Online Tool Store
Marathon Time Predictor
Predict race times from one recent result using both Riegel and Cameron, with the gap between them shown as the honest measure of uncertainty.
🔒 This tool runs entirely in your browser. Your files are never uploaded to a server.
1.06 is the classic value. Higher means you fade more over distance — realistic if your long runs are short.
Predicted marathon
—
| Distance | Riegel | Cameron | Pace | They differ by |
|---|
These formulas assume you have trained for the distance. They convert one performance into an equivalent one; they cannot know whether your longest run was 30 km or 10. A marathon predicted from a 5K is the time you could run if your endurance matched your speed, which for most people it does not — the further the extrapolation, the more the prediction becomes a statement about your training rather than your fitness. Not medical advice; if you are new to long-distance running, build up gradually and get checked out first.
How to use it
- Enter a race you have actually run, and its finish time.
- Read the whole table rather than one row — the spread matters.
- Raise the exponent if your long runs are shorter than the race you are predicting.
- Use the pace column to plan, not just the finish time.
Riegel, in one line
T₂ = T₁ × (D₂ / D₁)^1.06
If the exponent were exactly 1, you would hold the same pace at any distance and a marathon would be 8.4 times a 5K. The 1.06 is the empirical fact that people slow down — and because it is an exponent rather than a multiplier, the penalty compounds as the distance ratio grows. That is fine over a small stretch and increasingly speculative over a large one.
Where the formulas agree, and where they part
Riegel is a single power law; Cameron uses a curve fitted to race records. Comparing them on the same input is the cheapest possible uncertainty estimate:
from a 20:00 5K Riegel Cameron apart
10K 41:42 41:40 2s
half 1:32:00 1:31:51 9s
marathon 3:11:49 3:15:11 3:22
Two seconds apart at 10K and more than three minutes apart at the marathon, from identical input. Neither formula became less correct along the way — the distance ratio grew, and with it the amount each one is guessing. The table in the tool shows that gap on every row and highlights it once it passes two minutes.
The exponent is doing more work than the formula
marathon predicted from a 20:00 5K
exponent 1.05 3:07:46
exponent 1.06 3:11:49 ← the standard value
exponent 1.08 3:20:11
exponent 1.10 3:28:54
exponent 1.15 3:52:25
Four hundredths on the exponent is seventeen minutes on the finish time — 8.9%. Most calculators hide this behind a fixed 1.06 and present the result to the second, which implies a precision that simply is not there. Exposing it is not a cop-out: the right exponent genuinely depends on you, and specifically on how far your long runs go. Someone racing a marathon off 15 km long runs will fade more than the average runner in Riegel's dataset, and a higher exponent says so.
Why the marathon breaks the pattern
Every distance up to about the half is limited by the same things: how much oxygen you can use and how fast you can run near your threshold. Those transfer, which is why the formulas work well across that range. The marathon adds constraints no shorter race tests at all — stored carbohydrate running out, the accumulated mechanical damage of hours of footstrikes, fuelling and hydration, and a far greater sensitivity to heat.
None of that is visible in a 5K time, so a marathon prediction from a 5K is really a statement about your speed plus an assumption about your endurance. Predicting from a half marathon is much safer — the extrapolation is only twice the distance, and the old rule of thumb of doubling your half and adding ten minutes lands at 3:40:00 from a 1:45:00 half, inside the range the formulas give.
FAQ
How accurate is a marathon prediction from a 5K?
Treat it as a ceiling rather than a forecast. The formulas convert one performance into an equivalent one, assuming your endurance matches your speed — which is exactly the assumption a 5K cannot test. From a 20:00 5K, Riegel predicts 3:11:49 at the standard exponent and 3:28:54 at 1.10. That 17-minute spread is not the formula being vague; it is the honest uncertainty of extrapolating eight times the distance.
What is the Riegel exponent and should I change it?
It controls how much you are assumed to slow as distance grows: time scales with distance raised to that power. 1.06 came from fitting race records and suits a well-trained runner over nearby distances. If your longest run is well short of race distance, a higher exponent — 1.08 to 1.10 — is more honest, because you will fade more than the average of the runners the formula was fitted to.
Why show two formulas?
Because their agreement is the most useful accuracy signal available. Predicting a 10K from a 5K, Riegel and Cameron land within two seconds of each other. Predicting a marathon from that same 5K, they separate by several minutes. When they agree you can trust the number; when they diverge, that gap is roughly the size of your uncertainty.
Is the "double your half and add ten minutes" rule any good?
Better than it sounds. From a 1:45:00 half it gives 3:40:00, against Riegel's 3:38:55 and Cameron's 3:43:08 — comfortably inside the spread of the formulas. It works because the half is close enough to marathon distance that the extrapolation is short. It fails badly applied to a 5K.
Why does my actual marathon come in slower than every prediction?
Usually because the marathon is limited by things a shorter race never tests: stored carbohydrate, the mechanical wear of three or four hours on your feet, fuelling, and heat. These are training and race-day problems rather than fitness problems, and no formula built on race times can see them. The prediction tells you what your speed is worth; whether you can hold it for 42 km is a separate question.
Does it account for hills, heat or altitude?
No. It compares distances, nothing else. A hilly course, a hot day or thin air will all cost you time that these formulas know nothing about, so a prediction from a flat, cool race will flatter a hard one.
How we compare
| Feature | Online Tool Store | Other race predictors | A coach or training platform |
|---|---|---|---|
| Two formulas, with the gap shown | ✓ | One, to the second | ✓ |
| Exponent adjustable and explained | ✓ | ✗ | Handled differently |
| Warns when the extrapolation is a stretch | ✓ | ✗ | ✓ |
| Pace per km or per mile for every distance | ✓ | ✓ | ✓ |
| No account, nothing recorded | ✓ | Varies | ✗ |
| Knows your actual training volume | ✗ | ✗ | ✓ |
| Accounts for course, heat and fuelling | ✗ | ✗ | ✓ |
Honest about its own uncertainty, which is unusual in this category — two formulas, an adjustable exponent, and a warning when you are asking too much of the maths. It knows nothing about your training, the course or the weather, and those are exactly what decide a marathon, so a coach still beats a calculator.