You add the paces beside each kilometre, divide by the number of rows and obtain something different from the watch’s average for the run. The discrepancy is a few seconds, or sometimes more. This need not be a calculation error in the device. You may have averaged sections of different lengths, used already rounded figures or based the calculation on a different kind of time.

Pace is time divided by distance; speed is distance divided by time. Straightforward but easily misunderstood rules follow. Our checkable numerical examples show when a simple average of kilometre paces works, when distance weighting is needed and why speeds cannot be treated in the same way. A short final check helps distinguish rounding, different section boundaries and different time definitions.

The basic formula always uses total time and total distance

If the selected part of a run takes time T over distance D, average pace is T/D. Our example of thirty minutes and five kilometres gives six minutes per kilometre. Time and distance must refer to the same window. Do not mix the whole run’s time with a selected section’s distance, or moving time with total elapsed time from another view.

The official FIT Activity documentation describes separate activity and lap summaries, including time and distance. Individual records carry another level of information. This explains why a programme can generate an average through several routes. A summary field and displayed rounded lap paces are not necessarily equally precise copies of the same number series.

The selected kind of time and distance source therefore belong alongside the average’s name. The simplest initial check is dividing displayed total time by total distance. This may itself be an approximation calculated from rounded data, but can reveal the discrepancy’s scale. If that already produces another number, look for the app’s definition before constructing a more complex calculation from laps.

A simple pace average works for equal-length full kilometres

Each section of our three-part example run is exactly one kilometre. Times are five, six and seven minutes. Total distance is three kilometres and total time eighteen minutes, giving six minutes per kilometre overall. The simple average of the three paces is also six. This is no coincidence: times for equal distances can be combined with equal weight.

The important conditions are genuinely equal distance and consistent time handling. If one row is a shorter final section, or one includes a pause while another excludes it, the situation differs. A view called kilometre splits does not prove that every row is a full kilometre. Check the section-distance column too, especially at the end of the run.

The rule also applies to other equal distances: averaging several identical eight-hundred-metre section paces may be appropriate if calculated under the same conditions. The kilometre is not special; equal distance is. Including easy sections of differing lengths between repetitions changes the comparison’s contents. Boundaries and weights need to be considered together.

Do not weight the final four hundred metres as a full kilometre

The first kilometre of our next run takes five minutes. The last four hundred metres take two minutes forty seconds, corresponding to 6:40 min/km. Total time is 7:40 and distance 1.4 kilometres. Overall pace is 460 seconds / 1.4, approximately 328.57 seconds per kilometre: 5:28.6. That is the appropriate time-to-distance ratio.

The simple average of five-minute and 6:40 pace, by contrast, would be 5:50. This weights the short final section as though it represented the same distance as the first kilometre. A distance-weighted pace average gives the same result as total time divided by total distance: the five-minute value represents one kilometre, while 6:40 represents only 0.4 kilometres. Its weight depends on section length.

Removing the final part from a separate view may make sense if you only want to compare full kilometres. But you are then no longer calculating from the same dataset as the whole-run average. “Average of full kilometres” and “average of the whole run” remain different numbers. Their difference is not an error when the name accurately identifies the included sections.

A runner reviews their running week with a phone log and paper notes placed side by side.Enlarge image
AI-generated editorial illustration of reviewing weekly distance and time spent running; not a documentary photograph.

Averaging pace and averaging speed are not interchangeable

In our two one-kilometre sections, the first speed is six and the second twelve kilometres per hour. The first kilometre takes ten minutes and the second five, making fifteen minutes for two kilometres. Average pace is 7:30 min/km and average speed eight kilometres per hour. The simple arithmetic mean of the speeds is nine, which is not the correct overall speed for this equal-distance run.

In another example, the runner travels for ten minutes at six, then ten minutes at twelve kilometres per hour. The first section covers one kilometre and the second two, totalling three kilometres in twenty minutes. Average speed is now nine kilometres per hour, and average pace 6:40 min/km. The simple average of the two individual paces, ten and five minutes, would be 7:30, which is incorrect here. Equal time and equal distance imply different weighting.

These examples are not paradoxes: speed and pace are reciprocally related. A slower section takes longer over equal distance, but covers less distance in equal time. Always ask what receives equal weight: rows, time or distance? A reliable starting point is total time and distance with matching scope, rather than intuitively mixing different ready-made averages.

