent. Without that weight the ingredient's entire contribution — every calorie, all its protein and sodium — is simply absent. So the figure is not published as a smaller-but-confident number: it is withheld, with a note naming the ingredient that still needs a weight. A confident calorie count that silently omits a whole component is worse than an honest "not yet computed" — especially on a label a guest reads.

Declaring the missing weight once fixes it for good: weigh the batch, give the piece its grams, and the figure computes in full from then on. Nutrition is optional — a kitchen that doesn't care to publish figures simply never declares those weights, and nothing else is held up by it. The figures are kitchen-honest arithmetic over declared facts: weigh the batch, mark the drained and absorbed lines once, and every recomputation thereafter is reproducible — no estimating, no guessing, no re-judging.