Calculate running pace per km. What time for marathon?
Guide to Running Training
Running pace and training zones
Running pace can be divided into zones based on intensity. Zone 1 (recovery) is pace below 130% lactate threshold, Zone 2 (endurance) is 130-150% threshold, Zone 3 (tempo) is 150-175%, and Zone 4-5 (VO2 max and anaerobic) is above 175%. Running in lower zones builds endurance base, higher zones develop speed and power.
Race time prediction
- Riegel method - T2 = T1 × (D2/D1)^1.06
- Rule - double distance = about 2.2x time
- Example - 5km in 20min means ~3:52 marathon
- Note - predictions are approximate, depend on conditions
Weekly training structure
A typical week for a recreational runner should include 3-4 training days. One long day (varied), one interval day (e.g., 6x800m at 5km pace), one technical or tempo day (e.g., 3x1600m), and 2-3 active recovery days (walks, light mobility). Gradually increase volume by 10% weekly.
Thresholds and zones in practice
Lactate threshold (LT) is the intensity where lactate begins to accumulate. You can estimate it as half marathon or 10km pace. Heart rate threshold is about 85% of max heart rate. In training, the talk test is important - if you can't speak a sentence, you're going too fast. This is a good indicator for lower zones.
A 1:45 half predicts a 3:39 marathon, not 3:30
Doubling the distance does not double the time. Endurance decays as distance grows, and the standard prediction raises the ratio to the power 1.06 rather than 1.0. That small exponent is the difference between the marathon most runners plan and the marathon they actually run — about nine minutes, all of it arriving in the last hour.
How it works
- Converts between pace, speed, distance and finish time.
- Predicts a time at one distance from a known time at another, using the Riegel relation.
- Shows the per-kilometre pace each prediction implies, since that is the number you actually run to.
pace = time ÷ distance Riegel prediction: T₂ = T₁ × (D₂ ÷ D₁)^1.06 the exponent is above 1, so every doubling costs more than double at 1.0 it would be pure proportion — the 0.06 is the fade
Worked example
A 1:45:00 half marathon, projected forward, and two other common conversions.
- half 1:45:00 → marathon 3:38:55
- naive doubling would say 3:30:00
- the penalty is 8.9 minutes
- 5 km 25:00 → 10 km 52:07
- 10 km 50:00 → half 1:50:19, marathon 3:50:01
Half-marathon pace is 4:59 per kilometre; the predicted marathon pace is 5:11. Thirteen seconds per kilometre sounds trivial and is exactly what separates a race that holds together from one that does not.
Reading the result
- The prediction assumes you have trained for the distance. Riegel describes how a given engine fades over a longer effort, not how someone who has never run beyond 15 km will cope with 42 — for them the real result is far worse than the formula says.
- The exponent drifts with training. Well-trained endurance runners sit nearer 1.05, and runners with a short-distance background nearer 1.08, so the prediction is a centre of a range rather than a promise.
- Even splits beat going out fast, and the arithmetic is unforgiving. Banking four minutes in the first half typically costs eight to twelve in the second — a 3:38:55 attempt run that way finishes around 3:42:55, so the four minutes saved buys four minutes lost.
- Heat, hills and wind are absent from the formula entirely. All three cost more over longer distances, so a prediction made from a fast flat half will overstate what a hot hilly marathon returns.
Common questions
- Why can't I just double my half time?
- Because the exponent is 1.06, not 1. Doubling gives 3:30:00 for a 1:45:00 half, while the prediction is 3:38:55. The nine-minute gap is the physiological fade that the doubling assumption pretends does not exist.
- How accurate is this in practice?
- Good between neighbouring distances — 5 km to 10 km, 10 km to half — and progressively less reliable as the jump widens. Predicting a marathon from a 5 km stretches the relation across a factor of eight and should be treated as a rough ceiling, not a target.