Five calculators. No account, no catch.
These are the numbers most training plans are built on — threshold, zones, pace, aerobic ceiling, load. Every one of them runs on figures you type in yourself, which means they work the same whether your watch says Garmin, Coros, Polar, Suunto, Apple, or nothing at all.
Where our own research says a standard formula is wrong, we show you both answers rather than picking one quietly. Every formula names its source, and where a number is an approximation we say so instead of printing it to three decimal places.
- FreeAll five, all of them, forever.
- No signupNo account, no email, no wall.
- Nothing storedThe math runs in your browser.
- Any watchYou type the numbers. Brand is irrelevant.
Start here
What could I run for a marathon?
One honest recent race, converted into predicted times for every distance. You get two answers: the textbook formula everyone else uses, and the steeper fade we measured across the cohort. At the marathon they differ by about half an hour.
Open the race pace predictor- Textbook fade (Riegel, 1977)
- exponent 1.06
- Fade we measured (22,422 prediction-days)
- exponent 1.131
- Difference at the marathon
- about half an hour
The calculator shows both curves and lets you decide which to trust.
Pick the one you came for
Or compare them
What each one gives you, what you need to hand before you start, and the method underneath it.
| Calculator | What it gives you | You need | Method |
|---|---|---|---|
| Race pace predictor | One honest recent race, converted into predicted times for every distance. You get two answers: the textbook formula everyone else uses, and the steeper fade we measured across the cohort. At the marathon they differ by about half an hour. | One recent race distance and finish time | Riegel 1.06, against the 1.131 our data implies |
| VO2 max estimator | A field test instead of a lab. Cooper, Rockport or a 1.5-mile run, reported as a range rather than a false-precision figure, and placed against the cohort in our own drift study. | Age, sex, and one field-test result | Cooper 1968 · Kline 1987 · ACSM, each with its error band |
| FTP estimator | A hard effort turned into Functional Threshold Power, given as a range because the conversion is an approximation. Tell it your training level and the range narrows to the coefficient measured for riders like you. | Average watts from one hard effort | 0.88–0.96 of 20-min, by training level (Sitko 2023) |
| Heart rate zone calculator | Five zones by Karvonen, % of max, or MAF 180 — side by side, so you can see how much the method changes the answer. | Age, and resting HR if you know it | Karvonen · % Max · MAF 180 (Tanaka max HR) |
| Training load calculator | Fitness, fatigue and form from four weeks of training stress — running the real daily recursion rather than the shortcut most simple calculators use, which overstates form swings roughly twofold. | TSS for your recent workouts | 42-day and 7-day exponential averages |
What a formula can and cannot tell you
Every calculator here is a published equation applied to numbers you supply. Riegel was fitted to race results from 1977. Karvonen assumes a max heart rate that an age formula guessed at. None of them know how you slept, how hot it was, or that you were chasing someone up the last hill.
That is not a reason to skip them — a well-chosen estimate beats a vibe. It is a reason to treat the output as a starting range and let a few weeks of your own data narrow it. We write about where these formulas hold up and where they fall apart.
Read next
Where the numbers come from, how far you can trust them, and what the training-technology industry does with them once they leave your wrist.