Energy Availability vs Calorie Deficit for Athletes
Standard tracking apps optimise for weight loss, but ignoring your true energy availability silently shuts down the cellular machinery required to adapt.
~5 MIN · KEXBI RESEARCH
Standard tracking apps optimise for weight loss, but ignoring your true energy availability silently shuts down the cellular machinery required to adapt.
The training block did not stall because you were lazy. It stalled because the machinery that turns training into adaptation ran out of the raw material it needs to work. You cannot build on empty. That is not a slogan. It is a measured biochemical fact and the calorie app on your phone was not built to notice.
When a well-designed training block works, the visible result is a faster time, a heavier lift, a leaner shape. The invisible result underneath is a signal at the tissue level: repeated bouts of loading, plus available fuel and protein, drive up the rate at which the body builds new contractile and structural protein. This is called muscle protein synthesis, and it is measurable in a lab in a way that a stopwatch cannot capture.
Three independent controlled studies, using tracer methods that let researchers watch synthesis directly, converge on the same finding. Sustained energy deficit turns synthesis down.
There is a rescue lever, and it is the part that matters for anyone trying to build during a period of restricted intake. A single bout of resistance exercise plus a protein feed after it lifts synthesis back toward energy-balance rates, even inside a deficit [1,3]. Higher protein intake, in the region of twice the RDA, roughly 1.6 to 2.4 g per kg body weight per day, preserves fat-free mass during a deficit far better than RDA-level intake, with little added benefit beyond about twice [4]. In other words: training and correctly distributed protein defend the very thing the deficit is dismantling. A tracker that is only watching the calorie number cannot know it is switching one system off and cannot help you switch it back on.
The number the field uses to describe "enough" is energy availability. It is not TDEE minus intake, and it is not the daily calorie budget in a weight-loss app. It is a specific calculation.
Two nuances the app in your pocket will not tell you. The original studies expressed the denominator as lean body mass; most modern reviews and the IOC consensus use fat-free mass. The numeric difference is small but the terminological inconsistency is real. And "45 optimal, 30 low" are useful reference values, not physical constants. Only the sub-30 low-EA band was derived from a controlled experiment. The others are working conventions.
The 30 kcal per kg line that shows up in every REDs discussion has one origin: a 2003 study by Loucks and Thuma [6]. Twenty-nine regularly-menstruating, habitually-sedentary young women, in a five-day controlled feeding-and-exercise setting, produced the finding that luteinizing-hormone pulsatility was unaffected at 30 kcal per kg lean body mass per day and fell below it.
That is the entire foundation of the number. A five-day laboratory manipulation, in one specific population, using one female reproductive marker. Extrapolating it as a universal daily floor for men, for trained athletes, or for chronic real-world exposure goes past what the data show.
A companion study on bone metabolism landed nearby but not identically. Ihle and Loucks 2004 [8] found bone-formation markers suppressed by all the restricted-EA conditions they tested, with formation declining even at 30 kcal per kg. So the neat "fine above 30, broken below" story is true for reproductive hormones. It is not true for bone. There is no single universal cut-off.
The sex difference is where this gets sharpest. The one controlled study in men used the same protocol and the same low EA, and the results diverged.
The honest sentence is this: men appear to tolerate lower energy availability before the same disruptions appear, and a precise male floor has not been established. Any app that hard-codes 30 as a universal line for both sexes is defending a female-derived reproductive threshold as if it were a physical constant. It is not, and the reader deserves to know that.
The muscle-synthesis effect is the mechanism. The rest of the cost shows up as things you can see and things that take you out of training.
Bone stress injury. The most concrete performance cost is the injury that removes you from training entirely. Bone-formation markers are suppressed dose-dependently as EA falls [8]. Prospectively, cumulative Female Athlete Triad risk factors, of which low EA is the driver, track with bone-stress-injury incidence. In a multi-site cohort the highest-risk combination reached a 46 percent BSI incidence over the season, against a background rate of about 11 percent [12]. A separate cumulative-risk score predicted BSI, with moderate-risk athletes about twice and high-risk athletes about four times more likely to be injured than low-risk peers [13].
