Calorie Counting While Cooking: The Per-Ingredient Logging Method

Home cooking is where calorie tracking is theoretically easiest. You control every ingredient. You have a scale. You know exactly what went in the pot. In practice, most people abandon precision by the third ingredient and enter “pasta with tomato sauce, homemade” as a single line item in a generic database — accepting whatever figure the app suggests for a dish that may be nothing like what they made.

The failure mode is choosing the wrong moment to measure. Tracking cooked food is harder than tracking raw ingredients, because cooked food has lost water, absorbed oil, and changed density in ways that vary by temperature, duration, and cooking method. A cup of cooked rice weighs differently depending on how much water the grain absorbed. A chicken breast loses 20–30% of its raw weight when baked, but the caloric content of the meat doesn’t disappear with the water — it concentrates. Measuring cooked food without a reference for the specific cooking loss introduces systematic error that compounds across multiple components in a recipe.

The per-ingredient logging method flips the sequence. Instead of measuring the finished dish, you measure and log each ingredient before it enters the pan. The total calorie count for the meal is the sum of the raw ingredient inputs, divided by the number of portions served. This method is faster, more accurate, and more reusable than any post-cook estimation approach. Once you understand how it works and why it outperforms the alternatives, building a recipe template takes under five minutes.

Why raw beats cooked for calorie accuracy

Nutrition databases — including USDA FoodData Central, the reference for most credible food tracking apps — provide calorie values for foods in a defined state. For most raw ingredients, that state is unprocessed: raw chicken breast, dry pasta, uncooked oats, raw potato. For others it’s the processed state: canned tomatoes, olive oil, cooked beans.¹

The problem with measuring cooked food and looking it up as “cooked” is that cooking loss percentages vary more than most people assume. The USDA provides a yield factor table — a correction factor for water loss, fat gain or loss, and other changes during cooking. These factors differ by cooking method and duration:

• A chicken breast baked at 180°C for 25 minutes loses approximately 25% of its raw weight in moisture.
• The same breast pan-fried in oil for 12 minutes may lose 20% moisture while absorbing 5–10 g of oil, changing net calorie content in the opposite direction.
• Vegetables lose 5–60% of their water content depending on roasting temperature and cut size — a whole roasted carrot behaves very differently from matchstick-cut carrots on a high-heat sheet pan.

These differences are not captured when you enter “chicken breast, cooked, 150 g” into a database. You’re applying a single cooked-state calorie density to a preparation that may have gained or lost significant energy content relative to that reference.²

Raw ingredient measurement bypasses this entirely. Chicken breast, raw, has 165 kcal per 100 g regardless of what you do to it next. The oil you add is measurable and loggable before it hits the pan. The total calorie count is the sum of raw ingredients, and it remains accurate across cooking methods because you measured before the transformations occurred.

The five-minute recipe template method

A recipe template is a saved log of raw ingredient weights for a dish you cook repeatedly. Once built, it applies to every future cook with a single tap. The five-minute estimate is accurate: building a template for a four-ingredient curry takes less time than looking up each ingredient individually in a database.

Step 1: Gather and weigh before you start cooking. Place your prep bowls on the scale and tare to zero. Weigh each ingredient into its bowl or directly into the pan before heat is applied. As you weigh each item, log it immediately in your tracking app — weight in grams, food item matched to the database. Don’t wait until the end of cooking to reconstruct what went in. Memory underestimates oil and sauce additions by 30–50%.³

Step 2: Log oils and fats with a dedicated spoon or measure. Oil poured directly from a bottle is among the hardest ingredients to estimate accurately. One technique: zero the scale with your pan on it, add oil, read the weight before adding anything else. One gram of olive oil is 9 kcal. Underestimating 10 g of oil means missing 90 kcal — significant across a week. Alternatively, measure with a tablespoon or teaspoon (one tablespoon olive oil ≈ 13 g ≈ 119 kcal) and log the measured volume.

Step 3: Note the total raw weight of all ingredients. This is your yield basis. After cooking, the dish will weigh less (due to water loss) than the sum of raw weights. If you want to serve by weight of the cooked dish, you need to know the ratio of cooked to raw — typically 0.65–0.85 for meat-based dishes, 0.70–0.90 for grain and legume dishes.

Step 4: Determine the number of servings and divide. If the dish yields four portions and you eat one portion, divide the total raw-ingredient calorie sum by four. This is your meal calorie count. You don’t need to weigh your individual cooked serving — just know your fraction of the total batch.

Step 5: Save as a recipe in your tracking app. Name it specifically (“red lentil dal — 4 servings, my recipe”) rather than generically (“lentil soup”). Specific names prevent you from accidentally applying the template to a different preparation of the same dish with different ingredient ratios.

How to handle variable ingredients

Recipes are rarely identical across cooks. The onion is smaller than last time. The chicken portions varied. You added an extra handful of spinach. These variations don’t require rebuilding the template from scratch — they require adjusting for the variable ingredients only.

Fixed-ratio ingredients: Spices, herbs, baking powder, and condiments in fixed recipe quantities contribute few enough calories that variation of a few grams is negligible. Log these at the template weight.

Variable-ratio ingredients: Proteins, grains, oils, and starchy vegetables are where serving-to-serving variation matters. Weigh these fresh for each cook. In practice, you’ll find your portions stabilize quickly because the act of weighing trains consistent portioning.

Additions and substitutions: Log additions as separate line items when you cook. If you swapped Greek yogurt for cream in a pasta sauce, delete the cream entry and add the yogurt entry with its weighed amount. Template maintenance takes 30 seconds.

