Add up the calories of everything that goes in the pot. Weigh the finished dish. Divide total calories by total grams — that is the calories per gram for the pot. Stir before serving, weigh your own bowl, multiply.
Boiling changes the weight, not the calories: water leaves, energy stays. That is precisely why the per-gram figure has to be taken after cooking, not from the recipe.
Why a bowl of stew has no serving size
Packaged food comes with the problem already solved. The label states a serving, the app knows the serving, and the only thing you supply is how many you had.
A pot of anything is different in a way that breaks that whole approach. There is no serving, because the dish was never divided until you divided it. The recipe might say "serves four", but that is the author guessing at appetites, not a measurement of your bowl. Two people eating from the same pot can take portions that differ by three hundred calories and both believe they ate a quarter of it.
It gets worse with a ladle. The solids sink, the liquid does not, and the first bowl out of a curry is not the same food as the last one. Stirring helps. It does not make the ladle a unit.
The fix is not to guess better. It is to stop dividing by servings and start dividing by grams.
The finished-weight method, in four steps
The whole method is one idea: the pot has a fixed number of calories and a measurable weight, so it has a calories-per-gram figure — and that figure is the honest average for the whole pot.
Average is the right word, and it is why the ladle paragraph above matters. Weighing your bowl fixes how much you took. Stirring the pot before you serve is what stops it also being the wrong half. Do both and the number holds; do neither and you have swapped one guess for another.
None of this is our invention. It is how people who track seriously have always handled shared dishes, and every serious recipe editor lets you enter the finished cooked weight for exactly this reason. It is simply under-explained: the pages that come up for this question mostly tell you to divide by the number of servings, which is the one step that does not work for a pot.
- Log every ingredient as it goes in. Weigh it, or use the packet weight. Everything counts — the oil in the pan, the stock cube, the splash of cream at the end. Only water and plain spices are free.
- Add up the calories. One number for the whole pot. Your app will do this if you build it as a recipe; a scrap of paper works equally well.
- Weigh the finished dish. Put the pot on the scale when it is done, then subtract the empty pot. Or ladle it into a container you have already weighed. This number is not the same as the sum of the raw ingredients, and that is the point.
- Divide, then weigh your bowl. Total calories ÷ total grams = calories per gram. Multiply by whatever you actually served yourself.
Do it once, not every time. Write the calories per 100 g on the lid with a marker. Every portion after that — tonight, tomorrow, the one you take to work — is a weight times a number you already have. This is why the method survives contact with a normal week and portion-guessing does not.
A worked example: chicken curry for the week
Real numbers, so you can see the shape of it. Generic ingredients are rounded from USDA FoodData Central entries. Two of them depend on the jar rather than on any database: curry paste runs from roughly 125 to 300 kcal per 100 g depending on how much oil is in it, and coconut milk from about 150 to 230. Use your own labels for those two — that is the whole point of the method.
Everything that went in the pot
Raw ingredients weighed 1,990 g going in — a 400 ml tin of coconut milk weighs almost exactly 400 g, so it can be added to a gram total without cheating. The finished curry weighed 1,750 g — 240 g lighter, because that much water left as steam. None of the calories left with it.
So the arithmetic is:
That last line is the whole argument. Both of those are "a bowl of curry". Guessing which one you had is a 207-calorie coin flip, every time you eat it.
Work out your own pot
Three numbers. Nothing is stored, and no account is needed.
Boiling it down does not remove calories
This is the part almost nobody says out loud, and it is the reason the per-gram figure has to be taken after cooking.
Water carries no energy. When a stew simmers for an hour and reduces, what leaves the pot is steam — and steam has no calories in it. The pot ends up lighter, with exactly the same number of calories in it as when you put the lid on. Every remaining gram is therefore denser.
The same pot, before and after an hour on the hob
Nothing was added. The soup did not become more fattening. You are simply eating a more concentrated 300 g than you would have been an hour earlier — which is why a figure worked out from the raw recipe is wrong for the dish that actually reaches the table.
The same logic runs the other way. Add a mug of water when reheating and the pot has the same calories spread over more grams, so per-gram falls. Rice, pasta and dried pulses do this too as they absorb liquid: the starch that went in dry is the same starch, now diluted across two or three times as many grams.
What changes the calories and what only changes the weight
This is where most of the confusion lives, and it resolves into a single question: did anything with energy in it enter or leave the pot?
| What happens | Calories in the pot | Weight of the pot | Effect on per-gram |
|---|---|---|---|
| Water boils off as steam | unchanged | down | rises |
| You add plain water at the end | unchanged | up | falls |
| You add stock at the end | up — a cube or a carton is not water | up | roughly unchanged |
| Rice, pasta or pulses absorb water | unchanged | up | falls |
| You fry in oil first | up | up | rises |
| Fat renders out of the meat | unchanged — it is still in the pot | unchanged | unchanged |
| You skim the fat off and throw it away | down | down | falls |
| It sits overnight in a covered container and thickens | unchanged | unchanged | unchanged — thickening is texture, not evaporation |
| It sits uncovered and dries out | unchanged | down | rises |
Two rows are worth reading twice. Rendered fat does not disappear — it pools on the surface and gets served with the rest, so it stays in the count. And skimming it off is one of the only ways to genuinely remove calories from a finished dish: if you lift 30 g of fat off the top of a stew and bin it, you have removed about 270 calories from the pot. Re-weigh afterwards if you do.
