How big should your calorie deficit be?

The 500-calorie rule is a ceiling, not a starting point#

The usable answer is a share of your own maintenance rather than a number you inherited from a magazine: somewhere around 20% below what you burn, and lower if you are already lean or training hard. What makes that more than a vibe is where the ceiling sits. A meta-regression of randomized trials found that an energy deficit of about 500 kcal per day was the point at which lean-mass gains stopped entirely, and its authors' recommendation is blunt — people resistance training to hold onto muscle during weight loss "should avoid energy deficits >500 kcal day⁻¹"¹.

Sit with that for a second, because it inverts the folk wisdom. The 500-calorie deficit is not the moderate on-ramp everyone treats it as. It is roughly the edge. Almost every "lose a pound a week" plan starts you at the maximum the evidence supports and calls it conservative — and the plans that go bigger are past the point where the trials stop showing a benefit and start showing a cost. If you already know why a deficit is the mechanism, this article is about the one dial you actually get to set.

Two methods, a decade apart, land on the same number#

The striking thing is that this threshold shows up twice, from two research designs that share nothing but the topic.

The first is the meta-regression above. Murphy and Koehler pooled randomized trials of resistance training performed in an energy deficit for at least three weeks. Against control groups training without a deficit, lean-mass gains were clearly impaired (effect size −0.57, p = 0.02) while strength gains were statistically indistinguishable (ES −0.31, p = 0.28). Pooling everything into a meta-regression of deficit size against lean mass, the line crossed zero at roughly 500 kcal per day.

The second is a Norwegian randomized trial that never mentions a kcal threshold at all. Garthe and colleagues took 24 elite athletes — all doing four resistance sessions a week — and assigned them to lose bodyweight either slowly (0.7% per week) or fast (1.4% per week), each on an individually built diet². The trial reports what those rates cost in energy:

The slow group's deficit was 469 kcal/day. Murphy and Koehler's ceiling was ~500. One is an RCT in Olympic-level athletes measured by DEXA; the other is a meta-regression across the literature published eleven years later. Neither cites the other's number as a target. They agree anyway, and that convergence is worth more than either result alone.

Look at what the two Garthe groups bought for that difference. They lost the same bodyweight — 5.6% versus 5.5%. But the slow group lost half again as much fat (31% versus 21% of fat mass) and gained 2.1% lean mass, while the fast group gained nothing. Same scale reading, different bodies underneath. Among the men specifically, the fast group actually trended toward losing lean mass (−2.0% ± 1.0%, p = .1) while the men losing slowly gained 1.7% ± 0.4% — though with six and five men respectively, that subgroup is a hint, not a finding.

And Garthe's own authors flag the confound rather than hiding it: the slow group also spent about three weeks longer in the study, so they simply trained longer, which the paper calls "likely the most important explanation" for the lean-mass gap. That matters, and it is the reason this trial alone would not settle anything. The answer to it is inside the same paper: measured per week, lean mass rose 0.3% ± 0.0% in the slow group and 0.0% ± 0.1% in the fast one (p = .02). Time alone does not explain a difference that survives dividing by time.

Strength holds while muscle goes — which is why you won't notice#

Here is the part of Murphy and Koehler's result that should change how you self-monitor. Lean mass was impaired by the deficit. Strength was not: ES −0.31 with p = 0.28, meaning the trials could not distinguish strength gains in a deficit from strength gains without one. Garthe found the same dissociation from the other side — 1RM squat improved in both groups, by 11.9% and 8.9%, even though only the slow group added lean mass. (The separation appeared in upper body: bench press rose 13.6% ± 1.1% slow versus 6.4% ± 3.3% fast, p = .01.)

So the instrument most lifters trust to tell them their deficit is safe — the weight on the bar — is the one instrument that keeps reading fine while the thing you are trying to protect erodes. Early strength gains are substantially neural, and neural adaptation does not need a calorie surplus. A deficit that is too big will not announce itself in your training log for a long time.

