StratAlp's nutrition model, in the open.

Once you've imported a route and broken it into segments, StratAlp proposes a full hydration and fueling plan. It reads the profile — distance, gain, loss, the slope of each segment — and derives hourly targets: water, sodium, carbs, sometimes caffeine and protein. We laid the foundations at the beta launch; here's the machinery behind the plan, its sources as much as its limits.

01 / The white box

Plenty of tools "generate" a plan and stop there: you get numbers, never their reasons. We made the opposite choice. Every decision the engine makes points back to a source — sixteen scientific references, their DOIs, listed in the app next to the plan and gathered at the foot of this article. The sheet also shows the rules it triggered ("reinforced intake before the big climb", "switch to liquid after four hours") and the input values it used. Nothing comes out of a hat.

02 / The real cost of the climb

Two 30 km races aren't run — or fueled — the same way if one climbs 2,500 m and the other 600. To estimate the cost, the engine doesn't look at distance alone: it computes the energy cost of each segment from its slope, using the equation from Minetti et al. (2002) — the standard measured on mountain runners, from −45% to +45% grade. A steep climb costs a lot; a long descent, far less, but never zero. From there come calories, a duration, and needs tied to the real terrain rather than a flat average.

03 / How many carbs, really

The right carb dose depends on how long the effort lasts. StratAlp places the race using the ITRA km-effort — distance plus gain divided by a hundred — then sets an hourly target: 45 g on a short race, 60 g on a medium one, 75 g on a long one, within the ranges recommended for ultra by the International Society of Sports Nutrition (Tiller et al., 2019), in line with the reference carbohydrate targets (Burke et al., 2011). You stay free to adjust it, from 20 to 120 g/h. Past 60 g, a single sugar is no longer enough: its intestinal transporter saturates. The engine then keeps the 2-to-1 glucose-fructose ratio described by Jeukendrup (2014), which opens a second absorption pathway and lets up to 120 g/h through — a rate Urdampilleta et al. (2020) measured on the trail, and one that assumes a gut trained for it (Jeukendrup, 2017).

04 / Water, sodium, timing

Hydration starts at 500 ml/h, then rises when it's hot or you climb high: heat and altitude dry you out faster. These settings follow the work of Hoffman & Stuempfle (2014) on a 161 km race and of Périard et al. (2015) on heat acclimation. Sodium follows, more generous on long distances and in strong heat. Placement matters as much as amounts, and the ultra literature is clear on it (Costa et al., 2019; Arribalzaga et al., 2021; Kerksick et al., 2017): first intake at T+30 min, a gel placed ten minutes before a big climb so the fuel is there when demand rises, a switch to liquid past four hours when sweet turns sickening, a salty-sweet alternation to keep the palate fresh — the very shift toward salt and fat that Stuempfle et al. (2011) observe in finishers.

On the longest formats, two more levers come in. Caffeine, framed by the IOC consensus (Maughan et al., 2018) and by Hodgson et al. (2013) on endurance: a load before the start, micro-doses on the move, a final hit before the finish. Then protein, past six hours of effort, to limit muscle breakdown (Knechtle & Nikolaidis, 2018; Phillips & Van Loon, 2011).

05 / You keep control

A plan that has you swallow brands you've never tested is useless. StratAlp draws from your inventory: your gels, your bars, your drink, your water, your recovery products. Intake is set to the aid stations you've already placed on the route. Above all, nothing is written into your plan until you've seen the preview and said yes. The engine prepares a proposal; the last word is yours.

06 / What the model can't do yet

We care as much about stating the limits as about showing the strengths. They're written in plain sight in the app, right next to the sources. Here they are.

• Heuristics, not learning. The model doesn't adjust from your past races yet — that tuning is coming in a later version.
• Estimates, not measurements. Totals equal hourly target × duration. What your body actually absorbs varies from one runner to the next.
• An average sweat rate (~1 L/h). Hydration isn't individualized: adjust it to your own experience and the real weather.
• Population averages. Caffeine, protein and gut-tolerance thresholds rest on declared markers, not on your physiology.
• Above 3,000 m, and on the bike. The altitude modulation isn't calibrated at high elevation. The model is built for running — Minetti's equation doesn't transfer as-is to cycling.
• Allergen filtering by heuristic. It relies on product names and categories, not certified labels. Double-check every product.
• An indicative plan, not medical advice. It's a suggestion to validate, not an instruction. Stay tuned in to your body.

