Race Predictor Marathon Time: Combine 5K, 10K & Half Marathon Results
Compare marathon scenarios from recent 5K, 10K and half-marathon results. This tool converts each supplied result and combines the estimates using fixed weights. The range shows selected model assumptions; combining results has not been validated as a more accurate forecast.
Combined marathon race predictor
The calculator uses a Riegel-style power-law conversion for each result, with fixed weights of 1 for a 5K, 2 for a 10K and 5 for a half marathon. Review whether each race represents your current performance before including it.
Enter your race results
Leave out old or unrepresentative races. More inputs do not guarantee a better estimate.
Related prediction pages
If you want to understand how each result behaves on its own, use these dedicated pages:
- Marathon pace from 5K time
- Marathon pace from 10K time
- Marathon pace from half marathon time
- VDOT marathon pace explained
What combining results does
Different recent races can reveal inconsistent inputs or different performance patterns. Combining them here is a transparent convenience, not evidence that the result is more accurate than a single-race model.
On a small screen, swipe the table to see all columns.
| Input | Faster-scenario exponent | Slower-scenario exponent | Weight |
|---|---|---|---|
| 5K | 1.06 | 1.08 | 1 |
| 10K | 1.05 | 1.07 | 2 |
| Half marathon | 1.055 | 1.08 | 5 |
Each supplied result is converted to a marathon time using the power-law formula. The weighted mean of those times is calculated separately for each endpoint. The centre is the arithmetic mean of the two endpoints. Missing inputs are omitted, and remaining weights are renormalised.
Read the result as scenarios
The displayed interval is a span between two sets of assumptions. It is not a confidence interval, and the centre is not a measured “most likely” result. The fixed weights have not been fitted to your history or validated as improving this tool’s accuracy.
Agreement among inputs can be useful context, but shared model assumptions can make several estimates wrong in the same direction. Investigate conflicting results before averaging them away.
Input quality matters more than filling every field
Prefer representative recent races with known distances. Note weather, elevation, pacing and whether the performance reflects current fitness. Do not include a result merely to fill every input.
This calculator gives a half marathon weight five and a 10K weight two by design. It cannot recognise that a particular half marathon was old, interrupted or unrepresentative. That judgment remains with you.
What to do after you get your result
- Compare the scenarios with your training and representative recent races before choosing a target.
- Convert it into exact pace per kilometre with Marathon Pace Calculator.
- Review your key checkpoints in Marathon Splits Explained.
- Check whether your pacing plan is fragile with The cost of starting too fast.
- Build your next block with Monthly Training Plan.
FAQ
Does combining races guarantee greater accuracy?
No. This tool uses fixed illustrative weights, not an individually fitted prediction model.
Is the displayed range a confidence interval?
No. It is the span between two selected model scenarios. Actual performance may lie outside it.
Should I include all my races?
Use only representative inputs. Older results or races in very different conditions may make the comparison less useful.
References
- Riegel PS. Endurance power-law background: ARHE PDF.
- Review of performance prediction models: Predictive Performance Models in Long-Distance Runners.
- Large dataset prediction example: PLOS ONE performance prediction.
- Training and race history prediction approach: Case-based marathon prediction (PDF).
What a prediction missed in my own build
The scenarios from my 1:50:57 half marathon all suggested sub-four, which I did not achieve at Sydney. The personal review records that comparison; its selected race-split chart shows the later execution problem. Neither the formula nor this single race measures another runner’s probability of hitting a target.