Worked example: 10K in 50:00
- Pace per kilometre = 3,000 s ÷ 10 km = 300 s = 5:00 / km
- Pace per mile = 3,000 s ÷ (10,000 ÷ 1,609.344) = 3,000 ÷ 6.2137 = 482.8 s = 8:03 / mi
- Speed = 10,000 m ÷ 3,000 s = 3.333 m/s × 3.6 = 12.00 km/h, which is 7.46 mph
Nothing is rounded until the last step, and the mile conversion uses the exact 1,609.344 metres in an international mile rather than 1,609 or 1,600. Over a marathon the difference between 1,609 and 1,609.344 is about nine metres, which is roughly three seconds at this pace — small, but larger than the precision people quote paces to.
The formula, and why this page names it differently
Pace is time ÷ distance. It is a definition, not a fitted model: there is no Riegel, no Karvonen, no Mifflin here, no population it was regressed on, and no residual error. That makes this the only calculator on the site whose headline number is exactly right, and it is worth stating plainly because every other page here carries an error bar and this one honestly cannot.
What it does carry is an error in the input, and that error is bigger than most runners assume. A pace is only as good as the distance behind it.
Where the real error lives: your distance
A GPS watch does not measure distance. It samples position once a second and adds up the straight lines between the samples, which is why the same route gives different answers on different days. Three effects dominate:
Corner-cutting and sampling. On a 400 m track, a watch typically reads a lap long, because it interpolates a sequence of chords around a curve and the noise on each fix adds length rather than cancelling. Over ten laps that can be 40–80 m, which is 12–25 seconds of apparent pace at 5:00 / km.
Urban multipath. Signals bouncing off buildings put fixes tens of metres off the true position. In a city centre with tall buildings the error can run either way and can be large.
The tangent problem in races. A certified course is measured along the shortest legal path, to within +0.1% by design — and deliberately measured slightly long rather than short. A runner who does not take the tangents runs further than the certified distance, so their watch reads 42.4 km for a marathon and their pace looks better than the official one. The course is not wrong.
The practical consequence: pace from a certified race and pace from a training run are not the same quality of number. Compare like with like. If you want a consistent training figure, run the same loop and compare times, not paces.
Even splits, negative splits, and what the table is for
The split table shows what an even effort would look like: the same pace every kilometre. Almost nobody runs that. The pattern behind most well-executed distance races is a negative split — the second half slightly faster than the first — because starting conservatively costs a few seconds early and saves a great deal late.
The classic failure is the opposite. Going out 10 seconds a kilometre quick in a marathon feels free for 15 km and is not: the extra glycogen spent early is repaid at three or four times the rate in the last 10 km. Read the table as a reference line to check yourself against at each marker, not as a plan to execute exactly.
Converting between kilometres and miles in your head
Multiply a per-kilometre pace by 1.61 to get the per-mile pace. 5:00 / km × 1.61 = 8:03 / mi. 4:00 / km is about 6:26. 6:00 / km is about 9:39. The reverse is a division by the same figure. For speeds, 10 km/h is 6:00 / km and 6.21 mph; halving and doubling from there covers most of what a runner needs.
Limitations, stated plainly
The arithmetic on this page has no error. Everything that can go wrong is upstream of it: the distance may not be what your watch says, the time may include a stop you forgot, and a pace from a hilly route is not comparable with one from a flat route regardless of how accurately both were measured. Hills, wind, heat and surface change the effort behind a pace by far more than any measurement error — a 5:00 / km on a windy trail is a much harder run than a 5:00 / km on a still track, and no calculator can see that difference.
Pace also says nothing about whether a given effort is appropriate for you. This is a fitness tool for planning and comparing runs. It is not medical advice, and nothing here is a judgement about what you should be able to run.