Marathon Time Predictor
Predict your marathon finish time from a recent 5K, 10K, or half marathon result using empirical scaling factors tuned to typical recreational runners. Use it to set a realistic goal before your first marathon or to validate a training cycle.
Last updated: May 2026
Compare with similar
About this calculator
The predictor returns marathonTime = referenceTime * scaleFactor, where scaleFactor depends on the reference distance: ~9.3 from a 5K, ~4.66 from a 10K, ~2.1 from a half marathon. These factors are calibrations of the empirical Riegel formula (newTime = oldTime * (newDistance / oldDistance)^1.06), which models how endurance running times scale with distance for well-trained athletes. Variables: distance is the reference race distance (5, 10, or 21.1 km); time is the reference race time in minutes. The 1.06 exponent in Riegel's formula reflects that performance degrades faster than linearly with distance because fuel substrate shifts from glycogen to a mix of glycogen and fat, lactate buffering capacity matters more, and accumulated muscle damage compounds. Edge cases: predictions are most accurate from half marathon time (within 3-5% for trained marathoners) and least accurate from 5K time (often optimistic by 5-15% for runners who lack long-run training base). Predictions assume you have done marathon-specific training - long runs of 30+ km, a structured taper, and adequate fueling practice; without those, real marathon time often runs 10-20% slower than the formula predicts because the wall hits at 30-32 km from inadequate aerobic base. Age, gender, and training background do not enter the formula but matter in practice: older runners and women typically have flatter fatigue curves and run closer to or slightly faster than predicted; younger speed-trained runners often run slower than predicted because they lack endurance base. The formula also assumes flat, dry conditions on a measured course; hills, heat, altitude, and poor pacing can each add 5-10 minutes.
How to use
Example 1 - Recent half marathon time 1:45:00 (105 minutes). Compute: 105 * 2.1 = 220.5 minutes = 3:40:30 predicted marathon time. Verify against Riegel directly: 105 * (42.2 / 21.1)^1.06 = 105 * 2^1.06 = 105 * 2.085 = 219.0 minutes = 3:39:00 - the simplified 2.1 scaling adds about 90 seconds of conservatism. This is a realistic target for a runner with adequate long-run training (peak long runs of 32+ km and a structured 16-week build); without that base, expect to add 10-20 minutes. Example 2 - Recent 10K time of 45:00. Compute: 45 * 4.66 = 209.7 minutes = 3:29:42 predicted marathon. Verify via Riegel: 45 * (42.2 / 10)^1.06 = 45 * 4.553 = 204.9 min = 3:24:54 - the simplified 4.66 factor again adds a small conservatism buffer. For most runners without marathon-specific endurance, this prediction is optimistic; add 15-25 minutes if your longest long run is under 28 km, and use the half-marathon prediction as a cross-check before committing to a goal pace.
Frequently asked questions
How accurate is the marathon prediction from my half marathon time vs my 5K time?
Half marathon predictions are typically accurate within 3-5% for runners who have done at least 12 weeks of marathon-specific training, because the half marathon distance (21.1 km) shares aerobic demands with the marathon and reveals fueling and fatigue patterns that shorter races hide. 10K predictions are accurate within 5-10% for trained marathoners but tend to be optimistic by 10-15% for runners with a strong 10K base but limited long-run history. 5K predictions are the least reliable, often optimistic by 10-20%, because 5K performance reflects VO2 max and lactate threshold but says little about your ability to maintain pace for 4+ hours. The reliable rule: take your most recent half marathon time, multiply by 2.1, then add 5-10 minutes if your longest long run is under 32 km, or subtract 2-3 minutes if you have a strong endurance base and have run marathons before.
What is the Riegel formula and why is it the basis of most marathon predictors?
The Riegel formula, published by Pete Riegel in 1981, predicts a race time from a previous race time using newTime = oldTime * (newDistance / oldDistance)^1.06. The 1.06 exponent was derived empirically from racing performance data across distances from 800 m to ultramarathons and represents the average rate at which running performance degrades with distance for trained athletes. It is the basis of most marathon predictors (and many race calculators in general) because it is simple, has held up well against 40+ years of additional data, and only requires one reference race rather than physiological inputs like VO2 max or lactate threshold. The exponent varies slightly by ability level - elite athletes have an effective exponent closer to 1.04 (less fatigue), while less-trained runners have exponents of 1.08-1.10 (more fatigue). Some modern tools use Cameron's formula or Daniels' VDOT tables, which add small refinements but agree with Riegel within 1-2% for typical recreational runners.
Why does my actual marathon time often come in slower than the predicted time?
The most common reason is inadequate long-run training - the formula assumes you have done multiple long runs of 30+ km, but many first-time marathoners cap their long runs at 25-28 km and then hit the wall around 30-32 km on race day, losing 10-20 minutes in the final stretch. Fueling failures also commonly cost 5-15 minutes; the formula assumes you have practiced taking in 30-60 g of carbohydrate per hour during long runs, but many runners discover their GI system cannot tolerate that intake mid-race. Race-day conditions matter too - temperatures above 15 C, headwinds, hilly courses, and crowded starts each can add 3-8 minutes vs the formula's flat-and-cool assumption. Pacing mistakes (going out too fast in the first 10 km) commonly cost another 5-15 minutes via accelerated glycogen depletion. The cleanest fix is to set your race goal at the formula's prediction plus 5-10 minutes of buffer, then run conservatively for the first 25 km and accelerate only if you feel strong - the asymmetric cost of going out too fast vs too slow makes conservatism the right default.
What are common mistakes when using a marathon time predictor?
The most common mistake is using a fast 5K or 10K time to set a marathon goal without considering your long-run training base - speed-trained runners with weak endurance routinely set goals 10-15 minutes too fast and then collapse in the final 10 km. Another frequent error is using an old reference race time (more than 8-12 weeks old) when your fitness has shifted; predictions assume the reference time reflects current condition. People also commonly forget that the formula assumes ideal race conditions - flat, cool (5-15 C), low wind, and a familiar course - and apply it to a hilly destination marathon on a hot day, then are surprised by the result. Using a track time-trial or untrained race effort instead of a true all-out race produces predictions that look great but are physiologically unsupported. Finally, ignoring the formula's accuracy band (3-5% from a half marathon, 10%+ from a 5K) and treating the predicted time as a precise goal sets up a failure pattern; treat it as the center of a range and plan to nail the lower end through pacing rather than chasing the precise number.
When should I NOT rely on a marathon time predictor?
Skip the predictor entirely for your first marathon - novice runners do not have a stable enough physiological baseline for the formula to predict reliably, and the best target is simply to finish comfortably rather than hit a time goal. Do not use it after a major training disruption (illness, injury, three or more weeks of reduced volume) because the reference race no longer reflects current fitness. The formula is unreliable for trail marathons, hilly marathons (1,000+ m of climb), or hot-weather events (regularly above 20 C) - all of which can add 10-30% to flat-course-equivalent times that the formula simply does not model. Older runners (60+) and very young runners (under 18) often deviate from the formula by 5-10% in either direction, so use individual race history rather than the standard scaling. Finally, for ultramarathons (50 km+), the Riegel exponent of 1.06 stops applying - ultras have entirely different physiology (walking strategy, nutrition, sleep deprivation) and need specialized prediction tools.