Why 1.06?

If running were purely aerobic and fatigue-free, times would scale linearly with distance (exponent = 1.0). The exponent 1.06 captures the reality that longer races demand proportionally more effort: your pace will slow as the distance grows, not just linearly but slightly more.

A runner who finishes a 10K in 42:00 won't run a marathon at exactly 4× that pace — they'll be a bit slower due to energy depletion and cumulative fatigue. The 1.06 factor models this empirically.

Using Your Prediction in Training

Once you have a predicted marathon time, work backward to find your training paces. If your predicted marathon is 3:30, your target marathon pace is approximately 4:58 min/km (7:59 min/mi). Use that pace for your long run progression and marathon-specific workouts in the final 8 weeks before race day.

Predictions are most useful when updated after each key race. As your fitness improves through training, your 10K time will drop, and your marathon prediction will improve alongside it. Tracking this progression is a reliable way to monitor training effectiveness even without racing a full marathon every cycle.

⚠️ Limitations to Keep in Mind

The formula assumes both races are run at maximum effort and under similar conditions.

It doesn't account for terrain, weather, elevation, or training specificity.

Predictions are most accurate when the two distances are relatively close (e.g., 5K → 10K). Extrapolating from a 5K to a marathon carries more uncertainty.

Newer runners often outperform predictions; highly trained athletes tend to be closer to the model.

Race Prediction FAQ: Getting the Most From Your Results

The Riegel formula (1977) is the most widely used model for predicting running race times across distances. It captures a fundamental truth about endurance: as distance increases, pace slows at a predictable, non-linear rate. The formula is T2 = T1 × (D2/D1)^1.06, where the exponent 1.06 reflects the fact that longer races demand proportionally more effort — your pace will slow faster than linear as distance grows.

Accuracy is highest when predicting between adjacent distances on flat courses from a genuine all-out race effort. For a marathon from a 10K, expect results within 2–4 minutes under ideal conditions. Common sources of error include using a time from a training run rather than a race, mismatched course profiles (flat → hilly), or terrain and weather differences between the known and target event.

The formula does not account for training specificity. A runner whose entire recent training is 5K-focused may have a 10K time that implies a marathon pace their endurance base cannot sustain. Use the prediction as a realistic ceiling and plan a conservative start strategy. For large distance extrapolations (e.g. 5K to marathon), compare with the Cameron formula output as a cross-check.

How does this work?

How accurate is the Riegel formula?

Within 2–4 minutes for a marathon predicted from a 10K under ideal conditions (flat course, good weather, genuine maximal effort). Accuracy improves for adjacent distances (10K to half marathon) and decreases for large gaps (1 mile to marathon). Works best for paces between 3:30 and 5:30 per km. Elites and beginners tend to see larger errors.

Why does my predicted time seem too fast or too slow?

Most often the input was not a genuine race effort — a comfortable long run underestimates fitness, a PB on a fast course overestimates it. Course and weather differences between known and target events matter too. Training specificity is another factor: strong 10K speed does not guarantee the aerobic endurance base needed for a marathon.

Can I predict across very different distances, like 5K to marathon?

Technically yes, but reliability decreases as the gap widens because the energy system requirements shift significantly. For a 5K-to-marathon jump, also run the Cameron prediction — it applies a larger correction for the anaerobic component of short races and will give a more conservative, often more realistic, marathon estimate.

How the Race Predictor Works