One result, three visible equivalents
A race-equivalence formula answers a narrow question: if the endurance relationship encoded by the model continues to 21.0975 km, what time is equivalent to the performance you entered? It does not know whether your current training, race course or execution will support that equivalence.
| METHOD | WHAT IT CONTRIBUTES | WHAT IT CANNOT SEE |
|---|---|---|
| Riegel 1.06 | A transparent, reproducible power-law conversion | Your individual endurance curve or training support |
| Daniels VDOT | An equivalent performance at the same estimated aerobic score | Whether you have trained specifically for the target duration |
| Cameron | A distance-varying correction instead of one fixed exponent | Course, weather, race execution or recent training continuity |
| Optional two-race curve | The endurance exponent implied by two of your own results | Whether both races were equally fresh, all-out and comparable |
Riegel uses T₂ = T₁ × (D₂/D₁)1.06. Daniels combines a running oxygen-cost curve with the fraction sustainable for a given duration, while Cameron changes its correction with distance. You can compare Daniels-style equivalents with the official VDOT calculator and inspect a published overview of the formulas in this race-predictor reference.
Why a half marathon needs different logic from a marathon
The half marathon is not simply a marathon chart cut off at 21.1 km. In a study of 7,258 half-marathon finishers and 1,853 marathon finishers, both groups slowed on average, but half-marathon pacing was more even. The authors reported average speed change of 4.10% in the half versus 5.71% in the marathon and recommended less variable half-marathon pacing. Read the pacing study.
That is why this tool starts with even effort. The pacing chart does not automatically choose a fade based on your confidence score. Instead, you can deliberately compare an even race, a controlled start, a gentle positive split and a fast opening. These are consequence models for execution choices, not claims about the kilometre where fatigue must begin.
What the training adjustment can honestly use
Recent race performance carries the speed signal. Weekly distance, longest run, consistency and recency supply evidence about whether that signal can extend to a half marathon. A cohort of 556 half-marathon runners grouped weekly volume into below 20, 20–32 and above 32 km, and longest runs into below 15, 15–21 and above 21 km. Higher volume and longer runs were associated with faster finishes and less late-race decline. Read the training study.
The study supports the direction of those signals; it does not supply a universal penalty for a runner on the other side of a cutoff. Rayvik therefore keeps every named model intact and shows the exact percentage of its separate planning buffer. The range becomes wider and more cautious when evidence is thin, but it is never presented as a calibrated confidence interval.
- Weekly volume. A broader aerobic base is relevant to both finish time and the ability to keep pace stable.
- Longest run. A longer single run supplies distance-specific evidence, but the calculator does not prescribe a mandatory 21 km training run.
- Training continuity. A recent PB can survive a disrupted block better than half-marathon readiness does.
- Result recency. An older race leaves more room for current fitness to have moved in either direction.
Predicting a half marathon from 5K versus 10K
A 5K is a strong speed test, but the calculator has to infer more endurance on the way to 21.1K. A 10K demonstrates more sustained performance and receives more confidence. A recent 10-mile race is closer still.
If you enter two distances, the calculator solves for the exponent that connects your own results. A flatter curve suggests stronger observed endurance; a steeper one suggests more drop-off as distance increases. Treat it as a diagnostic signal. Two races on different courses, in different weather or at different effort levels can produce a misleading personal curve.
How to choose a race target from the range
- Use the right source. Prefer a recent all-out 10K or 10-mile race when available. Training runs usually contain unknown effort and recovery.
- Respect the slower end. Formula disagreement is uncertainty, not an invitation to pick the fastest answer.
- Plan an even race first. Use the other pacing profiles to understand execution risk, not to normalize a late fade before the race begins.
- Adjust for race day. Heat, hills, wind, crowding, taper and health can outweigh the small difference between two formulas.
Frequently asked questions
How can I predict my half-marathon time from a 10K?
Enter a recent all-out 10K result. The calculator applies Riegel, Daniels VDOT and Cameron independently, keeps their equivalent times visible, then uses your weekly volume, longest run, consistency and result recency to create a cautious planning range. A 10K is close enough to be useful, but it still cannot prove 21.1 km readiness by itself.
Is a 5K or 10K better for predicting a half marathon?
A recent 10K is generally the stronger single input because it demonstrates more speed endurance and leaves a smaller distance gap. A 5K still helps, but the calculator gives it less confidence. If both races were recent, all-out and run in comparable conditions, the optional two-race curve can show whether your observed endurance agrees with the standard models.
Why do half-marathon predictors give different times?
Riegel uses a fixed endurance exponent, Daniels VDOT maps equivalent performance through oxygen-cost and sustainable-duration equations, and Cameron changes its correction with distance. The disagreement is useful: it shows the uncertainty hidden when a calculator returns only one number.
Should a half-marathon prediction assume a late-race fade?
No. A half marathon has a more even pacing profile than a marathon, and a well-chosen target should not require a built-in collapse. This calculator uses even effort as the baseline. Its other pacing profiles are manual stress tests for a controlled start, gradual drift or an over-fast opening—not forecasts of when you must slow down.
How do weekly mileage and the longest run affect the result?
They change confidence and a modest planning buffer, not the named formula results. A cohort study grouped half-marathon runners by less than 20, 20–32 and more than 32 km per week, and by longest runs below 15, 15–21 and above 21 km. Higher volume and longer runs were associated with faster times and less pace decline, but those associations do not justify a universal number of minutes for every runner.
What does the optional two-race curve do?
It calculates the endurance exponent implied by two performances at different distances. It is useful context when both races were recent, all-out and comparable. Because weather, course and race execution can distort either result, the personal curve does not replace the three named models or control the planning range.
Is the predicted range a guarantee or probability interval?
No. The formula results are fitness equivalents, and the training-supported range adds a transparent planning heuristic. It is not a statistically calibrated probability interval. Weather, course, health, taper and pacing can still move the finish time.