Marathon Time Predictor With Weekly Mileage
Use a recent race result and your typical weekly training distance to estimate a marathon time with the Vickers–Vertosick model. Add a second result at a different distance to include your own rate of slowing as races get longer.
Predicted marathon time3 hr 54 min 10 sec
- Predicted pace per mile
- 8:56 per mile
- Predicted pace per kilometer
- 5:33 per kilometer
- Standard Riegel prediction
- 3 hr 40 min 26 sec
- Mileage-adjustment difference
- 13 min 44 sec slower than standard Riegel
- Observed fatigue exponent
- Unavailable in one-race mode
- Model used
- One prior race (Vickers–Vertosick model)
The one-race Vickers–Vertosick model combines a Riegel-based marathon velocity with typical weekly mileage.
How to use this calculator
- Choose one prior race or two prior races at different distances.
- Choose miles or kilometers, then enter your typical weekly marathon-training distance rather than your highest week.
- Select the recent non-marathon race distance and enter each finish-time component.
- Use the predicted time and two pace outputs for planning, then compare the result with the standard Riegel projection.
How this marathon prediction uses weekly mileage
A conventional race predictor extends one result with the Riegel equation. That can be useful for nearby distances, but the original Vickers–Vertosick study found that it was poorly calibrated for recreational marathon prediction: it often projected a faster marathon than runners actually completed. Their models add typical weekly training distance, which is a practical proxy for the volume supporting marathon-specific endurance.
For one prior race, the calculator first projects a Riegel marathon velocity with vR = M / [t1 × (M/d1)^1.07]. It then calculates marathon velocity as vM = 0.16018617 + 0.83076202vR + 0.06423826(W/10), and converts it to time with TM = M / (60vM). Here, M is 42,195 meters, d1 is the prior-race distance in meters, t1 is its time in seconds, and W is weekly distance in miles.
What changes with two races
With two results, the calculator orders them from short to long and estimates your observed fatigue exponent: kObserved = ln(tLong/tShort) / ln(dLong/dShort). It uses that with mileage to estimate the marathon exponent: kM = 1.4510756 − 0.23797948kObserved − 0.01410023(W/10). The marathon time is then TM = [tLong × (M/dLong)^kM] / 60.
The Riegel comparison uses TR = [tLong × (M/dLong)^1.07] / 60 from the longest race entered. The difference is not a verdict that one forecast is right. It makes visible how the mileage-adjusted model differs from a fixed 1.07 exponent. A lower weekly mileage generally makes the Vickers–Vertosick projection slower; two comparable races can also move it according to your observed change in pace with distance.
Choosing useful input races
A recent half marathon is usually the strongest single input here because it is closer to the marathon than a 10K. Two well-raced results at different distances can be more informative because they reveal a personal fatigue pattern, but only if the efforts and conditions were broadly comparable. Enter a typical training week, not a brief peak or a week disrupted by illness.
What the prediction leaves out
This is a population-based estimate, not a tailored race plan. It does not model elevation, weather, altitude, surface, crowding, injury, illness, fueling, pacing errors, taper quality, long-run preparation, or the time since the input races. Treat the pace as a planning reference to test in training, not as a target that overrides symptoms, safety conditions, or qualified professional advice.
Worked example
Suppose you usually run 30 miles per week and recently finished a half marathon in 1 hr 45 min. The standard Riegel calculation is 6300 × (42195 / 21097.5)^1.07, or 13,226 seconds. That gives a Riegel marathon velocity of 3.1902 m/s.
The one-race model gives 0.16018617 + 0.83076202 × 3.1902 + 0.06423826 × 3, or 3.0032 m/s. Dividing 42,195 by that velocity gives 14,050 seconds: about 3 hr 54 min 10 sec, with a pace of about 8:56 per mile or 5:33 per kilometer. The standard Riegel result is about 3 hr 40 min 26 sec, so the mileage-adjusted result is roughly 13 min 44 sec slower.
Common questions
Why does weekly mileage change a marathon prediction?
Marathon performance depends on endurance preparation as well as shorter-race speed. The Vickers–Vertosick models use typical weekly mileage because it improved marathon estimates in their recreational-runner data. Mileage is not a complete training description, so the result remains an estimate rather than a prescription.
Is a half marathon or 10K better for predicting a marathon?
For a one-race prediction, use your longest recent non-marathon race, which is usually a half marathon. A half marathon asks more of the endurance that matters for a marathon and requires less extrapolation. A strong 10K can still be useful, especially when paired with a second result at another distance.
Why is this result slower than a standard Riegel calculator?
Riegel applies a fixed 1.07 fatigue exponent to every runner. The Vickers–Vertosick approach adjusts the projection for typical weekly mileage and, with two races, the way your own pace changed across distance. In validation data from the study, ordinary Riegel marathon predictions were commonly too fast for recreational runners.
How recent should the input races be?
Use races that reflect your current fitness and were run hard under reasonably comparable conditions. A result from before an injury, major training change, pregnancy, long layoff, or a very different climate may be a poor input. When two races disagree because conditions differed, the fatigue exponent can be misleading.
What does the personal fatigue exponent mean?
In two-race mode, it describes how sharply your time increased as distance increased between the entered races. It is not a fixed physiological trait and can move with course profile, weather, pacing, fitness, and measurement error. An unusually low or high result is a reason to check the inputs rather than a diagnosis of endurance ability.
Can this predict a first marathon accurately?
It can give a useful starting range when you have recent races and a representative weekly distance, but a first marathon has extra uncertainty. Long-run experience, fueling, pacing discipline, and how well your training transfers to 42.195 km can matter more than the formula captures. Plan conservatively and revise with training evidence.