Every method here — dieting styles, fasting, meal timing, training — is backed by the randomised trial or meta-analysis that tested it, with the sample size and study design shown alongside the result. No single "best" diet; the evidence points somewhere more useful than that.
This isn't a quiz with a score — it's context that shapes which parts of the evidence below are actually relevant to you. Answer as much or as little as you like.
Losing fat is the visible outcome. The lever behind it — the one that also affects your energy, your disease risk, and how hard the weight is to keep off — is insulin sensitivity. This is the mechanism worth understanding, not just the diet name.
Insulin's job is to move glucose out of your blood and into your cells. When cells stop responding well to insulin — usually driven by carrying excess body fat, especially around the organs — the pancreas compensates by pumping out more of it. This review lays out how that chain of events (insulin resistance) is directly implicated in obesity, type 2 diabetes, fatty liver disease, cardiovascular disease, and polycystic ovary syndrome — not as five separate problems, but as different downstream expressions of the same upstream issue.
Eleven women recently diagnosed with type 2 diabetes (HbA1c 8.9% at baseline) followed a ketogenic diet (under 30g carbs/day) for 90 days. Average weight loss was 9kg, HbA1c fell to 5.6% — below the diagnostic threshold for diabetes in 9 of the 11 women — and triglycerides roughly halved while HDL rose. This is a small, uncontrolled case series (no comparison group, self-selected volunteers), so it can't isolate how much of the effect was the carbohydrate restriction itself versus the weight loss it produced — but the effect sizes are large enough to be a genuinely useful data point.
This study came out of Benjamin Bikman's lab at Brigham Young University — he's a metabolic scientist whose research career has focused on insulin resistance and how chronically elevated insulin drives fat storage and disease risk. He's also a prominent public communicator on this topic (books, podcasts), so it's worth reading his popular claims alongside — not instead of — the primary research, since public-facing summaries tend to state things more confidently than the underlying papers do.
Because insulin resistance sits upstream of so many conditions, improving it tends to pay off in more places than just body weight: steadier energy through the day, better blood pressure and triglycerides, lower type 2 diabetes risk, and — for women with PCOS specifically — improvements in cycle regularity and androgen-related symptoms, since insulin resistance is a core driver there too. This is also why the "which diet" question matters less than the pattern underneath most of the effective approaches above: reducing refined carbohydrate and processed food, eating enough protein, moving regularly (particularly resistance training, covered in the next category), and sleeping enough — all of which independently improve insulin sensitivity.
Tap a card to see the mechanism, the actual studies, and where the evidence is genuinely strong versus thin. Evidence strength badges reflect how much high-quality randomised-trial data exists for fat loss specifically — not whether a diet "works" in principle.
Any macronutrient mix, eaten in a consistent deficit
Strong evidenceAny structured eating pattern — low-fat, moderate-carb, whatever suits the person — held at roughly 15–25% below maintenance calories. This is the baseline every other "diet" is really being compared against.
Adults with BMI 28–40 randomised to a healthy low-fat or healthy low-carb diet, both coached toward whole-food quality rather than a specific macro target. Average loss was 5.3 kg (low-fat) vs 6.0 kg (low-carb) — not a statistically significant difference. Pre-trial genotype and insulin-secretion patterns did not predict who did better on which diet.
Reducing total fat intake produced about 1.6 kg more weight loss than usual eating patterns across dozens of pooled trials — a real but modest effect, consistent with "any consistent deficit works."
The single biggest predictor of outcome in these trials isn't the macronutrient split — it's whether the person could stick to it for 12 months. Choose the version of "eat less" you can actually sustain, and build the rest of your approach around that.
Under ~130g carbs/day (low-carb) or under 50g/day (ketogenic)
Strong evidenceKetogenic diets restrict carbohydrate to roughly 50g/day or less (about 10% of energy), shifting the body toward fat oxidation and ketone production. "Low-carb" is a looser version of the same idea.
Very-low-carb ketogenic diets produced 0.91 kg more weight loss than low-fat diets over a year or more (95% CI −1.65 to −0.17 kg) — statistically real, but small. Triglycerides and diastolic blood pressure also improved slightly more; LDL cholesterol rose more on keto. When the analysis was limited to trials running 24 months, most differences stopped being significant.
Across 14 named diets, low-carb and low-fat approaches produced almost identical weight loss at 6 months (4.63 kg vs 4.37 kg vs usual diet). By 12 months, the weight-loss advantage of every diet pattern — low-carb included — had largely faded.
Ketogenic diets can work well for people who find fewer, more satiating meals easier to stick to, and the appetite-suppressing effect of ketosis is real for many. The weight-loss edge over other approaches is small and shrinks over time, so this is a "pick it because it suits your lifestyle," not "pick it because it's metabolically superior."
~27–35% of calories from protein, rest flexible
Strong evidenceKeeps fat and calorie targets the same as a standard diet but raises protein's share of intake, usually by trading it against carbohydrate.
