Most runners think about hydration in fairly simple terms: drink when thirsty, take an electrolyte capsule or two, and hope it's enough. Buried in that approach is an assumption about sweat profiles that rarely gets questioned. Many athletes assume that if they learn their sweat profile from a couple of data points, they're in the clear, but it's often overlooked how drastically someone's sweat profile can change day to day based on a myriad of factors.
Sweat rate, electrolyte loss rate, and sweat composition aren't fixed numbers. They shift by workout, by environment, and by effort, sometimes in ways that don't match intuition at all. This is one athlete's training block in preparation for the Old Port Half Marathon in Portland, Maine, used to show how sweat profiles can change, and how the Nix Hydration Biosensor and Nix Solo app can help pinpoint hydration needs based on personalized data and forecasted conditions.
The Old Port Half Marathon, officially the Shipyard Old Port Half Marathon & 5K, is held each June in Portland, Maine. The single-loop course passes Casco Bay and Back Cove before finishing in the city's historic Old Port district. Because the race falls in early June, training for it means building through a Maine spring that can swing from raw and damp to unexpectedly warm and humid within the same week.
Using the Nix Hydration Biosensor, one runner logged 16 training runs between April and early June ahead of the race: 10 short runs and 6 long runs, each tracking fluid, electrolyte, and sodium losses throughout.
When reviewed individually, no single training run says much. When the workouts are reviewed together, a pattern starts to emerge based on the type of run being completed.
Across the shorter runs, sweat rate tracked closely with the Nix Index environmental score. The Nix Index revealed hotter, more humid conditions meant a noticeably higher fluid loss rate, even when the sessions were under an hour in duration. Across the long runs, the Nix Index barely moved the needle.
What predicted a higher sweat rate on long runs was simply time on feet. The longer the run, the higher the loss rate, regardless of how tough conditions were that day. This is a good example of a finding that's likely specific to this individual rather than universal. While the Nix Index had little bearing on sweat rate for this athlete's long runs, it could just as easily be the biggest driver in someone else's data. It's one more reminder of just how differently one athlete's sweat profile can look from another's.
On average, the long runs in this training block produced:
There's a seasonal wrinkle worth noting, too. The average Nix Index across these runs climbed steadily over the training block, from the low 50s in April to the low 60s by late May. That means a simple chart of “loss rates rising as training progressed” could easily be misread as an acclimation or fitness story, when it's really conditions warming up and long runs getting longer at the same time. Separating cause from coincidence is exactly why this kind of analysis needs real volume, not just a small handful of runs, to interpret the data meaningfully and correctly.
Sweat rate answers how much fluid a runner is losing. Sweat composition, how many milligrams of electrolytes are packed into each ounce of that sweat, answers a different question: how concentrated an electrolyte drink needs to be to actually match the composition of the athlete's sweat. In this case, the runner's electrolyte drink of choice was LMNT, which packs 1,000 mg sodium, 200 mg potassium, and 60 mg magnesium (1,260 mg of electrolytes total) into each stick, recommended to be mixed with 16–32 oz of water.
Across the same 16 runs, sweat composition ranged from 38 to 90 mg/oz, more than a two-fold spread, and unlike sweat rate, it did not track meaningfully with the Nix Index, temperature, or even run duration. The hottest short run in the block (Nix Index 70) produced one of the lowest sweat compositions recorded (43 mg/oz); the highest composition of the entire block (90 mg/oz) came on a mild, low-Nix-Index evening. Rate and composition were simply answering two different questions.
The split between short and long runs was notable too. Long runs stayed inside a fairly tight composition band (62–73 mg/oz), which translated into a fairly consistent electrolyte mix of about 17–20 oz of water per stick of LMNT. Short runs swung much more widely (38–90 mg/oz), calling for anywhere from 14 to 33 oz of water per stick depending on the day, variability that a single go-to ratio would have missed entirely.
The data shaped how the training block was run; the Nix Calculator tools shaped how it was fueled.
Race Day Calculator: used periodically throughout the training block to check in on the expected Nix Index conditions for the day of the race, using the last 10 years of historic weather data to predict the environmental conditions and expected sweat profile for the day.
Priming Calculator: used the night before every long run, to plan what and how much to drink in the two hours before going out for a long training run.
Custom Drink Calculator: used regularly to understand the exact product-to-water ratio needed to match the electrolyte composition of a given beverage to the athlete's own sweat composition.
Recovery Calculator: used after each long run to highlight the specific replenishment needs in the two-hour period after finishing a run.
Predictive Calculator: used to check in starting 10 days out from race day, to see what the expected sweat profile would look like heading into June 6 based on upcoming weather forecasts and historic Nix data.
Race-day conditions, recorded on June 6 in Portland, landed at a Nix Index of 68, within the range already logged in training, toward the tougher end of it.
Finalized Race Data:
The finish time came in at 1:43:02, a 7:52-per-mile pace, faster than any training pace logged over the previous ten weeks. That extra intensity showed up directly in the data: race-day sweat rate and sodium loss rate were both the highest recorded across the entire data set, training included, even though the race wasn't the longest single effort logged. Pace, not just heat or distance, moved the numbers.
The composition number stands out the most. At 82 mg/oz, race-day sweat composition worked out to an ideal LMNT ratio of about one stick per 15–16 oz of water, which lands almost exactly on LMNT's own standard 16 oz recommendation. The training-block average (63.8 mg/oz) had called for a noticeably milder mix, closer to one stick per 20 oz. Knowing that difference ahead of time, instead of discovering it mid-race, is exactly what the Predictive and Custom Drink Calculators are built for.
Had this athlete relied on training averages alone heading into race day, the expectation would have been to consume about 3,290 mg of electrolytes over the race's 1:43:02 duration. Actual race-day losses came out closer to 5,560 mg, a shortfall of roughly 2,270 mg, or 41%, if the plan had been based on averages instead of individual data. That gap could easily have been the difference between a personal record and simply finishing. Used as a training tool, the Nix Hydration Biosensor surfaced this kind of individual insight well before race day, arguably the difference maker in this race, something mileage alone couldn't solve.