Most nutrition advice is too vague to be truly helpful.

“Eat healthy” does not tell you whether salmon is a better default than chicken, whether lentils beat rice, or berries are better than other fruits. Calories and macros matter, but they do not tell the whole story. You can hit both targets with nutrient sparse foods.

The more useful question is:

How much nutrition comes with each calorie?

This system answers it in two steps.

First, look at the food itself: how close is it to its original form, and how much industrial processing was required to make it? Then look at what it delivers: protein, fiber, vitamins, minerals, fats, and all other things that make a food useful as a daily default.

Processing sets the starting range. Nutrient density decides the ranking within it.

Start with processing

Group	What it means	Examples
NOVA 1	Unprocessed or minimally processed	Vegetables, fruit, eggs, fish, meat, beans, plain milk
NOVA 2	Ingredients used to cook with	Olive oil, butter, sugar, honey, cream
NOVA 3	Processed foods made from tier 1 and 2 ingredients	Cheese, canned fish, bread, tofu, olives
NOVA 4	Industrial formulations of refined ingredients	Soda, sweets, crisps, packaged cakes, cured meats

The tiers are not moral categories. They are a useful starting heuristic.

Tofu is more processed than an egg but can still be a strong food. Canned fish is processed but remains nutritionally dense. Fortified corn flakes can contain more added vitamins than fruit while still being built on refined grain. The tier matters because food structure and formulation matter. It does not settle the question by itself.

Then count what it delivers

Within a tier, the ranking is driven by nutrient return.

A sardine carries protein, omega-3 fats, calcium, vitamin D, vitamin B12, and selenium. A blueberry carries fiber, vitamin C, vitamin K, and plant compounds. Both are good foods. They are not equally dense.

The model asks how much useful nutrition a food provides in two ways:

Per 100 calories: what does this food give back for the calories it costs?

Per realistic serving: what does this food give back in an amount a person is likely to eat?

Those two views are blended. Per-calorie scoring alone rewards watery foods in unrealistic amounts. Per-serving scoring alone rewards large portions. Together, they produce a more grounded comparison.

The score rewards:

Vitamins, minerals, protein, and fiber

Nutritional breadth: many nutrients beat an extreme amount of just one

Food that delivers more nutrition for fewer calories

The score subtracts:

Added sugar

Sodium

Saturated fat

The nutritional cost of an industrial formulation

Natural sugar in fruit and milk is not treated like added sugar.

The scores

These scores are best used for substitution.

Compare proteins with proteins, vegetables with vegetables, fruits with fruits, oils with oils, etc. A score tells you what each food brings for its calories. It does not mean you should eat nothing but spinach because spinach scores highly per calorie.

Vegetables

Score	Food	Tier	Rationale
96	Spinach	1	Folate, vitamin K and magnesium at very few calories.
96	Watercress	1	Vitamin K and carotenoids, with almost no calories to offset them.
94	Asparagus	1	Folate and vitamin K, with fibre.
94	Collard greens	1	Vitamin K and calcium.
91	Kale	1	Carotenoids and vitamins C and K.
89	White mushrooms	1	B vitamins, copper and selenium — unusual among vegetables.
89	Brussels sprouts	1	Vitamin C and K with fibre.
88	Red bell pepper	1	More vitamin C per calorie than citrus.
84	Broccoli	1	Vitamin C, folate and fibre.
83	Cabbage	1	Vitamin C and K at low cost.
82	Cauliflower	1	Vitamin C and choline.
81	Tomatoes	1	Potassium and vitamin C. Carotenoids concentrate when cooked.
73	Sweet potato	1	Dense in carotenoids, on more starch than most vegetables.
73	Carrots	1	Carotenoids for vitamin A, plus fibre.
71	Potato, with skin	1	Potassium and vitamin C when the skin is eaten.
70	Garlic	1	Manganese and B6, but eaten in very small amounts.
68	Beets	1	Folate and nitrates, with intrinsic sugar.
67	Onions	1	Modest density. Used for flavour more than nutrition.
39	Olives	3	Monounsaturated fat. Canning brings sodium.
31	Sauerkraut	3	Fermented cabbage, with the sodium that preserves it.
30	Dill pickles	3	Sodium dominates what is otherwise a cucumber.

