# Training Fueling Calculator — Validated Spec > A tool to calculate how much **carbohydrate** and **sodium** to prep for a training session, driven by session type, duration, and weather. This document is the fact-checked formulation and feature spec; the tool is **built** at `index.html`. _Last updated: 2026-10-29_ --- ## 1. What it is A **single self-contained `index.html`** page (HTML + CSS + JS, no dependencies) that you can open on your phone before a session. You enter the session and conditions; it outputs the total carbs and sodium to prep, broken down into actual ingredients to weigh out. ## 2. Inputs | Input | Type | Notes | |---|---|---| | Session type | dropdown | Recovery/Easy, Long (steady), Tempo/Threshold, Intervals/VO₂, Race/Time trial | | Duration | minutes | primary carb driver | | Date | date picker | used to fetch forecast | | Location | lat/lon or geolocation | required for forecast | | Temperature | °C | auto-filled from forecast, editable | | Humidity | % | auto-filled from forecast, editable | | Body weight | kg | scales sweat rate; notes projected deficit >2%, warns >5% body-mass loss | | Gut-trained toggle | on/off | raises carb ceiling 90 → 120 g/h | | Salty-sweater toggle | on/off | raises sweat sodium concentration | | Carb format | dropdown | Glucose / 1:0.8 glucose:fructose / Sucrose | | Sodium format | dropdown | Table salt / Sodium citrate | **Weather:** date + location → **Open-Meteo** forecast API (free, no key, hourly temp + relative humidity, ~16-day horizon). The forecast **auto-fetches when a location is selected** (pick a suggestion, type + blur, or 📍 current location) and re-fetches when the date or start time changes — no button. If date is beyond the forecast window, fall back to manual entry / seasonal averages. **Workout entry (optional):** instead of picking session type + duration by hand, paste a planned workout (time- or distance-based) and the tool derives both and auto-fills them — see §6. --- ## 3. Carbohydrate model **Two main drivers: duration × session type.** Duration sets the band (from the guidelines); session type positions within the band (higher intensity → faster glycogen turnover → fuel toward the top). Intake never exceeds the gut absorption ceiling. **Duration bands (non-overlapping):** `< 45` · `45–75` · `75–150` · `> 150` min. The tool **interpolates by exact duration within each band** to avoid hard jumps. **Sub-65-min cutoff + 65–75-min taper (non-race).** For any session that is **not** a race and lasts **≤ 65 min**, the tool recommends **zero in-session carbohydrate and zero sodium**. Between **65 and 75 min** both are **tapered in linearly** (0% at 65 → 100% of the model value at 75 min), so there is no hard jump at the cutoff (previously a 65-min tempo gave 0 g and a 66-min tempo ~47 g). This also matches the guideline band, which puts in-session need at ~0 for efforts up to ~45–75 min. Endogenous glycogen (~400–500 g muscle + liver, ~1,600–2,000 kcal) comfortably fuels sub-maximal efforts up to ~75–90 min, so exogenous carbohydrate adds little for shorter training bouts; likewise sub-hour electrolyte losses are modest and easily replaced at the next meal. **Races are exempt:** for hard ~45–75 min efforts a carbohydrate mouth-rinse / small intake can improve performance via a *central-nervous-system* mechanism (oral CHO receptors), independent of any metabolic fuel need — see §8. ### Standard matrix — ceiling 90 g/h | Session type | < 45 m | 45–75 m | 75–150 m | > 150 m | |---|---|---|---|---| | Recovery / Easy | 0 | 0 | 30 | 60 | | Long (steady) | 0 | 15 | 40 | 70 | | Tempo / Threshold | 0 | 25 | 50 | 90 | | Intervals / VO₂ | 0 | 25 | 50 | 90 | | Race / Time trial | 0–30 | 30 | 60 | 90 | ### Gut-trained matrix — ceiling 120 g/h Rule: cells that are **absorption-limited (≥ 45 g/h)** scale ~1.33× toward the 120 ceiling; cells that are **demand-limited (≤ 40 g/h)** stay unchanged (no metabolic need to eat more just because the gut can absorb more). | Session type | < 45 m | 45–75 m | 75–150 m | > 150 m | |---|---|---|---|---| | Recovery / Easy | 0 | 0 | 30 | **80** | | Long (steady) | 0 | 15 | 40 | **95** | | Tempo / Threshold | 0 | 25 | **65** | **120** | | Intervals / VO₂ | 0 | 25 | **65** | **120** | | Race / Time trial | 0–30 | 30 | **80** | **120** | **Total carbs to prep (g) = chosen rate (g/h) × duration (h).** ### Carb format dropdown — affects ceiling AND ingredient breakdown | Format | Absorption ceiling | Breakdown of total carb (g) | |---|---|---| | Glucose (single SGLT1 transporter) | **~60 g/h** | 100% glucose powder (dextrose monohydrate ≈ 91% carb) | | 1:0.8 glucose:fructose (multi-transportable) | up to 90 (120 gut-trained) | glucose = 1/1.8 (55.6%), fructose = 0.8/1.8 (44.4%) → e.g. 90 g = 50 g glucose + 40 g fructose | | Sucrose / table sugar (glucose+fructose disaccharide) | up to ~90 g/h | 100% carb → grams = total carb | **Rule:** when the recommended rate **> 60 g/h**, the **Glucose** option is disabled/greyed out, with a note: _"Rate exceeds glucose-only absorption — use 1:0.8 glucose:fructose or sucrose."_ --- ## 4. Fluid & sodium model > **Key correction (v2):** total sweat *losses* are NOT what you consume during the run. > You only *partially* replace fluid and sodium while running, capped by what you can > actually drink. The tool reports **recommended intake** as the headline number, with > total losses shown only as context. (v1 wrongly presented total losses as intake.) **Estimated sweat loss — context only, informational:** - `Sweat loss (L) = Sweat rate (L/h) × Duration (h)` - `Sodium lost (mg) = Sweat loss (L) × Sweat sodium concentration (mg/L)` - **Sweat rate** anchored to the validated **0.5–2.0 L/h** range (Baker 2017; ACSM 2007), scaled by intensity and weather (temp ↑, humidity ↑ → more sweat); kept inside the envelope, directionally consistent with Gonzalez 2009. - **Sweat sodium concentration:** ~**1,000 mg/L** default (42.9 mmol/L marathoner mean), salty-sweater **~1,500 mg/L**. **Recommended fluid intake — actionable:** - `Intake rate (L/h) = min(sweat rate, 1.0 L/h cap)` → `Fluid to carry (L) = intake rate × duration` - Guidance is **drink to thirst**, up to **~1.0 L/h in the heat** — the gastric-emptying limit. Warm conditions (≤35 °C) don't impair emptying unless you're already dehydrated; only extreme heat (~49 °C) or hypohydration cut it ~20–25%. Never exceed your sweat rate — over-drinking causes hyponatremia. - Finishing **2–3% of body mass down is normal** on long/hot efforts; rehydrate afterwards. The tool flags only large (>5%) projected deficits. **Recommended sodium intake — the "to prep" number, actionable:** Scales with losses so it responds to heat, then bounded by tolerance and palatability: - `Loss rate (mg/h) = sweat rate × sweat-sodium concentration` - `Target (mg/h) = clamp(loss rate × 60%, 300, 1500)` — partial replacement (top up after the run); floored 300, capped 1,500 mg/h for GI tolerance. - `Sodium to add (mg) = min(target, intake rate × 1,500 mg/L) × duration` — delivered in the drink up to a **1,500 mg/L** palatability cap. - **Salty-sweater** raises sweat-sodium (1,000 → 1,500 mg/L) → higher loss rate → higher target. - Sodium matters most for efforts **> ~2 h**, **high sweat rates (>1.2 L/h)**, or **salty sweaters**; for short cool runs it's optional — the tool shows an "optional" note when the session is **< 2 h, sweat rate < 1.2 L/h and not salty-sweater**. - The 65–75-min non-race taper (§3) applies to the sodium target