Second-rounded rows have already lost detail

In our five-kilometre example, the first four kilometres each take 300.4 seconds and the fifth 302.4. Total time is 1,504 seconds, or 25:04. Exact average pace is 300.8 seconds per kilometre, rounded to the nearest whole second as 5:01. The final display calculated from the original time data may show that.

If the lap list first rounds every kilometre separately to whole seconds, the first four rows become 5:00 and the last 5:02. Their sum is 1,502 seconds and their mean 300.4, which rounds again to 5:00. The two routes differ by one second per kilometre. This worked example shows precisely how a discrepancy can arise without changing total time or any actual section.

Rounding error need not always create an increasing bias: one row can round upwards and another downwards. Differences may partly cancel out. However, the original fractions of a second cannot reliably be restored from displayed numbers. Greater accuracy requires preserved time and distance data, rather than displaying on-screen rows with extra decimal places.

A pause creates a different average with a different time definition

The moving part of our six-kilometre run lasts thirty-six minutes, with four minutes standing still. Moving time gives six minutes per kilometre; forty minutes of total elapsed time gives 6:40. Both numbers can belong to the same run while answering different questions. Better weighting of paces cannot resolve a comparison when the views count different time.

The FIT decoding guide treats recorded data points separately from timer start and stop events. Time between missing records is not necessarily a pause. Appropriate time handling may require events and the recording mode too. Do not automatically infer the run’s seconds from the number of exported points.

In a comparison, state whether you mean elapsed time, timer-running time or time the service classifies as moving. The lap list and final summary may use different definitions, especially across several systems. If the discrepancy roughly matches the total duration of stops, that is a reasonable investigative clue. It is not proof, but shows where to look for the documented calculation rule.

A distance change also changes final pace

In our example, the same thirty-minute recording covers five kilometres according to one processing method and 4.95 according to another. The first average is six minutes per kilometre; the second approximately 6:03.6. Running time did not change, but the denominator did. If kilometre rows and the final summary use different distance sources, the pace discrepancy is not merely rounding.

The official GPX 1.1 schema describes geographical points and optional associated timestamps. From such a sequence, a receiver may generate its own distance and sections, which is not necessarily the same data path as the device’s summary distance. Format and the particular processor’s behaviour therefore also matter when comparing pace. A minutes-per-kilometre label alone cannot reveal them.

Matching distance matters more than how appealing the two numbers look. The same time over ten kilometres on one platform and 9.9 on another does not produce identical average pace. Compare time and distance separately first, then their quotient. This avoids treating a difference in the division result as an independent, inexplicable pace error.

A female runner reviews a notebook beside a café window, looking for patterns in her running log.Enlarge image
AI-generated editorial illustration of reviewing a running log; not a documentary photograph.

A time string is not an ordinary decimal number

5:30 min/km means five minutes and thirty seconds: 330 seconds or 5.5 minutes. It does not mean 5.30 decimal minutes. In a table, 5:30 and 6:30 average six minutes when both concern equal distances. Simply replacing the colon with a decimal point gives 5.30 and 6.30, averaging 5.80, which might incorrectly be read as 5:80 or as five minutes forty-eight seconds.

A reliable approach is to convert times consistently to seconds for calculation. Convert back to minutes:seconds only at the end. A spreadsheet’s own time format can also work if you understand precisely how it stores and divides the cells. An attractive colon-formatted display alone does not prove that a cell contains appropriate time data rather than plain text.

Pay particular attention to display when a run’s duration includes hours. Total time and pace have different units: one is duration, the other time per unit distance. Our worksheet therefore separates section seconds, kilometres and the resulting seconds per kilometre into different columns. Clear structure prevents an incorrectly formatted cell from later producing a discrepancy resembling a trend in improvement.

A short cross-check for comparing two apps

First check that you are looking at the same run and window. Then record total time, time type and total distance using the most precise available values. Calculate the quotient in seconds per kilometre and round only at the end. Next inspect the sums of section times and distances. Use an average of the ready-made pace column only after those checks.

If time and distance agree but the displayed average differs by a few tenths or one second, rounding is a possible cause to investigate. If distance differs, clarify its source first. If time differs around pauses, inspect time handling. If the short final section strongly influences a simple average, weighting may be wrong. The pattern of the discrepancy helps choose the next question.

Rounded kilometre paces therefore reproduce the whole-run average only under appropriate conditions. Equal distance, a consistent time definition and preserved precision are needed, while averaging speed and pace may require different weights. Total time divided by total distance provides a clear basis for checking. The log can then concern consistent comparisons of equivalent data, rather than a search for a more favourable number.

Sources

Helpful? Pass it on.