Fruit

Score	Food	Tier	Rationale
82	Blackberries	1	Fibre and vitamin C well above most fruit.
79	Strawberries	1	Vitamin C at low calorie cost.
77	Raspberries	1	Among the highest fibre of any fruit.
73	Orange	1	Vitamin C and folate.
73	Mango	1	Carotenoids and vitamin C, with more sugar.
69	Blueberries	1	Vitamin K and manganese. Less dense than darker berries.
67	Banana	1	Potassium and B6, carried on more starch.
65	Grapes	1	Mostly water and sugar.
62	Apple, with skin	1	Fibre in the skin. Thin on micronutrients.
60	Avocado	1	Potassium and folate, diluted by its own fat.

Protein

Score	Food	Tier	Rationale
97	Beef liver	1	Exceptional across B12, copper, folate and iron.
90	Oysters	1	Zinc, B12 and copper in unusual concentration.
81	Wild salmon	1	Omega-3s, vitamin D and B12.
78	Canned tuna, in water	1	Protein, selenium and B12. Canning costs little.
75	Canned sardines, with bones	1	Omega-3s and vitamin D, plus calcium from the bones.
74	Cod	1	Lean protein with B12 and selenium.
71	Chicken breast	1	Lean protein, niacin and B6.
65	Tofu, firm	3	Calcium-set soy protein.
62	Ground beef, 90% lean	1	Zinc, iron and B12, with saturated fat.
61	Pork loin	1	Thiamin and protein, with saturated fat.
61	Egg	1	Choline and B12. The fat dilutes density per calorie.
60	Tempeh	3	Fermented soy. More protein and B vitamins than tofu.
50	Shrimp	1	Selenium and B12. The sodium comes from processing.
9	Beef bologna	4	Sodium and saturated fat with modest protein.
8	Beef frankfurter	4	Protein carried on sodium and saturated fat.

Legumes and grains

Score	Food	Tier	Rationale
83	Lentils	1	Folate, fibre and protein.
79	Chickpeas	1	Fibre, folate and manganese.
78	Black beans	1	Fibre and folate.
73	Quinoa	1	Magnesium and protein, with complete amino acids.
72	Brown rice	1	Manganese and magnesium, retained in the bran.
68	White rice	1	Refining removes most of the bran's minerals.
68	Oats	1	Beta-glucan fibre and manganese.
52	Whole-wheat bread	3	Fibre and B vitamins the white loaf loses.
48	White bread	3	Enriched B vitamins, little fibre.
39	Corn flakes, fortified	4	Fortification supplies the vitamins. The grain is refined.

Nuts and seeds

Score	Food	Tier	Rationale
81	Sunflower seeds	1	Vitamin E, selenium and magnesium.
70	Chia seeds	1	Fibre and calcium. Measured on a partial panel.
68	Almonds	1	Vitamin E and magnesium, at high calorie density.
65	Flaxseed	1	Fibre and plant omega-3s, eaten in small amounts.
62	Walnuts	1	Plant omega-3s. Fat dilutes the rest.
47	Peanut butter, unsalted	3	Niacin and magnesium, dense in calories.

Dairy

Score	Food	Tier	Rationale
53	Whole milk	1	Calcium, B12 and iodine, with saturated fat.
50	Plain yogurt, whole milk	1	Calcium and B12, offset by saturated fat.
40	Cottage cheese	3	Protein and calcium, with sodium.
39	Mozzarella, whole milk	3	Calcium and protein, with saturated fat and salt.
38	Cheddar	3	Calcium, but concentrated saturated fat and sodium.
6	Vanilla ice cream	4	Sugar and saturated fat with a little calcium.

Cooking fats and sweeteners

Score	Food	Tier	Rationale
31	Canola oil	2	Vitamin E and K, with little saturated fat.
29	Olive oil	2	Vitamin E and monounsaturated fat.
21	Lard	2	Some vitamin D. Mostly saturated fat.
20	Heavy cream	2	Vitamin A carried on saturated fat.
20	Maple syrup	2	Manganese, but overwhelmingly sugar.
18	Butter, unsalted	2	Vitamin A and concentrated saturated fat.
18	Brown sugar	2	Trace minerals do not change what it is.
18	Granulated sugar	2	Calories with nothing attached.
15	Honey	2	Trace minerals, but sugar per calorie all the same.
15	Margarine	4	Fortified vitamins on industrial fat.
14	Coconut oil	2	Almost entirely saturated fat.

