How to adjust nutrition based on HRV for recovery

Look at your phone's home screen. In one folder, you have an intelligence engine built on millions of data points, tracking your heart rate variability, respiratory rate, and sleep architecture down to the minute. In another folder, you have a digital ledger from 2014 asking you to scan a barcode.

This is the structural deficit in modern endurance training. We measure the strain in fine detail. We fuel the engine using static math.

Recovery is the rate-limiting feature of human performance. You cannot out-train your ability to adapt, and you cannot adapt without matching substrate to strain.

The signal is already in your pocket

A modern wearable produces more physiological data in a single day than most elite competitors generated in an entire career a decade ago. The algorithms are trained on millions of nights. The problem is not the signal.

The problem is where the signal ends up.

Apple Health, Whoop, Oura, Garmin, and Polar all expose this data. They all ship beautiful dashboards. None of them can tell your kitchen what to do about it. The signal ends in a graph.

An open loop and a dead end

Every control system is either open loop or closed loop. An open loop produces a reading and stops. A closed loop takes the reading, acts on it, and measures the response. Most consumer health tech is open loop by design. You are supposed to be the controller.

In the open loop, the signal dies in your dashboard. You scroll, nod, and decide to "take it easy today." What that means for your protein distribution, your carb timing, or your evening hydration is left entirely to intuition. That is not a system. That is a mood.

Why 800 kcal is never just 800 kcal

The flaw in a calorie-first model is that every workout is treated as a single lumpy number. It isn't. Two workouts that burn the same total energy can leave behind completely different biochemical debts.

In this illustration, the Zone 2 ride leaves roughly a 12-hour recovery window. It is forgiving. A modest protein meal, normal carbs, and you are ready again. The threshold workout leaves something nearer 48 hours. It calls for prompt carbohydrate, protein spread through the day, and a pre-bed protein feed that can add to overnight muscle repair.

Your food log treats both identically. Your wearable can already tell the difference in the first ten minutes of the workout.

The three live signals KEXBI actually reads

Of the ten or so channels a modern wearable produces, three can move the plan. They fire rarely, the changes are small, and each one is offered, not applied: it maps to a concrete kitchen decision you can accept or ignore.

None of these adjustments require a human to read a graph. None of them require you to know what cortisol is doing. The engine reads the overnight numbers, does the math, and offers the updated plan before you finish your coffee.

The hydration blind spot

Fluid replacement is the cleanest example of the gap. The research guidance for post-exercise rehydration is 1.5× the fluid volume lost (Shirreffs et al., 1996), because the kidneys keep excreting urine while you drink. If you lose 1 kg of water weight in a hot workout, you need roughly 1.5 litres to restore homeostasis.

Static nutrition apps give you an arbitrary daily goal, usually around 2.5 litres. They don't know you ran in 28°C heat. They don't know your sweat rate. They don't know you finished at 19:00, which means that 1.5 litres needs to land between 19:00 and 21:00 if you want to protect deep sleep. Drink it at 22:00 instead and waking in the night to urinate can cut into the sleep your recovery depends on.

This is not complicated math. It is simply math that few apps have bothered to build, because the two apps that would need to talk to each other do not.

One day, worked through

Picture a Tuesday. You log a 55-minute threshold ride at 17:30, finishing at 430 watts average, burning 780 kcal. Your Whoop flags it as a strain score of 17.2. You wake up Wednesday and HRV is 14% below baseline. Sleep efficiency came in at 82%.

A standard nutrition app sees: yesterday you earned 780 extra calories, today you have your normal 2,400 kcal target.

KEXBI sees the same data and offers a set of changes. Tuesday's post-workout window gets a 70g fast carbohydrate plus 30g whey top-up. Dinner carbs get lifted by 40g, leaning low-GI. A 40g casein dose lands 90 minutes before bed. After the broken night, Wednesday's plan suggests an easier ride and offers to rebuild the day around recovery. Hydration targets rise by 600 ml for the day and finish by 20:30.

Same person. Same wearable. Same calorie target. A noticeably different plan by Thursday morning, on the uncommon day when all three signals fire at once. Most days, none of them do.

The point

The wearable industry solved sensing. It did not solve acting. Every major consumer device now produces data of a fidelity that would have required a sleep lab twenty years ago, and most consumer nutrition apps treat that data as if it did not exist.

Every wearable on the market has solved sensing. KEXBI is built to work out what to do about it.

• Shirreffs SM, Taylor AJ, Leiper JB, Maughan RJ. Post-exercise rehydration in man: effects of volume consumed and drink sodium content. Med Sci Sports Exerc 1996.
• Res PT et al. Protein ingestion before sleep improves postexercise overnight recovery. Med Sci Sports Exerc 2012.
• Stanley J, Peake JM, Buchheit M. Cardiac parasympathetic reactivation following exercise: implications for training prescription. Sports Med 2013.
• Plews DJ, Laursen PB, Stanley J, Kilding AE, Buchheit M. Training adaptation and heart rate variability in elite endurance athletes. Sports Med 2013.
• Kerksick CM et al. ISSN position stand: nutrient timing. J Int Soc Sports Nutr 2017.
• Spiegel K, Knutson K, Leproult R, Tasali E, Van Cauter E. Sleep loss: a novel risk factor for insulin resistance and type 2 diabetes. J Appl Physiol 2005.