These limits aren't excuses — they're appointments. As the beta moves forward, we calibrate what isn't yet, and learning from your races will come. In the meantime, you know exactly what you're holding: a serious, sourced starting point that you adjust to how you feel. One less question to carry on race day.

The sixteen references

Here they are, in the order the engine calls on them. Each carries its DOI — the same list shown in the app, next to your plan.

1. Minetti, A.E. et al. (2002). Energy cost of walking and running at extreme uphill and downhill slopes. Journal of Applied Physiology, 93(3), 1039-1046. DOI: 10.1152/japplphysiol.01177.2001.
2. Jeukendrup, A.E. (2008; 2014). Carbohydrate feeding during exercise; A step towards personalized sports nutrition: carbohydrate intake during exercise. European Journal of Sport Science; Sports Medicine. DOI: 10.1080/17461390801918971; 10.1007/s40279-014-0148-z.
3. Tiller, N.B. et al. (2019). ISSN Position Stand: nutritional considerations for single-stage ultra-marathon training and racing. Journal of the International Society of Sports Nutrition, 16, 50. DOI: 10.1186/s12970-019-0312-9.
4. Urdampilleta, A. et al. (2020). Effects of 120 vs. 60 and 90 g/h carbohydrate intake during a trail marathon. Nutrients, 12(7), 2094. DOI: 10.3390/nu12072094.
5. Costa, R.J.S. et al. (2019); Arribalzaga, S. et al. (2021). Nutrition for Ultramarathon Running; Carbohydrate intake during ultra-trail and GI problems: a systematic review. IJSNEM; IJERPH. DOI: 10.1123/ijsnem.2018-0255; 10.3390/ijerph18115737.
6. Hoffman, M.D., Stuempfle, K.J. (2014). Hydration strategies, weight change and performance in a 161 km ultramarathon. Research in Sports Medicine, 22(3), 213-225. DOI: 10.1080/15438627.2014.915838.
7. Périard, J.D. et al. (2015). Adaptations and mechanisms of human heat acclimation. Scandinavian Journal of Medicine & Science in Sports, 25(S1), 20-38. DOI: 10.1111/sms.12408.
8. ITRA. km-effort convention = distance (km) + gain (m) / 100. Trail-community standard: itra.run.
9. Maughan, R.J. et al. (2018). IOC consensus statement: dietary supplements and the high-performance athlete. British Journal of Sports Medicine, 52(7), 439-455. DOI: 10.1136/bjsports-2018-099027.
10. Hodgson, A.B., Randell, R.K., Jeukendrup, A.E. (2013). The metabolic and performance effects of caffeine compared to coffee during endurance exercise. PLOS ONE, 8(4), e59561. DOI: 10.1371/journal.pone.0059561.
11. Knechtle, B., Nikolaidis, P.T. (2018). Physiology and pathophysiology in ultra-marathon running. Frontiers in Physiology, 9, 634. DOI: 10.3389/fphys.2018.00634.
12. Phillips, S.M., Van Loon, L.J.C. (2011). Dietary protein for athletes: from requirements to optimum adaptation. Journal of Sports Sciences, 29(S1), S29-S38. DOI: 10.1080/02640414.2011.619204.
13. Stuempfle, K.J. et al. (2011). Race diet of finishers and non-finishers in a 161 km mountain footrace. Journal of the American College of Nutrition, 30(6), 529-535. DOI: 10.1080/07315724.2011.10719999.
14. Jeukendrup, A.E. (2017). Training the gut for athletes. Sports Medicine, 47(S1), S101-S110. DOI: 10.1007/s40279-017-0690-6.
15. Burke, L.M. et al. (2011). Carbohydrates for training and competition. Journal of Sports Sciences, 29(S1), S17-S27. DOI: 10.1080/02640414.2011.585473.
16. Kerksick, C.M. et al. (2017). ISSN position stand: nutrient timing. Journal of the International Society of Sports Nutrition, 14, 33. DOI: 10.1186/s12970-017-0189-4.

— The StratAlp team