too. > **Why concentration scales (not just fluid):** in heat you can't drink more than ~1 L/h, so to replace more sodium the *drink* gets saltier — up to 1,500 mg/L (Johan's chosen cap). Palatability and voluntary intake drop above ~50 mmol/L (~1,150 mg/L), so the tool flags when the mix is strong. Beyond the drink's capacity, the excess should come from **salt capsules taken with water** — *parked; drink-only for now*. Replacement fraction (60%), the 1,500 mg/h ceiling, and the 1,500 mg/L cap are all tunable. ### Sodium source dropdown — conversions | Format | Sodium (mg) → grams | |---|---| | Table salt (NaCl, 39.3% Na) | grams = mg ÷ 393 | | Sodium citrate (23.4% Na) | grams = mg ÷ 234 | **Output format:** `Total sodium = 1,000 mg → 2.54 g table salt` _(or)_ `→ 4.27 g sodium citrate` > Note: dropped "dihydrate" label from the UI; ÷234 still corresponds to the common food-grade form. (Anhydrous form, 26.7% Na / ÷267, was removed as an option.) --- ## 5. Outputs 1. **Total carbs to prep (g)** + ingredient breakdown per chosen carb format (+ warning if rate exceeds chosen format's ceiling). 2. **Sodium to add to drinks (g)** of table salt **or** sodium citrate — recommended *intake* (partial replacement), mixed across your planned fluid. 3. **Recommended fluid intake** (mL/h + total L, drink-to-thirst), with estimated sweat loss and projected deficit shown as context. ### Phone layout (≤ 1000 px wide) - Single column, **top-to-bottom flow**: You & your fuel → Your paces → Plan from workout → Session → Conditions → **Result** (result at the bottom). - **Sticky result bar** at the bottom (carbs g · salt/citrate g); tap → scroll to the full result. Hides while the result cards are on screen. - All controls ≥ 16 px text (no iOS focus-zoom), ≥ 46 px tall; bigger hit area on the gut/salty toggles; safe-area insets respected; can be added to the home screen. - **Remembered on the device** (browser localStorage, per-device convenience only): body weight, carb/sodium source, paces, gut-trained, salty-sweater, time/distance mode. Session, duration and conditions are not remembered. --- ## 6. Workout-based session entry (optional) Instead of manually choosing session type + duration, paste a planned workout and the tool **derives both** and auto-fills them. Fully client-side — no LLM, no backend. ### Input mode toggle: time-based or distance-based The toggle sets the reading of the ambiguous `m` unit (see grammar) and the input placeholder; classification and all downstream carb/sodium logic are identical in both modes. Either mode accepts any mix of time and distance segments — a distance segment is always timed from the paces panel. - **Time-based** — segments carry a duration: `15' easy, 5x3' hard w/2' jog, 10' cd`. Duration is read directly; no paces needed. - **Distance-based** — segments carry a distance: `14km easy, 3km HMP`. Duration = distance × the pace for that segment's intensity, drawn from the **My paces** panel (or an inline `@m:ss`). ### Paste syntax (grammar) - Segments separated by `,`, a new line, `+`, or the word `into`. - **Time:** `15'`, `15min`, `15m`, `90s`, `1:30` (mm:ss). **`m` is ambiguous** (minutes vs metres): in **time mode** `m` is **always minutes** (`15m` = 15 min); in **distance mode** `m` is **always metres** (`400m`). One rule per mode, no number thresholds. To give a distance inside a time-based workout, use `km` (e.g. `6x0.4km`). `'` and `min` are always minutes. - **Distance:** `14km`, `10k` (= 10 km), `400m`, `1.5km`, `1.5k`, `1mi`. A segment's own distance token is never read as an intensity word, so `5k easy` = 5 km easy; `5k`/`3k` elsewhere in the segment (`8k @5k pace`) still mean 5K/3K race pace. - **Reps:** `5x1km`, `8x90s`, `6x3'` — optional recovery `w/