Do you weigh food cooked or raw for calories?

Here's the rule, and where it actually trips people up.

Why the same chicken breast gives two different numbers

Picture a 200g raw chicken breast. You cook it, and it comes out of the pan at roughly 150g. Nothing was added. So where did the other 50g go? Water: cooking drives moisture out through evaporation, and lean meat loses a meaningful chunk of its raw weight this way. The chicken breast you're about to eat has the same total calories it had raw (cooking doesn't add or remove energy, it just changes water content), but that energy is now packed into less weight. Per gram, cooked chicken reads more calorie-dense than raw chicken. Log 150g of cooked chicken against a raw chicken entry, or 200g of raw chicken against a cooked entry, and you've quietly introduced an error, in opposite directions, but both real.

Rice, pasta, oats, and dried legumes do the reverse. Dry rice absorbs water as it cooks and roughly triples in weight, which means the calories per gram go down, not up. A 50g serving of dry rice and a 50g serving of cooked rice are wildly different amounts of food.

The rule: match your weight to your database entry

This is really the whole answer. Every nutrition database, USDA's, and the ones apps like Bocco are built on, lists foods by their as-measured state: "chicken breast, raw" is one entry, "chicken breast, roasted" is another. The rule is simple: weigh the food in the same state the entry describes, and don't swap mid-log. Weighing isn't the hard part. Matching the number to the right entry is.

The USDA's own documentation on building its cooked-food values, the Table of Nutrient Retention Factors, exists specifically to account for what cooking does to weight and nutrient concentration. If the federal government needs a dedicated methodology for this, it's a safe bet it's not something you'll get right by instinct alone.

Meat, poultry, and fish: weigh raw when you can

For anything you're cooking yourself, weighing raw, before it hits the pan, grill, or oven, is usually the more reliable habit. One weigh-in, no guessing how much water cooked off, and most recipes and database entries for proteins default to a raw baseline anyway. If you're logging something you didn't cook (restaurant chicken, a rotisserie breast from the deli), you're stuck weighing it cooked, just make sure whatever you're logging against is also a cooked entry, not a raw one.

Rice, pasta, oats, and legumes: the opposite problem

Here the water goes in, not out, and the swing is bigger: a cup of dry rice turns into roughly three cups cooked. Weighing dry is usually easier (less sticky, less mess, and you often measure it dry anyway before it goes in the pot), but it only works if your cooking ratio is roughly consistent. If you're logging a cooked portion from a container of rice you made earlier in the week, weigh it cooked and log it against a cooked entry, don't try to reverse-engineer the dry weight.

Vegetables: usually simplest raw

Most vegetables lose a modest amount of water when cooked, nowhere near rice's swing, but enough to matter if you're chasing precision. The easiest fix: weigh vegetables raw, before they go in anything. It's one less step while you're actually cooking, and raw is how most vegetable entries in nutrition databases are logged anyway.

The #1 mistake: mixing a raw weight with a cooked entry

This is the actual failure mode, more than "raw vs. cooked" in the abstract. Someone weighs a cooked chicken breast at 150g, searches "chicken breast" in their tracker, and picks whichever entry comes up first, sometimes that's the raw one. Now they've logged 150g of raw-chicken-breast calories for a piece of chicken that's actually denser than that, because it's cooked. The gap isn't trivial: it's the entire reason raw and cooked get separate entries in the first place.

The fix isn't memorizing conversion ratios. It's a two-second check: does the food in front of me match the state described in what I just logged? If you weighed it sizzling out of the pan, make sure the entry says "cooked," "roasted," or "grilled," not "raw."

It matters less than you think for your daily total

Worth saying plainly: this is a real source of error, but it's a smaller one than most people fear, and it's not a reason to abandon tracking over. If you're consistently off by 10-15% on a chicken breast because of a raw/cooked mismatch, that's one line item in a day with a dozen others, some of which overestimate, some of which underestimate. The goal is a trustworthy weekly picture, not a lab-grade single meal. Get the big habit right (pick a state, weigh that state, match the entry) and the smaller imprecisions wash out over a week.

That said, if you cook the same few meals on repeat, and most people do, it's worth getting those specific entries right once. A correctly matched chicken breast you log fifty times is worth more than a perfect one-off.

Skip the debate entirely: log what's on your plate

This whole raw-vs-cooked question only exists because you're weighing food before you eat it. A photo sidesteps it completely. Bocco reads the meal as it actually sits on your plate, already cooked, already dressed, already portioned, and estimates the calories and macros from that. No raw weight to remember, no "was this entry cooked or raw" guessing. I built it this way on purpose, because the raw/cooked mismatch is one of the quieter ways a well-meaning tracker goes wrong, and the easiest fix is removing the decision entirely.

If a photo estimate looks off, bigger portion, extra oil, a different cut, you correct it in a sentence and the numbers recalculate properly, instead of just getting relabeled.