How to count calories without going crazy: complete 2026 guide

Counting calories works, but it does not have to be perfect

Calorie counting has a reputation for being tedious and obsessive, and it earned part of it: open an app, search each food among thousands of results, work out the portion and repeat it 3-5 times a day. To get a feel for what a typical dish holds, check the dish calories: paella, döner kebab, schnitzel or lentil stew, with the recipe gram by gram. What almost nobody tells you is that you do not have to do it perfectly. Xu and colleagues measured it in a real digital programme (WeightWatchers, 153 participants; Obesity, 2023): logging food on 28.5% of days already predicted losing 3% or more of body weight at 6 months; 39.4% of days predicted 5%; 67.1% predicted 10%. Their conclusion is literal: perfect tracking adherence is not necessary.

And logging helps. In the Burke systematic review (2011), studies consistently found that self-monitoring is associated with greater weight loss, although the authors themselves rate the evidence as weak because of methodological limitations. In the Weight Loss Maintenance trial (Hollis et al., 2008, 1,685 participants), the behaviours measured — with food records at the top — explained most of the gap between people who lost a lot and people who lost little. What makes people quit is the effort: in the Harvey study (Obesity, 2019), logging took 23 minutes a day in month one and 15 in month six, and those who lost 10% or more did not spend more minutes but opened their diary more times a day (2.7 vs 1.7). They logged in the moment instead of rebuilding the day at night.

How wrong we are when we eyeball it

Very, and almost always on the low side. The most cited case is Lichtman (NEJM, 1992): ten people with obesity who said they could not lose weight despite eating under 1,200 kcal a day. Their energy expenditure was measured and it was within the expected range; what failed was the record: on average they reported 47% less than they ate and overstated their physical activity by 51%. They were not lying: what you do not see, you do not write down. And it is not only memory: in restaurants, 19% of 269 dishes analysed in a lab contained at least 100 kcal more than the restaurant stated (Urban, JAMA 2011). So rather than chasing absolute precision, it pays to know where the error hides and which method reduces it.

What is a calorie and why does it matter

A calorie is a unit of energy: technically, the energy needed to raise the temperature of 1 gram of water by 1 degree Celsius. When we talk about calories in nutrition, we actually mean kilocalories (kcal), although by habit we just call them "calories". Your body needs energy for everything: breathing, thinking, digesting, walking, sleeping. Even at complete rest, an adult burns somewhere around 1,200 to 2,000 kcal a day just to stay alive. This is called the basal metabolic rate (BMR).

Energy balance explained

If you eat more calories than you burn, the excess is stored, mostly as fat. If you eat less, your body draws on its reserves. This is energy balance, the first law of thermodynamics applied to nutrition. For weight it does not matter whether you eat "clean" or "dirty": 3,000 kcal of salad (almost impossible, but it makes the point) would make you gain weight just like 3,000 kcal of pizza. Quality matters for health and hunger; quantity matters for weight.

How to calculate your daily calories step by step

Your total daily energy expenditure (TDEE) has three parts: basal metabolic rate (60-70%), the thermic effect of food (around 10%) and physical activity (20-30%). The equation most used in clinical research is Mifflin-St Jeor, published in 1990. If you would rather not do the maths by hand, the calorie calculator applies the same equation to your data.

• Step 1 — Calculate your BMR: men = (10 × weight in kg) + (6.25 × height in cm) − (5 × age in years) + 5. If you weigh it cooked, the cooked to raw weight calculator tells you what it weighed raw. Women = (10 × weight in kg) + (6.25 × height in cm) − (5 × age in years) − 161
• Step 2 — Multiply by your activity factor: sedentary (office, no exercise) = BMR × 1.2. Lightly active (exercise 1-3 days) = BMR × 1.375. Moderately active (3-5 days) = BMR × 1.55. Very active (6-7 days) = BMR × 1.725
• Step 3 — Adjust for your goal: to lose weight, subtract 300-500 kcal. To gain weight, add 300-500 kcal. To maintain, eat your TDEE

A practical example

Sarah is 30 years old, weighs 65 kg, is 165 cm tall and exercises 3 times a week. Her BMR = (10 × 65) + (6.25 × 165) − (5 × 30) − 161 = 650 + 1,031 − 150 − 161 = 1,370 kcal. Her TDEE = 1,370 × 1.55 = 2,124 kcal. To lose weight, she should eat between 1,624 and 1,824 kcal a day. At first she would lose about 0.3-0.5 kg a week; over the months somewhat less, because a lighter body burns less (Hall, Lancet 2011): that is why it is worth recalculating every few kilos.

The photo method: fast, but not exact

Today's vision models are good at recognising what is on the plate; the hard part is how much. Studies put numbers on it, and not all of them are flattering. In Sweden, ChatGPT-4o and Claude 3.5 Sonnet estimated the calories in 52 standardised photos with a mean error of 36%, and all three models tested fell further short as the portion grew (Fridolfsson, Curr Dev Nutr 2025). In Ireland, ChatGPT-4 correctly identified 93% of foods in 114 meal photos from a national dietary survey, but performed poorly at estimating the weight of medium and large portions (O’Hara, Nutrients 2025). In other words: a photo gives you a fast estimate worth checking, not the accuracy of a scale.