The per-ingredient method’s major advantage over recipe databases is that it reflects what you actually made. Generic database recipes assume ingredient proportions that may not match your cooking. A “chicken tikka masala” in a database might assume 2 tablespoons of oil per serving; your version might use 1 tablespoon. That discrepancy, across three weekly cooks, is an untracked 1,000 kcal difference.

Managing cooking oils accurately

Oil is the most underlogged calorie source in home cooking. Research on self-reported dietary intake consistently finds that fat and oil intake is underreported at higher rates than any other macronutrient — in some studies, reported oil consumption is less than half of measured intake.³

The mechanisms are straightforward: oil is poured in a continuous stream, not in discrete units. It’s invisible in the finished dish. It coats ingredients at the surface, so visual inspection of the cooked food doesn’t reveal the quantity used. And oil is high-calorie by volume — a single tablespoon is 120 kcal, which is more than most people guess.

Three practical techniques for accurate oil logging while cooking:

Scale-zero method: Place pan on scale before heat. Add oil while watching the weight display. Log the measured grams before turning on heat. This is the highest-accuracy method and takes under 10 seconds.

Measured pour: Use a dedicated tablespoon measure rather than pouring from the bottle. Measure into the spoon, add to pan, log the spoon measurement. More accurate than an unguided pour, less accurate than the scale method.

Oil mister calibration: If you use an oil mister/sprayer, pump it 10 times into a small bowl on a scale and weigh. Divide by 10 to get grams per pump. Log pumps used when cooking. This works well for high-heat roasting where misting is the oil delivery method.

The spray-oil category labeling deserves particular attention. Some “zero-calorie” cooking sprays carry a label claim based on a serving size of 0.25-second spray — a quantity that delivers approximately 0.3 g of oil. In practice, a realistic spray duration for coating a pan is 1–3 seconds, delivering 1–4 g of oil (9–36 kcal). The label is not lying; the serving size is just implausibly small. Log actual spray duration fractions or use the gram method to avoid a systematic zero-calorie misattribution.

Batch cooking and portion accounting

Batch cooking — preparing large quantities of a base ingredient (rice, roasted vegetables, cooked protein) for use across multiple meals in the week — creates a specific tracking challenge. The total raw ingredient weight is logged at cook time, but individual portions are extracted throughout the week without being weighed at extraction.

The cleanest solution is to establish a per-gram calorie density at cook time and apply it to weighed portions at meal time. Calculate this as:

Calorie density (kcal/g cooked) = Total raw ingredient calories ÷ Total cooked batch weight

Measure the cooked batch weight immediately after cooking (weigh the container, zero, add food, read weight). Then, at each extraction, weigh your portion and multiply by the calorie density. This remains accurate regardless of how many portions you extract over the week because each is measured against the cooked-state reference weight.

For dishes where reheating with added liquid changes the cooked weight (soups, porridges, some stews), remeasure the batch weight after each addition and recalculate the density. In practice, most batch-cooked dishes don’t require this unless the liquid addition is substantial.

When to use a photo log instead of per-ingredient tracking

The per-ingredient method is optimal when you control the cooking. It is not applicable to food cooked by someone else, food assembled from partially processed components (a takeaway wrap with ingredients you can’t weigh), or food served at a meal where weighing would be socially disruptive.

Photo logging fills these gaps. For meals you didn’t cook, a photograph of the plated dish captures the composition and portion geometry in a way that a text search cannot. CalEye processes the photograph against USDA FoodData Central references to produce a per-component breakdown with explicit uncertainty ranges.⁴

The two methods are complementary, not competitive. Use per-ingredient logging when you cook. Use photo logging when you don’t. A hybrid workflow — per-ingredient for dinner, photo logging for the lunch you bought, manual entry for the coffee and snacks — covers the full day with the most accurate method for each occasion.

Building the five-minute template: a worked example

Here is the full workflow for building a recipe template using a simple meal: one portion of pan-fried salmon with roasted broccoli and brown rice.

Ingredients weighed raw before cooking:

• Salmon fillet: 200 g (raw) → 330 kcal
• Olive oil for salmon: 10 g → 90 kcal
• Broccoli: 200 g (raw) → 68 kcal
• Olive oil for roasting: 8 g → 72 kcal
• Brown rice, dry: 75 g → 272 kcal
• Soy sauce: 15 g → 9 kcal
• Garlic (minced): 5 g → 7 kcal

Total raw calorie sum: 848 kcal. Servings: 1 (entire recipe). Logged meal total: 848 kcal.

This takes under four minutes to build, including weighing. The salmon is the only item that varies materially — fillet sizes typically run 150–250 g. Save the template and adjust the salmon entry by the actual gram weight for each cook. The oil quantities are measured each time (30 seconds on the scale). Everything else is close enough to template weight that the variation is below 10 kcal.

The next time you cook this meal, it takes 90 seconds: open the template, adjust the salmon weight, confirm the oil measurements you just took, log. That’s the payoff for the five minutes you spent building it the first time.

References

1. U.S. Department of Agriculture, Agricultural Research Service. FoodData Central. Nutritional data for raw vs. cooked food states. https://fdc.nal.usda.gov/

2. USDA Agricultural Research Service. “USDA Table of Cooking Yields for Meat and Poultry.” Revised 2012. https://www.ars.usda.gov/

3. Dhurandhar NV, Schoeller D, Brown AW, et al. “Energy Balance Measurement: When Something Is Not Better than Nothing.” International Journal of Obesity 39, no. 7 (2015): 1109–1113.

4. CalEye. “Photo Recognition Methodology and USDA FoodData Central Integration.” Internal documentation, 2024.

5. Wansink B, Chandon P. “Slim by Design: Redirecting the Accidental Drivers of Mindless Overeating.” Journal of Consumer Psychology 24, no. 3 (2014): 413–431.