Direct performance decrement. Softer evidence than the mechanism, clearly labelled as such. In a controlled study, inducing low EA at around 22 kcal per kg for 14 days in trained men significantly reduced explosive power before hormonal changes appeared [14]. Over a 12-week training block, ovarian-suppressed low-EA swimmers showed a nearly 10 percent decline in 400 m velocity while cyclic peers improved by about 8 percent [15]. Amenorrheic endurance athletes showed lower knee strength and slower reaction time than eumenorrheic peers [16]. The performance case is real; it is thinner than the biochemistry, and we present it as such.
Availability. In elite athletes preparing for Rio 2016, low EA sat among the top correlates of illness that pulled athletes out of training days [17]. Days you cannot train are days you cannot adapt.
The mainstream trackers are not lazy about this. They were built for a different job. When their goal-setting logic is read literally, the design intent is clear: help someone lose weight by hitting a daily calorie target. That is a coherent product. It is not the product an athlete needs.
| App | Core design (per their own docs) | Computes EA? |
|---|---|---|
| MyFitnessPal | Sets goals by subtracting a fixed deficit tied to a weekly rate of loss. Frames the day as "a daily budget of calories to spend." Net Calories = Calories Consumed − Calories Burned (Exercise). Exercise is added back so you can "eat more." | No |
| Noom | Centres a single personalised daily calorie budget. Overlays a red / yellow / green calorie-density system that sorts foods but does not replace the calorie target. | No |
| Cronometer | Deep micronutrient tracking. In their own words, the "Energy Target is derived from your Weight Goal." Organising number remains a calorie target. | No |
There is a specific inversion worth noticing. Energy availability subtracts exercise energy from intake and judges what is left against fat-free mass. MyFitnessPal's Net Calories adds exercise energy back to your budget, letting you eat more on hard days. The two formulas point in opposite directions. For weight loss in a sedentary or lightly active user, the net-calories model works. For an athlete trying to protect adaptation, it steers straight into the deficit that stops adaptation.
Even the calorie total the apps do produce carries known limits. A peer-reviewed validation of MyFitnessPal found it broadly accurate for total energy and macronutrients but significantly underestimating sodium and cholesterol, and recommended budgeting a 5 to 10 percent data-quality loss on user-contributed entries [18]. A single number, even tracked diligently, is a blunt instrument for a job that needs two more numbers underneath it.
The prevalence data lands where the reader lives. Across athlete populations, low energy availability sits between about 22 and 58 percent, in both sexes [19]. A 2024 systematic review of 46 studies covering roughly 6,100 athletes put pooled low-EA prevalence at about 45 percent. Recreational, non-elite exercisers show the same pattern: a cohort study of competitive but recreationally-trained male endurance athletes found a mean EA of 28.7 kcal per kg FFM, below the low band as a group, and the authors concluded low EA "is not restricted only to elite male athletes."
Read plainly: if you train seriously as an amateur, the base rate of running under your own floor is somewhere near coin-flip. This is not an edge case. It is the modal outcome of following a weight-loss tracker while training like an athlete.
The counter-move is not a different diet philosophy. It is a different organising number. KEXBI is built around energy availability, not a daily calorie target.
Rather than defend a borrowed 30, the engine calibrates your availability floor to your sex and your body composition, and holds your intake above it. Session energy cost is estimated from the workout on your calendar and applied to that day's plan, so the budget moves with the training rather than sitting at a flat number the app cannot see through. Protein is distributed across meals in a way that supports muscle protein synthesis; the rescue lever from Chapter 01 is built in rather than assumed. Fat-free mass is re-estimated as you go, so the calculation does not silently drift.
If you try to configure a deficit that would strip the energy your body needs to adapt, the engine raises the floor instead. That is the honest expression of what "you cannot build on empty" means in software: refusing to hand you a plan that turns off the machinery you are training to switch on.
Training is the request. Fuel is the permission. The number that decides whether the request becomes progress is not the calories in your app. It is the energy left for your body after training has taken its cut, judged against the mass that is doing the work. Track that number, calibrate it to yourself rather than to a borrowed constant, and the block you are grinding through starts to give back what you put in.