Energy-matched high-protein diets produced 0.79 kg more weight loss than standard-protein diets (95% CI −1.5 to −0.08 kg), plus greater fat-mass loss and better preservation of lean mass and resting energy expenditure — the last two being the more clinically meaningful findings.
The extra weight-loss effect is small on its own, but the lean-mass-sparing effect is one of the more consistent findings in this literature — especially valuable when combined with resistance training (see the Strength Training category). A reasonable target is roughly 1.6–2.2g of protein per kg of body weight per day during a fat-loss phase.
Eating windows or alternating fast/feast days rather than daily restriction
Moderate evidenceRestricting eating to a window (e.g. 16:8 time-restricted eating) or alternating very-low-calorie "fast" days with normal or higher-calorie "feast" days, rather than a steady daily deficit.
Obese adults were randomised to alternate-day fasting (25% of energy needs on fast days, 125% on feast days), daily calorie restriction (75% of energy needs every day), or no intervention. Weight loss was statistically the same between alternate-day fasting and daily restriction at both 6 months (−6.8% vs −6.8%) and 12 months (−6.0% vs −5.3%). The fasting group had the highest dropout rate (38%, vs 29% for daily restriction) and tended to eat more than prescribed on fast days and less on feast days — in effect drifting toward ordinary calorie restriction anyway.
Fasting protocols are a scheduling strategy, not a separate fat-loss mechanism — the controlled trial evidence shows no metabolic advantage over an equivalent daily deficit, and adherence was actually harder for participants in this trial. It's worth using if compressing your eating window naturally reduces how much you eat; it's not worth adopting for a claimed metabolic edge.
Animal foods only — no plant foods, no fibre, no carbohydrate
Weak evidenceAn extreme elimination diet restricted to meat, fish, eggs and sometimes dairy, excluding all plant foods.
A social-media survey of self-selected carnivore dieters — not a controlled trial. 95% reported improved overall health and satisfaction was high, but the study has no comparison group, relies entirely on self-report, and is subject to substantial selection and recall bias (people willing to fill out a survey praising their diet are not representative). LDL cholesterol was notably elevated in respondents who reported it.
There are currently no randomised controlled trials of the carnivore diet for fat loss. Any weight loss reported is almost certainly driven by the same mechanisms as other high-protein, low-carb approaches (satiety, elimination of energy-dense processed foods, spontaneous calorie reduction) rather than anything unique to excluding all plants. This is the thinnest evidence base of any approach on this list.
Atkins, Zone, DASH, Mediterranean and 10 others, directly compared
Strong evidenceRather than testing one diet against a generic control, this review pooled every trial that compared named commercial or clinical diets against each other or against usual eating.
At 6 months, the diets with the largest weight-loss effect versus usual eating were Atkins (−5.5 kg), Zone (−4.1 kg) and DASH (−3.6 kg), all moderate-certainty evidence. By 12 months, weight-loss benefits diminished substantially across essentially every named diet, and cardiovascular risk-factor improvements largely disappeared too — with one exception: the Mediterranean diet retained its cardiovascular benefits at 12 months.
This is the closest thing to a definitive answer on "which diet is best": short-term results vary by a few kilograms between approaches, but almost none of that difference survives to 12 months except the Mediterranean diet's cardiovascular (not weight) benefits. The name on the diet matters far less than whether you're still doing it in a year.
The headline finding across this research: intensity matters less than total energy expended and whether you'll actually keep doing it. The differences between methods are real but small.
Short, hard intervals vs sustained moderate-intensity effort
Moderate evidenceHigh-intensity interval training (HIIT) alternates short bursts of hard effort with recovery periods; moderate-intensity continuous training (MICT/steady-state) holds one sustainable pace for the whole session.
Both HIIT and steady-state training produced significant reductions in body fat percentage, with no significant difference between them (HIIT −1.50% vs MICT −1.44%). For absolute fat mass, HIIT edged ahead by roughly 2.3kg on average — driven mainly by sprint-style interval protocols specifically. Worth knowing: other researchers publicly raised data-extraction concerns about some of the pooled studies in this review shortly after publication, so treat the absolute-fat-mass advantage as suggestive rather than settled.
Both approaches work. HIIT is more time-efficient per session but harder to recover from and easier to under-recover into; steady-state is easier to sustain for longer weekly volume. Pick based on which one you'll actually complete week after week — that predicts the outcome more than the protocol does.
Training on an empty stomach to "burn more fat" — does it actually?
Moderate evidenceThe theory: exercising after an overnight fast, with low circulating insulin, shifts the body toward burning more stored fat during the session.
Young women on a calorie-restricted diet were randomised to fasted or fed steady-state cardio, matched for exercise volume. Fat mass and lean mass changes were statistically the same between groups. What burns preferentially during a single session doesn't change how much fat is lost once you account for total daily calorie balance.
