Can you count calories without weighing food?

Yes, you can count calories without weighing food, and a scale was never the only error. A 1992 study of 10 people who reported eating little and not losing found portion estimates accurate but reported intake about half of actual. A 2016 study of 67 adults found hands and cups miss shapeless foods. You can check your own log.

Where the error sat in a 1992 study

In a 1992 study in the New England Journal of Medicine, researchers screened 224 adults at a hospital weight-control clinic. Ten of them, nine women and one man, said they ate little and still did not lose weight. The team measured their energy burn with a lab method and tracked body change, and worked out what they had really eaten.

Their reported intake was about 47 percent below actual intake on average. Their portion-size estimates, tested separately, were accurate and similar to a comparison group's. A day after a test meal they recalled eating about 20 percent less than they ate.

So in these 10 people the gap was not mainly about misjudging a portion. It sat in what was recorded and recalled. A scale cannot fix a snack that never reaches the log.

The limits matter. Ten self-selected people with a history of not losing on a low intake is a small, specific group, and the paper does not describe it as a finding about dieters in general. The authors say the mechanism is not well understood, and the gap did not look like deliberate deception. The study also does not test whether a steady error is safe to live with, so we do not claim that.

Where the error sat in a 2016 hand-ruler study

A 2016 study in the Journal of Nutritional Science tested the other half of the question: food shape. It gave 67 adults at one university a single 1.5-hour session. Seventy percent were women, the average age was about 33, and the average BMI was 23. Each person estimated pre-weighed foods placed in plain view. There were 21 foods in two sizes each, so 42 portions.

After brief training, they used a finger width as a ruler. For 24 block, wedge and cylinder items, 19 (80 percent) landed within 25 percent of true weight. That count comes after the paper set aside fish fillets, chicken breasts and beef steak, which were not estimated well.

Shapeless foods went worse. By the fist method, no shapeless food or muffin landed within 10 percent of its true weight. Household measures did poorly on them too, with exceptions: cups put both cereal-flake portions within 10 percent, and spoons did so for the larger portion of the oil-based spread. The authors say hand or household estimates of shapeless foods should be read with caution.

This is the case against the argument that you either weigh or give up. Our view is that the shape of the food decides how much a scale helps. A scale matters most for the foods the eye misses: rice, mashed potato, spreads.

Read the study's limits before you copy it. The people were young, mostly female and not overweight, and the training was brief. It was a controlled test with food in front of them, not real logging over days, and no app was tested. We are not giving you a palm, fist or thumb routine, because no paper we cite tested a palm, and the fist method did poorly in the hand-ruler study.

What the simplified-tracking pilot did not test

If you quit weighing, you may have met the idea of "simplified" tracking. A 2022 pilot trial randomized 72 adults, all with a child at home, to track calories in the standard way or to track only high-calorie "red" foods against a daily limit. After six months, weight loss was not significantly different between the two arms, and tracking days were similar.

That is not a trial of rough calorie estimates. The simplified arm counted red foods, not guessed portions. There was no group that tracked nothing, so it cannot show that either style caused weight loss. It was a pilot, and the group was mostly well-educated women who were parents.

We cite it for one reason: it shows researchers testing a lighter way to log. It does not show that rough counting costs less than quitting, and it does not show that a consistent rough log gives a usable trend. No paper here tests either claim.

A one-week check of your own log

Here is our own suggestion, not a finding from any study. For one week, log everything you eat, even roughly. Beside each entry, note anything that was hard to record: a snack standing at the counter, a meal someone else cooked, a handful from a bag, a drink.

At the end of the week, read the notes. The foods and moments you flagged are where your own log is most likely to miss. They show where you might spend effort, whether that means writing something down sooner or weighing one food that is hard to judge by eye.

Say a week's notes show that dinner is easy to log and the afternoon is not. The 1992 study suggests that what goes unrecorded is where the gap sits. One honest rough entry for it beats a precise dinner and a blank afternoon. That example is illustrative, not measured.

Missed days are fine. A gap does not spoil the check; log the days you can. It shows where your log is likely to miss, not why your weight did what it did. Reading the weight trend is a separate job.

When a scale is worth it, and when to stop

A scale is worth it for the foods the hand-ruler study handled worst: shapeless foods, spreads, fish fillets, chicken breasts and steak you eat often. Weigh those, estimate the rest, and keep the habit you can sustain. We would rather you log roughly for a month than weigh perfectly for a week.

Stop if logging stops being useful. If tracking starts to feel compulsive, or you feel anxious when you cannot do it, it is fine to stop and talk to someone: a doctor or registered dietitian is the right call, and especially so if you have a history of disordered eating, are pregnant or breastfeeding, or manage a medical condition with your diet.

Questions people ask

A calorie range on each portion

We make BurnCurve for people who want to log roughly. Each portion carries a calorie range, and your burn learns from that range. No study here tested BurnCurve.