How we calculate
Every pace, split and equivalent time on SplitPace, and the exact rules behind them.
- A marathon is 42,195 metres and a half marathon 21,097.5 metres — exact, not 26.2 and 13.1 miles.
- Values are computed at full precision and rounded once, at display.
- Cumulative times are never built by adding rounded splits.
- Equivalent race times use Riegel’s model, and we state the assumption it rests on rather than a confidence figure.
Every number on SplitPace comes from one library with a fixed set of conformance tests. The calculators, the split pages and the articles all call the same functions, which is why they cannot disagree with each other.
Distances
Race distances are exact rather than rounded:
| Race | Metres |
|---|---|
| Mile | 1,609.344 |
| 5K | 5,000 |
| 10K | 10,000 |
| Half marathon | 21,097.5 |
| Marathon | 42,195 |
The mile is the international mile of exactly 1,609.344 metres. A marathon is not derived from 26.2 miles — that would be 42,164.8 metres, roughly thirty metres short, or about nine seconds at eight-minute pace.
Rounding
Pace, time and distance are held at full precision throughout and rounded exactly once, when displayed. Rounding is half away from zero.
Cumulative times are computed from the underlying values, never by adding rounded splits. A marathon in miles is twenty-seven segments — twenty-six whole miles and the remainder — so accumulating rounded splits can drift by up to thirteen and a half seconds — which is precisely the error that makes a split table stop being trustworthy at the point you need it most.
Splits and strategy
Splits come from a single elapsed-time function, so the split table and the checkpoint list are always consistent. The final split of a race is the true remainder: a marathon’s last segment is 0.2188 of a mile, not a rounded fifth.
Negative-split plans hold a constant pace either side of halfway, with the second half faster by the percentage chosen. Total time is preserved exactly.
Equivalent race times
Predictions use Riegel’s model: T₂ = T₁ × (D₂ ÷ D₁)1.06, from Peter Riegel’s work on endurance performance.
It is a model, not a measurement. It assumes comparable training across the distances — and that assumption is the part most likely to fail, because most runners are not equally prepared for 5K and for a marathon. The further apart the two distances, the more work the assumption is doing. That is a statement about the assumption rather than about the model's accuracy, which is not something we can put a number on.
What we don’t do
- We do not store a computed value anywhere an editor could type over it.
- We do not adjust results to look more encouraging.
- We do not present a model’s output as a measurement.
Common questions
Why does SplitPace disagree with my watch?
Almost always distance. GPS drifts, and most runners cover slightly more ground than the course distance because they don’t run perfect tangents. Race timing uses the certified distance; your watch estimates the path you actually took.
Why is a marathon 42,195 metres and not 26.2 miles?
Because 26.2 miles is a rounded figure — it is 42,164.8 metres, about 30 metres short. At eight-minute-mile pace that is nine seconds of error, which is enough to matter to anyone chasing a time.
Tools for this
Sources
- Athletic Records and Human Endurance — Peter S. Riegel, American Scientist 69(3), 1981