Fasted cardio isn't harmful and isn't magic — pick whichever fits your schedule and doesn't make you feel light-headed or under-fuelled for the session. This applies to cardio specifically; resistance training fasted has a separate, more consistent performance downside (see Strength Training).
Resistance training doesn't burn as many calories per session as cardio, but it changes what the weight you lose is made of — and that's the part that actually determines how you look and how your metabolism holds up afterward.
Same time commitment, very different effect on what you lose
Strong evidenceStructured resistance training (machines/free weights, progressive load) compared directly against aerobic training and a combination of both.
Sedentary overweight/obese adults were randomised to aerobic training (AT), resistance training (RT), or both combined (AT/RT). AT and AT/RT reduced total body mass and fat mass significantly more than RT alone. But RT and AT/RT increased lean body mass significantly more than AT alone — resistance training was the variable that determined whether the weight lost (or gained on top of fat loss) was muscle.
Cardio alone is more time-efficient for pure scale weight loss. But without a resistance training stimulus, a meaningful share of what you lose in any deficit is muscle, not fat — and that's what makes weight regain come back almost entirely as fat afterward. If you only have time for one type of training during a cut, the evidence doesn't clearly favour dropping resistance training over cardio.
More isn't automatically better — how you combine cardio and strength training changes what you get out of each.
What the STRRIDE trial found when people did both
Strong evidenceThe combined AT/RT group did not lose significantly more fat mass or total body mass than the aerobic-only group, despite roughly double the weekly time commitment. Its advantage was entirely in lean mass preservation, matching the resistance-only group on that front while also getting the fat-loss benefit of cardio.
If time is unlimited, combining both is the best body-composition outcome on the table. If time is genuinely limited, doing cardio alone loses fat just as effectively — the tradeoff is purely about how much muscle comes along for the ride, not how much fat comes off.
When and why combining the two can work against your lifting progress
Moderate evidence"Concurrent training" is doing endurance and resistance training in the same program. The concern is that endurance work can blunt the strength/muscle-building adaptations from resistance training through competing signalling pathways and accumulated fatigue.
The interference effect on strength, power and hypertrophy was real but conditional — it scaled with how frequent and how long the endurance sessions were, not with concurrent training itself. Running interfered more than cycling. Shorter, less frequent cardio sessions produced minimal interference with lifting progress.
You don't have to choose one or the other. Keep cardio sessions reasonably short, favour lower-impact modes (cycling, incline walking) over running if strength is the priority, and where possible separate hard cardio and lifting sessions by a few hours or different days.
This is the category with the most popular mythology and the least supporting evidence relative to total daily intake.
Does the "anabolic window" right after training actually matter?
Moderate evidenceA simple pooled comparison showed a small hypertrophy benefit from consuming protein close to a workout — but once the analysis controlled for total daily protein intake, that benefit disappeared. Total protein intake across the day was the strongest predictor of muscle growth, not the timing of any single dose. (A later critique noted that most of the pooled trials weren't well-designed to isolate timing in the first place, which is a fair caveat on how much weight to put on either conclusion.)
Hitting your total daily protein target (roughly 1.6–2.2g/kg during a fat-loss phase — see the High-Protein Diet card) matters far more than eating it in any particular window. A protein shake within 30 minutes of training isn't wasted, but it's also not required.
Covers both cardio and resistance training
Moderate evidenceCovered in full under Cardio → "Fasted vs fed cardio": a controlled trial found no difference in fat or lean mass outcomes between fasted and fed steady-state cardio when total intake was matched. For resistance training specifically, fasted lifting hasn't shown a body-composition disadvantage either, but eating beforehand does appear to support better strength performance during the session itself — worth knowing if you're chasing progressive overload as well as fat loss.
Neither state is "better" for fat loss specifically. If you're lifting for strength or muscle gain alongside fat loss, eating something beforehand tends to support better session quality; if it's a cardio-only session, do whichever fits your day.
Of everything covered in this tool, this is the category most people skip and most likely to be quietly capping their results.
Same calorie deficit, very different composition of the weight lost
Strong evidenceOverweight adults followed the same moderate calorie restriction for 14 days under two conditions: 8.5 hours or 5.5 hours of sleep opportunity per night. Identical calorie deficit, but sleep-restricted participants lost 55% less body fat and 60% more lean mass — meaning most of their "weight loss" was muscle, not fat. They also reported more hunger.
In a real-world weight-loss program, every extra hour of sleep per night was independently associated with about 0.7kg more fat loss over the course of the program, and better sleep quality independently predicted greater fat loss too — even after adjusting for calorie intake.
Sleep isn't a soft "self-care" add-on to a fat-loss plan — the controlled trial shows it directly determines what a calorie deficit actually costs you: fat, or muscle. If you can only fix one non-diet, non-training variable, this is the one with the most direct evidence behind it.