Packaged snacks and drinks

Score	Food	Tier	Rationale
20	Potato chips	4	Potassium from the potato, undone by salt and fat.
19	Saltine crackers	4	Refined flour and salt.
17	Corn chips	4	Refined corn, fat and salt.
13	Chocolate chip cookies	4	Sugar and fat with trace minerals.
8	Doughnut, plain	4	Refined flour fried in fat.
5	Vanilla frosting	4	Sugar and fat, and essentially nothing else.
4	Milk chocolate	4	Some magnesium and iron, dominated by sugar and fat.
3	Cola	4	Sugar and water. The lowest density measured here.

Build meals, not a collection of high scores

A useful meal combines strengths.

Salmon brings protein, vitamin D, B12, selenium, and omega-3 fats.

Lentils bring fiber, folate, potassium, and slow-digesting carbohydrate.

Vegetables add volume, fiber, and micronutrient range.

Olive oil can improve the meal without becoming its nutritional foundation.

Build most meals from Tier 1 foods
Use Tier 2 foods to cook them
Judge Tier 3 and Tier 4 foods by what ingredients they contain

How the score works

The calculation is deliberately simple.

Each qualifying nutrient is measured as a percentage of its FDA Daily Value, capped at 100%, then averaged across the nutrient panel. Micronutrients receive full weight; protein and fiber receive reduced weight.

No single nutrient should win the entire argument. Beef liver contains far more than a daily value of vitamin A. Without a cap, that one number would distort the ranking. Each qualifying nutrient therefore counts up to its daily value and no further.

That makes breadth matter more than spikes. A food that provides a useful amount of protein, fiber, folate, magnesium, potassium, vitamin C, and iron should outrank one that supplies an extreme amount of one vitamin and little else.

Saturated fat, sodium, and added sugar are then subtracted. Natural sugar in fruit and milk is not treated like added sugar.

The result blends nutrient return per 100 calories with nutrient return per realistic serving. It is then placed inside the range allowed by its processing tier. Processing is not a veto; it is a constraint.

The math

For readers who want the mechanism.

Each qualifying nutrient is scored as a percentage of its FDA Daily Value, capped at 100%, then averaged across the whole panel of 23: 21 micronutrients at full weight, plus protein and fiber at 0.6.

Q = \frac{\sum_i w_i \cdot \min\!\left(100,\; 100\,\frac{a_i}{\mathrm{DV}_i}\right)}{\sum_i w_i}

The cap matters. It prevents one extreme nutrient value from carrying an otherwise thin food. An unmeasured qualifying nutrient contributes nothing while the denominator still counts the full panel, so incomplete data act as a floor rather than an advantage.

The limit nutrients—saturated fat, added sugar, and sodium—are scored against maximum recommended values. They are averaged only across values that were measured, so a known downside is never diluted by fields nobody checked.

L = \frac{1}{\lvert M \rvert} \sum_{j \in M} \min\!\left(100,\; 100\,\frac{a_j}{\mathrm{MRV}_j}\right)

Term	What it is	Example
Q	Mean capped % of daily value across the qualifying panel	Spinach 50.1 per 100 kcal
L	Mean capped % of maximum recommended value across the limits	Cheddar 10.2 per 100 kcal
Net	Q minus L, blended 60% per 100 kcal and 40% per serving	Spinach 28.4, cola -19.0

Net density is qualifying nutrition minus nutritional liabilities, measured on both bases and blended. The 60/40 mix favors nutrient return per calorie while keeping the result tied to a plausible amount of food.

\mathrm{net} = 0.6\,(Q - L)_{100\ \mathrm{kcal}} \;+\; 0.4\,(Q - L)_{\mathrm{serving}}

The final logistic curve places that net value inside the range allowed by the food’s processing tier. Its fitted constants spread the observed foods across each tier without forcing artificial 100s or 0s. Displayed scores are rounded to whole numbers, so foods close together can print the same score even when the underlying model distinguishes them.

\mathrm{score} = \mathrm{floor} + \frac{\mathrm{width}}{1 + e^{-(\mathrm{net} - 5.278)/8.249}}

What the numbers leave out

The score is a ranking tool, not a medical diagnosis.

It does not include allergies, affordability, food safety, medical conditions, personal goals, recipe variation, or total daily intake. A food absent from these tables is unscored, not necessarily low-scoring.

Source: nutrient data from USDA FoodData Central SR Legacy records.