How to use it well in practice

Take the photo before you start eating, from above and with the whole plate in view; if something is hidden (the oil, the sauce under the rice), add it by hand. Then look at the portion it suggests and correct it if it does not match what is in front of you: that is exactly the weak spot studies measure. Pay attention to how the food was cooked, because it changes the number: a large hard-boiled egg is 78 kcal and a fried one 90; 100 g of cooked white rice is 130 kcal and brown rice 123 (USDA FoodData Central). In Renzy, for example, the photo gives an estimate of calories and macros that you can correct, and for packaged foods there is the barcode scanner. If you want to see in more depth how accurate photos are, we cover it in how accurate counting calories with a photo is.

The 7 mistakes that hide the most calories

These are the mistakes that most often keep people from losing weight despite "counting calories". Figures come from USDA FoodData Central (SR Legacy), with each food's identifier; for anything else there is a page per food in the food calorie lookup:

• Not counting oil: one tablespoon of olive oil (13.5 g) is 119 kcal (FDC 171413). If you cook with 2-3 tablespoons a day, that is about 240-360 kcal you cannot see on the plate. Measuring it with the spoon instead of pouring it straight from the bottle already changes the count a lot
• Forgetting drinks: a coffee with 150 ml of whole milk and two teaspoons of sugar comes to about 130 kcal (FDC 171265 and 169655). A cup (248 g) of fresh orange juice is 112 kcal (FDC 169098), almost the same as that amount of regular cola, 104 (FDC 174852). And alcohol adds up fast: a 355 ml can of beer, 153 kcal (FDC 168746); a 150 ml glass of red wine, about 125 (FDC 173190); a gin and tonic with 50 ml of gin and 200 ml of tonic, about 190 (FDC 171920 and 171869)
• Eyeballing portions, especially big ones: in the Fridolfsson study the AI fell further short as the portion grew, and if you estimate yourself, be wary of your big plates too. Also watch out for mixing raw and cooked: 200 g of cooked pasta is 316 kcal (FDC 169737), but logged with the dry pasta entry (371 kcal/100 g, FDC 169736) it comes out at 742
• Forgetting nibbles: tasting while you cook, grabbing a few nuts in passing, finishing the kids' leftovers. A handful of 23 almonds (28 g) is 164 kcal (FDC 170567) and a tablespoon of mayonnaise 94 (FDC 171009): three or four of those a day already add up to a whole snack
• Not counting weekends: if from Monday to Friday you eat 500 kcal below maintenance (2,500 kcal in total), a Saturday 1,000 kcal above it and a Sunday 1,500 above it erase the deficit for the whole week
• Trusting labels and menus completely: the FDA treats a label as compliant as long as actual calories are no more than 20% above the declared value (21 CFR 101.9). In restaurants, 19% of 269 dishes analysed had at least 100 kcal more than stated (Urban, JAMA 2011)
• Obsessing over precision: chasing the exact number is tiring and leads people to quit, and for some it worsens their relationship with food (more on that below). A 100 kcal error on one day goes unnoticed; what moves the scale is the trend over weeks

When counting calories is NOT a good idea

Counting calories is a tool, not an obligation, and it does not suit everyone. Among 493 college students, those who used calorie apps showed more eating concern and dietary restraint, even after accounting for BMI (Simpson and Mazzeo, Eating Behaviors 2017); it is a cross-sectional study, so it does not show the app causes the problem, but the warning is there. Among 105 people diagnosed with an eating disorder, about three in four used MyFitnessPal, and 73% of those users felt the app had contributed to their disorder (Levinson, 2017). This is general information, not a diagnosis: if you have or have had an eating disorder, do not start counting without talking to your doctor or healthcare team. And stop and see a professional if you recognise yourself in any of these signs:

• You skip meals or compensate with exercise so the numbers "add up"
• Not being able to log a day makes you anxious or guilty
• You spend much of the day thinking about the numbers of what you eat or are about to eat
• You are pregnant, still growing or have a condition that shapes your diet: your target should come from your healthcare professional, not an equation

You do not need to log everything: the 80/20 rule

You do not need to record every gram you eat. The 80% does not come from any study: it is a way of saying "most days, most meals". What has been measured is that consistency matters more than perfection: in the Xu study, logging on 28.5% of days — about two in seven — already predicted losing 3% of body weight. The days you do not log — a dinner out, a trip, a day you simply do not feel like it — nothing happens. What matters is the pattern over weeks, not what you ate on a Tuesday at 3pm.

How long to count, and alternatives to doing it every day

Nobody has measured how many weeks it takes to learn to estimate portions by eye, so be wary of anyone who gives you an exact figure. What we do know is that the effort drops with practice (from 23 to 15 minutes a day in the Harvey study) and that logging on only part of the days is already associated with weight loss. With that, there are several ways to get most of the benefit for less work: log two or three days a week, one of them at the weekend; log thoroughly for two or three weeks to calibrate your eye, then come back for a full week when your average weight stalls for 3-4 weeks; or log only what slips past you most (oil, drinks, nibbles) and eyeball the rest. The best method is the one you are still using a month from now.