A flat 500 is a different diet for every body#

Which exposes the real problem with any fixed number. Run the same 500 kcal against different maintenance levels and it is not the same intervention at all — this next line is my own arithmetic on their figures, not something either paper prints: 500 kcal is 25% of a 2,000-kcal maintenance, 20% of 2,500, and about 17% of 3,000. One person's moderate cut is another person's crash diet, wearing an identical label.

Garthe's trial dodged this by prescribing a percentage of bodyweight per week rather than a flat number, which is why the smaller athletes in it got smaller deficits automatically. Do the same thing with your intake: pick your deficit as a fraction of your own maintenance estimate — the reasons that estimate is a range rather than a number are in TDEE explained — and you get scaling for free.

There is a physiological argument pointing the same way. A review of metabolic adaptation in athletes concludes that "large caloric deficits are also likely to induce greater losses of LBM" and that "it is likely that the magnitude of these adaptations are proportional to the size of the energy deficit," recommending "the smallest possible deficit that yields appreciable weight loss"³. Note the authors' own hedging — likely, twice. Proportionality here is a well-reasoned inference from the mechanisms, not a measured dose-response curve, and it deserves to be read at exactly that strength.

Setting yours, and checking it#

A workable procedure, in order:

1. Estimate maintenance, then take ~15–20% off it. For most people that is 300–500 kcal. If your maintenance is under about 2,000, the percentage keeps you honest where a flat 500 would not.
2. Do not go past ~500 kcal/day if you lift. That is the number both lines of evidence above converge on, and past it you are trading muscle for time.
3. Put protein and resistance training underneath it. Both Garthe groups trained four times a week and ate 1.4–1.6 g/kg of protein; the deficit decides how much you lose, and protein decides what you lose — see also how much protein it takes to build muscle.
4. Judge it on the scale trend, not the bar. Strength will not warn you. Bodyweight over weeks will, which is the subject of a safe rate of weight loss.
5. Remember the number is a hypothesis. Your maintenance estimate and your food log each carry real error — audited in how accurate calorie counting really is — so the deficit you set and the deficit you get are different quantities. Correct the target from the trend, not the other way around.

If you take one thing: the aggressive deficit is not the fast route to the same place. In Garthe's data it was the route to the same scale number with less of your muscle attached.

FAQ#

Is a 500-calorie deficit too much?#

It is at the limit rather than over it, for most people. A meta-regression of resistance-training trials put the point where lean-mass gains disappear at roughly 500 kcal per day and advised against exceeding it¹. If your maintenance is low — say 1,800–2,000 kcal — then 500 is a 25%+ cut and worth trimming to 300–400.

Should I set my deficit as a percentage or a fixed number of calories?#

A percentage, because a fixed number is a different diet for different bodies: 500 kcal is 25% of a 2,000-kcal maintenance but 17% of a 3,000-kcal one. Percentages scale automatically with your size, your activity, and your changing weight, which is exactly what a fixed number cannot do.

How do I know if my calorie deficit is too big?#

Not from your lifts — that is the trap. Strength gains in a deficit were statistically indistinguishable from strength gains without one, even in trials where lean mass was measurably impaired. Watch the rate instead: losing much faster than about 0.7% of bodyweight a week, feeling training recovery slide, or seeing the scale fall while your lifts stall flat are better signals than the weight on the bar.

Sources#

1. Murphy C, Koehler K. Energy deficiency impairs resistance training gains in lean mass but not strength: a meta-analysis and meta-regression. Scand J Med Sci Sports. 2022;32(1):125-137.
2. Garthe I, Raastad T, Refsnes PE, Koivisto A, Sundgot-Borgen J. Effect of two different weight-loss rates on body composition and strength and power-related performance in elite athletes. Int J Sport Nutr Exerc Metab. 2011;21(2):97-104.
3. Trexler ET, Smith-Ryan AE, Norton LE. Metabolic adaptation to weight loss: implications for the athlete. J Int Soc Sports Nutr. 2014;11:7.
4. Hall KD, Sacks G, Chandramohan D, Chow CC, Wang YC, Gortmaker SL, Swinburn BA. Quantification of the effect of energy imbalance on bodyweight. Lancet. 2011;378(9793):826-37.