I eat twice a day and I’m done by 2pm. Before you copy that, here’s what the trials say about meal count, in the reader’s units: about 1.5 to 2 kg over three months, from evidence the authors themselves call messy.
Fewer meals moved the scale, a little, across 29 trials
The biggest recent review on when and how often people eat pooled 29 randomised trials, 2,485 adults, all at least 12 weeks long, and split the interventions three ways. Time-restricted eating. Fewer meals per day. And shifting calories earlier in the day. Each was compared against people eating normally.
Liu HY, Eso AA, Cook N, O’Neill HM, Albarqouni L. Meal Timing and Anthropometric and Metabolic Outcomes: A Systematic Review and Meta-Analysis. JAMA Netw Open 2024;7(11):e2442163. doi:10.1001/jamanetworkopen.2024.42163
Read the authors’ own summary before you read anyone’s landing page: the effects were “small and of uncertain clinical importance,” with high heterogeneity and a high risk of bias across the trials. Two kilos over a season, from studies that disagree with each other. That’s the real size of it.
The other review says there’s no difference at all
A year earlier, a different team asked a narrower question: three or fewer meals a day versus four or more, in 16 randomised trials. Their answer was a mean difference of −0.62 kg with a confidence interval from −2.76 to +1.52. That interval crosses zero, so the honest reading is “no detectable difference,” and they graded it very low certainty because 15 of the 16 trials were at high risk of bias.
Blazey P, Habibi A, Hassen N, Friedman D, Khan KM, Ardern CL. The effects of eating frequency on changes in body composition and cardiometabolic health in adults: a systematic review with meta-analysis of randomized trials. Int J Behav Nutr Phys Act 2023;20(1):133. doi:10.1186/s12966-023-01532-z
So you have one review that finds a small effect for fewer meals and one that finds none, and both say the underlying trials are shaky. Anyone telling you two meals a day is “proven” is reading one paper. Anyone telling you it’s “debunked” is reading the other. The truthful sentence is that meal count, on its own, does a little or nothing, and the trials aren’t good enough to say which.
Where the effect actually comes from
If meal count is a weak lever, why does eating twice a day work for so many people, me included? Two reasons, and neither is metabolic magic.
You eat less without deciding to. Fewer eating occasions means fewer chances to eat, and the trials that let people eat freely are the ones that show a benefit. When researchers hold calories equal across arms, the advantage of eating earlier mostly disappears. The window does the portion control for you. That’s the whole point of not counting: you can get the intake reduction without doing arithmetic on a sandwich.
When you eat matters more than how many times. Look at the figure again. Earlier calories had the tightest confidence interval of the three. In a 14-week trial, 90 adults with obesity who ate between 7am and 3pm lost 2.3 kg more than a group with the same calorie cut and a 12-hour window (Jamshed et al., JAMA Intern Med 2022). Meanwhile 116 adults eating noon to 8pm for 12 weeks lost 0.26 kg more than control, which is nothing (Lowe et al., TREAT, JAMA Intern Med 2020). Same eight hours. Different clock. Different result. I went through the timing evidence in detail here.
Satchin Panda, whose lab at the Salk Institute did the animal work that started the time-restricted eating field, is careful about exactly this, including who it isn’t for. Worth an hour if you’re going to build a habit on it.
The two-meal study you’ll get sent, and why it doesn’t apply
Someone will link you a 2021 paper where eleven lean people ate one meal a day, in the evening, and it did worse than three meals. It’s a real paper (Meessen et al., Front Physiol 2021). It’s also one late meal in eleven lean volunteers, which is close to the opposite of two early meals in someone with weight to lose. A late single meal is a different, worse-timed intervention, and I don’t cite it for my case in either direction.
What two meals by 2pm looks like in practice
My version: a real breakfast, a real lunch, last bite done by 2pm most days. Two things about the “most days.”
- Dinner is meal two. If there’s a birthday dinner or you’re travelling, eat breakfast, skip lunch, and dinner is your second meal. You’ve kept the count at two. The cutoff moved, and that’s fine. The point is awareness, not a rule you fail.
- Weigh after the second meal, whenever it is. That number is your ceiling for the day, and it’s the only number the method asks you to manage. A late dinner shows up as a higher ceiling and a bigger overnight drop, and you just write both down.
Forty pounds in a year, at 0.77 lb a week, is what it did for me. That’s one person and one year. Your body is not my body, and the two reviews above are the reason I say that instead of promising you a number.
Who should not try this
Fewer, earlier meals are a mild intervention in the trials, with transient hunger and headaches in the first days and no serious adverse events reported. The long-term unknowns the reviewers name are lean-mass loss and adherence under social pressure. What the trials don’t cover: anyone under 18, anyone pregnant, and anyone with a history of disordered eating, for whom a cutoff is not data, it’s a trigger. The screening questions are here, and the app asks them before your first entry.
Every paper in this post is on the science page with its DOI, including the one that finds no effect. Want to see what two meals does to your own overnight number? The calculator takes ten seconds.