Race time predictor

How the prediction works

The core is the Riegel formula, the standard in race prediction for forty years: time at the new distance equals time at the known distance multiplied by the ratio of distances raised to a fatigue exponent. Riegel’s exponent was 1.06, and it holds well between neighbouring distances. The known weakness is the marathon. Vickers and Vertosick’s study of thousands of recreational runners found the standard formula overpredicts marathon performance, and that the error tracks training volume: runners on low weekly mileage fade far more than 1.06 allows, while high-mileage runners fade less. This calculator moves the exponent with your volume — 1.05 above 70km a week, 1.06 in the middle, 1.08 below 40km — so a fast parkrun stops promising a marathon the training has not built.

The second half is the weather. Every prediction assumes conditions, and most silently assume cool ones. Add the forecast temperature and humidity and the calculator derives the dew point and applies the temperature-plus-dew-point cost curve — the same one in the heat-adjusted pace calculator: no meaningful cost below a combined 38°C, about 3% at 24°C with an 18°C dew point, 8–10% or more in oppressive air. How to pace a race in the heat covers what to do with the adjusted number on the day.

The reference chart

Equivalent performances at the standard exponent (1.06), for a runner in the 40–70km-a-week band on a cool day. Read across a row to see what one race says about the others.

5K	10K	Half marathon	Marathon

Worked example: a 48:30 10K predicts a 3:43 marathon at the standard exponent. For a runner under 40km a week the same 10K predicts about 3:50 — and on a 24°C morning with an 18°C dew point, about 3:56. Three different honest targets from one race, depending on the training behind it and the air on the day.

What the number is for

A prediction is a planning tool, not a promise. Its best use is setting the opening pace: start the race at the predicted pace for the day you actually got, not the cool-day pace, and treat anything faster in the final third as a bonus. Its second use is honesty about the gap — if the marathon prediction from your 10K is a time you have never come close to holding on a long run, the exponent is telling you something about the training, and the training is what closes the gap. For the deeper story on why watch-based predictions run optimistic, read how accurate race predictors really are.

Frequently asked questions

How accurate is a race time predictor? Riegel-style predictors are reliable between adjacent distances — a 10K from a 5K, a half from a 10K — and become optimistic the further you extrapolate. A large study of recreational runners found the standard formula overpredicts marathon performance, especially for lower-mileage runners, which is why this calculator adjusts the exponent for weekly volume.

How does weekly mileage change the prediction? Endurance beyond your training is what the formula cannot see. Runners on high weekly volume slow down less as distance increases, so the calculator uses a gentler fatigue exponent for them and a steeper one for runners under about 40km a week, whose short-race speed converts less well to the marathon.

Why does the predictor ask about the weather? A prediction is only as good as the conditions it assumes. Adding the forecast temperature and dew point prices the race for the air you will actually run in — about 3% slower at 24°C with an 18°C dew point, 8–10% or more in oppressive conditions — using the same model as the heat-adjusted pace calculator.

Can I use a training run instead of a race? Yes, if it was an honest maximal effort at a known distance. A hard parkrun or a time trial works. An easy long run does not — the formula assumes the input was a race effort.

Fuelling the race too? The carb loading planner and marathon fuelling guide cover the other half.