The raw-versus-cooked trap
The single most common way to be badly wrong, and it has nothing to do with the pot.
Dry rice roughly triples in weight when cooked. Dry pasta roughly doubles. Dried beans and lentils roughly double or triple depending on how long they soak. The calories do not change at all — they simply arrive attached to two or three times as many grams.
So a database entry for "rice, white, raw" at around 360 kcal per 100 g and one for "rice, white, cooked" at around 130 are describing the same rice. Log 300 g of cooked rice against the raw entry and you book about 1,080 calories where the true figure is about 390 — an error of roughly 700, from one line in a diary.
While we are here: cooling rice does not halve its calories
You will have seen this one. Cook rice, chill it overnight, reheat it, and supposedly you absorb far fewer calories because some starch turns resistant.
The first half is true and the second is not. Chilling cooked rice does raise its resistant-starch content: in the trial this claim is usually traced back to, it went from about 0.6 g per 100 g freshly cooked to about 1.6 g after a day at fridge temperature and reheating — one small study, one variety of white rice, so treat the exact figures as an order of magnitude rather than a constant.
Now price it. Resistant starch is fermented in the large intestine rather than digested, so it yields roughly two calories a gram instead of four. One extra gram per 100 g is therefore worth about two calories per 100 g. On a 200 g serving of cooked rice — around 260 calories — that is a saving of about four. Anyone reporting a 50% cut is describing something other than a plate of rice.
It is worth knowing because it is the kind of belief that quietly corrupts a whole week of logging: not one wrong entry, but a standing discount applied to every meal with rice in it.
The five ingredients that cause most of the error
Get these right and the rest barely matters. Miss one and nothing else you weigh will rescue the total.
| Ingredient | Per 100 g | Per tablespoon | Why it gets missed |
|---|---|---|---|
| Cooking oil | ~884 kcal | ~120 kcal | Poured, never weighed. Two glugs is 240 calories |
| Butter | ~717 kcal | ~100 kcal | Melts into the dish and disappears from memory |
| Coconut milk (canned, full fat) | ~195 kcal | — | A 400 ml tin is most of a curry's calories |
| Heavy cream (heavy whipping, ~36% fat) | ~340 kcal | ~50 kcal | Added at the end, after the logging is done |
| Hard cheese | ~400 kcal | — | Grated on top, counted as a garnish |
The pattern is that all five are fats, and fat carries nine calories per gram against four for protein and carbohydrate. A dish can be entirely reasonable until somebody finishes it with cream.
When you did not cook it
Someone else's stew cannot be weighed, and pretending otherwise produces a precise-looking number built on nothing. What you can do instead:
- Estimate the fat, not the dish. Whatever else is in there, the calories live in the oil, cream or coconut milk. Assume more than looks reasonable — home cooks and restaurants both use more than you would.
- Anchor to something similar you have measured. If you have worked out one curry, you have a per-100 g figure for a whole category of dish. Use it and adjust.
- Weigh what you can. Even at someone else's table you usually know roughly what went on the plate. A 400 g bowl estimated at a plausible density beats a shrug.
- If a chain cooked it, the number may already be printed. US chains from twenty locations up have to post it — counting calories eating out covers what that number does and does not tell you.
- Be consistent rather than exact. A log that reads 10% high every day still shows the right trend, and the trend is the thing you act on.
If you have no scale at all, that is a different problem with a different answer — where estimating by eye costs you most is worked out separately.
A photograph helps here more than it helps with packaged food, because the alternative is nothing at all — but it is also the hardest case for any estimate, mixed dishes being exactly where a camera cannot see the oil. We wrote up how far a photo estimate can be trusted, including which dishes it gets worst.
The short version
- Count the calories going in, weigh the dish coming out, divide.
- Evaporation removes water, never calories.
- Write calories per 100 g on the container and the work is done for the week.
- Weigh raw and log raw, or weigh cooked and log cooked — never mix them.
- If you only get one thing right, get the fat right.
Please read. This is a method for counting, not dietary advice. If you are managing a medical condition, pregnant or breastfeeding, or working to a therapeutic diet, use the figures your clinician gives you. If counting is making your relationship with food worse rather than better, stopping is a reasonable answer — in the US the National Alliance for Eating Disorders helpline is 1-866-662-1235. See our health disclaimer.
You can earn the full version instead of paying for it
If you would rather not do the arithmetic by hand, a photo of the bowl gets you an estimate in seconds, and photo logging is on the free plan. Describing a dish in words instead — "chicken curry with coconut milk, about 400 grams" — suits a stew better than a photograph does, and that one is a Full-plan feature.
- Create an account and log a few meals. New accounts open with a two-day Plus trial — three days if somebody invited you — and then settle onto the free plan: no card, no end date, one meal scan a day.
- Publish an honest post, story, write-up or short video about what you found — or invite three people who go on to actually use it.
- Send us the link. A published piece unlocks the higher tier; three qualified invites unlock 90 days of the full version.
Your post has to say you got free access in exchange — that is an FTC requirement in the United States. Access is granted whether what you write is positive or critical. Full terms: promotional access.