How to count calories in soups, stews and curries

A packet tells you what a serving is. A pot of curry does not, and the ladle is not a unit of measurement. This is the method that works for anything cooked together and shared out — and the arithmetic is smaller than it looks.

Updated August 18, 2026Free to use · no sign-up for the calculators

Short answer

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.

  1. 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.
  2. 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.
  3. 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.
  4. 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

Chicken thigh, boneless and skinless, raw — 700 g1,050 kcal
Onion — 200 g80 kcal
Coconut milk, canned, full fat — 400 g780 kcal
Chopped tomatoes, canned — 400 g80 kcal
Vegetable oil — 30 g265 kcal
Curry paste — 60 g120 kcal
Spinach — 200 g46 kcal
Total in the pot2,421 kcal

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:

2,421 kcal ÷ 1,750 g1.38 kcal per gram
Per 100 g — the number to write on the lid138 kcal
A 400 g bowl553 kcal
A 250 g bowl346 kcal
Difference between those two bowls207 kcal

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

Calories in the pot — unchanged throughout1,200 kcal
Weight when you put the lid on2,400 g → 0.50 kcal/g
Weight when it has reduced1,800 g → 0.67 kcal/g
A 300 g bowl therefore goes from150 → 200 kcal

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 happensCalories in the potWeight of the potEffect on per-gram
Water boils off as steamunchangeddownrises
You add plain water at the endunchangedupfalls
You add stock at the endup — a cube or a carton is not wateruproughly unchanged
Rice, pasta or pulses absorb waterunchangedupfalls
You fry in oil firstupuprises
Fat renders out of the meatunchanged — it is still in the potunchangedunchanged
You skim the fat off and throw it awaydowndownfalls
It sits overnight in a covered container and thickensunchangedunchangedunchanged — thickening is texture, not evaporation
It sits uncovered and dries outunchangeddownrises

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.

The rule. Weigh raw, log raw. Weigh cooked, log cooked. Never weigh one and log the other. With the finished-weight method this mostly takes care of itself, because the pot gets weighed after cooking while the ingredients were counted before it — which is exactly why both numbers have to be recorded at the moment they are true.

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.

Rounded from USDA FoodData Central entries. Per tablespoon figures assume a level spoon.
IngredientPer 100 gPer tablespoonWhy it gets missed
Cooking oil~884 kcal~120 kcalPoured, never weighed. Two glugs is 240 calories
Butter~717 kcal~100 kcalMelts into the dish and disappears from memory
Coconut milk (canned, full fat)~195 kcalA 400 ml tin is most of a curry's calories
Heavy cream (heavy whipping, ~36% fat)~340 kcal~50 kcalAdded at the end, after the logging is done
Hard cheese~400 kcalGrated 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.

  1. 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.
  2. 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.
  3. Send us the link. A published piece unlocks the higher tier; three qualified invites unlock 90 days of the full version.
Create a free account →

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.

Frequently asked questions

Add up the calories of every ingredient that went into the pot. Weigh the finished dish, pot excluded. Divide the total calories by the total grams to get calories per gram, then weigh your own bowl and multiply. Evaporation removes water, not calories, so the total for the pot is the same before and after cooking — only the weight changed.
No. Boiling drives off water, which carries no calories. The pot gets lighter and every remaining gram gets denser, which is exactly why the per-gram figure has to be worked out after cooking rather than before.
Either, as long as the database entry matches. Raw rice roughly triples in weight when cooked and dried pasta roughly doubles, so 100 g of dry rice logged as 100 g of cooked rice understates the entry by about two thirds. The mistake is not the choice — it is mixing the two.
Then you are estimating, and the honest move is to estimate deliberately: identify the two or three ingredients that dominate the calories — oil, cream, coconut milk, cheese, nuts — and be generous with them. Restaurant and home cooks both use more fat than you expect.
The calories do not, and in a covered container neither does the weight — overnight thickening is starch and fat setting, not evaporation. Left uncovered it can dry out, and then the per-gram figure rises: re-weigh if that happens. Add water when reheating and it falls.
Once per recipe, not once per bowl. You do the arithmetic when you cook, write the calories-per-100 g on the container, and every serving after that is a number times a weight. For anything you cook regularly it is a few minutes that pays out for months.