Running Pace Calculator
Convert between pace, time, and distance. Predict marathon, half-marathon, 10K, and 5K finish times from a shorter race result.
What this calculates
Pace, distance, and time are the three numbers every runner juggles. Given any two, this calculator solves for the third. It also predicts race times across distances using the Riegel formula — useful for picking a realistic target pace for an upcoming race based on a recent shorter result.
Formula & how it works
Pace = time ÷ distance. Time = pace × distance. Race prediction uses Riegel: T2 = T1 × (D2 ÷ D1)^1.06, where T1 and D1 are your known time/distance and the exponent 1.06 captures how pace slows with distance. The formula is reasonably accurate up to ~3× distance scaling; it overpredicts beyond that for most runners.
Worked example
5K in 25:00 → pace = 25 ÷ 5 = 5:00/km. Predict marathon: T = 25 × (42.195/5)^1.06 ≈ 240.3 min = 4:00:18. Predict 10K: T = 25 × (10/5)^1.06 ≈ 51.7 min = 51:42. The marathon prediction assumes proper training for the distance — untrained runners will be slower.
Frequently asked questions
How accurate is the race predictor?
Within ~5 % for trained runners and modest distance jumps. The Riegel formula assumes you've trained appropriately for the longer race — predicting a marathon from a 5K without long-run training will overshoot reality.
What's a good 5K time?
Beginner: 30+ minutes. Recreational fit: 25 minutes. Trained recreational: 20 minutes. Competitive amateur: 18 minutes. Elite: under 14 minutes (men) / 15 (women). 'Good' is whatever's a personal record.
Should I race at predicted pace?
Use predicted pace as a ceiling, not a target. For an A-race, aim for 5–10 seconds per km slower than predicted to leave room for the final-kilometer kick. Going out at predicted pace often causes a blowup in the last quarter.
Why isn't pace × distance perfectly linear?
Because running pace decays at long distances — same effort can't be held forever. Glycogen depletion, thermal load, and form breakdown all conspire. Riegel's 1.06 exponent encodes this empirically.