Raw or Cooked? The Weighing Mistake That Hides 300 Calories a Day

TL;DR

• Meat loses 20-30% of its weight to water during cooking, rice and pasta gain 100-150%, so the same food has two very different weights.
• Weighing cooked chicken and logging it against a raw database entry under-counts by roughly 25-30%.
• Weighing cooked pasta and logging it against a raw entry over-counts by a similar margin in the other direction, which is why some people stall and others can’t gain.
• Pick one convention per food and stick to it: raw for meat and dry for grains is the easiest to keep honest.
• A 150 g chicken breast and a 150 g portion of dry rice are not the same meal cooked, and your log should know the difference.

You weigh your food. You log it properly. You’ve been in a deficit for six weeks and the scale has barely moved. Before you blame your metabolism, check which weight your database entry refers to, because this single detail moves more calories than almost anything else people get wrong. It’s one of the common macro tracking mistakes that never gets flagged because the log looks perfectly precise, it’s just precisely wrong.

Water is the whole story

Cooking does one thing to weight: it moves water. Nothing else meaningfully changes. The protein, fat and carbohydrate in a chicken breast are the same before and after you cook it. What changes is how much water sits alongside them.

Grill a chicken breast and it loses water, so the same protein is packed into less weight. Boil rice and it absorbs water, so the same carbohydrate is spread across more weight. The macros per 100 g swing dramatically in both directions even though the food itself is unchanged.

Typical shifts

• Chicken breast, grilled or pan-fried: loses 20-30% of raw weight
• Beef mince, browned and drained: loses 25-35%, more if it’s higher fat and you drain the rendered fat off
• Salmon fillet, baked: loses 15-20%
• White rice, boiled: gains 150-200% over dry weight
• Dried pasta, boiled: gains 100-130% over dry weight
• Lentils and dried beans, boiled: gain 120-150%

Those ranges are wide on purpose. How hot you cook, how long, and whether you cover the pan all shift the number. That variability is exactly why the raw or dry weight is the more stable reference point.

The maths of getting it backwards

Take a raw 200 g chicken breast, roughly 46 g protein and 220 kcal. Grill it and it comes off the pan at about 150 g. If you weigh those 150 g and log them against a raw database entry, you record 35 g protein and 165 kcal. You’ve just lost 11 g of protein and 55 kcal from your log on one item, which makes it that much harder to hit your protein goal every day even when you think you’re eating enough.

Repeat that across two protein servings a day, add a cooked-weight logging error on mince at dinner, and you are under-recording by 150-250 kcal daily without a single dishonest entry. Over a week that is roughly a full day of food unaccounted for, which is enough to turn a 500 kcal deficit into something closer to 300, and enough to explain a stalled weight loss plateau that has nothing to do with metabolic adaptation.

Carbohydrates run the error in reverse and the consequences are less obvious but just as real. Weigh 200 g of cooked rice, log it against a dry entry at 350 kcal per 100 g, and you’ve recorded 700 kcal for a portion that actually contains around 260. Your log now says you ate far more than you did. People in this situation cut portions they didn’t need to cut, feel constantly under-fuelled, and conclude that tracking makes them miserable.

Pick a convention and hold it

There is no correct answer between raw and cooked. There is only a consistent answer. Two workable systems:

System one: weigh everything raw or dry

Weigh the chicken before it hits the pan. Weigh the rice dry, before water. Log against raw and dry database entries, which is what most entries default to anyway.

This is the most accurate approach because raw weight doesn’t depend on your cooking method. The downside is that it only works when you’re cooking for yourself from the start, and it needs a bit of planning: once the food is cooked, the data is gone. This is where a bit of meal prep pays off, since raw weighing happens once at the start of the week instead of at every meal.

System two: weigh cooked, log cooked

Use database entries explicitly labelled “cooked”, “boiled”, “grilled” or “roasted”. These exist for most common foods and are perfectly usable.

The trade-off is precision. A cooked entry assumes an average water loss that may not match yours. Someone who grills chicken hard and dry will have less water in the finished product than the database assumes, so their cooked entry under-counts slightly. It’s a smaller error than mixing conventions, but it’s there. If you’re logging without a scale at all, the same logic applies loosely to tracking macros without weighing food: pick a reference point and stay consistent with it.

The rule that keeps it simple

Raw for meat and fish. Dry for rice, pasta, oats and pulses. Cooked for anything you didn’t cook yourself.

That last one matters. If someone hands you a plate, raw weight isn’t recoverable, so a cooked entry is the only honest option. Mixing conventions across different foods is fine; mixing them within the same food is what causes drift. It’s the same principle behind tracking homemade meals and recipes: consistency in how you log matters more than chasing perfect precision on any single entry.

The batch-cook workaround

Weighing raw falls apart when you cook 1.2 kg of chicken on Sunday and eat it across five meals. You can’t weigh raw portions that don’t exist yet.

The fix is a two-step weigh. Log the total raw weight of what goes into the oven, say 1,200 g of chicken. When it comes out, weigh the whole cooked batch, say 880 g. You now know the yield: every 100 g of cooked chicken represents about 136 g raw. Portion out 176 g cooked and log it as 240 g raw.

Do the same with a pot of rice or a chilli. Weigh the dry ingredients going in, weigh the finished pot, and you have a conversion factor for the whole batch. It takes two extra weighings on cook day and removes the guesswork from every meal that follows.

Where photo logging fits

A photo scanner sees a cooked plate, because that’s what plates look like. When you photograph a finished meal in Chowdown, the estimate is working from the cooked portion in front of it, which is the right reference for that image. Knowing how AI food scanning actually works helps explain why: the model is reading what’s on the plate, not guessing at what went into the pan before you cooked it.

That makes photo logging genuinely useful for the meals where raw weight is unavailable: someone else’s cooking, a takeaway, a plate at a restaurant. It’s less useful as a replacement for the scale on foods you cooked yourself and could have weighed dry, because a scale beats an estimate every time. Use the camera for the meals where a scale was never an option, and the scale for the ones where it was.

The fix, in one sitting

Open your saved foods and check the five items you log most often. For each, decide whether the entry is raw or cooked, then rename it so future you can’t get it wrong: “Chicken breast RAW”, “Basmati rice DRY”. Two minutes of admin.

If your protein always looks lower than you’d expect for what you eat, you’re probably weighing cooked meat against raw entries. If your carbs look absurd for a modest plate of pasta, you’re weighing cooked against dry. Both are fixable today, and neither requires eating differently. It just requires your log to describe the same food your body received.