Should heart rate count on a running leaderboard

What a leaderboard is actually ranking

A board ranks a number, not a run.

Whatever number you put on it becomes the thing people compete at, and everything that number cannot see stops counting. The first question is which metric every entrant can produce. A metric that needs a device ranks the device as much as the runner, and it does so invisibly, because the board shows a clean number with no note about what produced it.

Three tests sort the candidates.

• Can every entrant produce it
• Is it produced the same way for everyone, or does it depend on the equipment they already carry and on which sensor is on the wrist
• Does it move only for reasons the runner controls, or does yesterday's sleep move it too

Distance passes all three for anyone carrying a phone. Heart rate fails the first outright, and the coaching literature below says it struggles with the third. The wider question of which apps assume a wrist device at all is covered in running apps that need a watch.

Why heart rate is hardware before it is effort

A phone has no heart rate sensor.

Heart rate can only reach a phone from a paired watch, a Bluetooth chest strap, or the platform health store where another device already wrote it. A run with zero heart rate samples is the normal case. That is not a failure. The missing sensor is the first problem, and on its own it decides the ranking question: a board scoring heart rate is a board people without a strap cannot enter, and the entry fee is not skill.

A second problem shows up even among people who own the hardware. Coaches have long argued that the number moves for reasons unrelated to fitness. Runners Connect puts day to day variation at 2 to 4 bpm with no change in fitness or fatigue, a lack of sleep at 5 to 10 bpm, workplace stress at 4 to 6 bpm, and, citing a study of cyclists exercising in a dehydrated state, readings 5 to 7.5 percent higher than normal, fetched 2 August 2026.

There is no predictable relationship between heart rate and lactate threshold.

Runners Connect, 3 Reasons You Need to Ditch Heart Rate Training

The third problem is the sensor.

Across eighteen activities spanning walking, running, cycling, gym work, household tasks and sitting still, a wrist optical monitor stayed within a mean absolute error of 3 bpm against a chest strap, with limits of agreement between minus 12.3 and 13.3 bpm, in a sample of 199 volunteers that deliberately mixed healthy adults with pregnant women and coronary artery disease patients (Sartor et al., BMC Sports Science, Medicine and Rehabilitation, 2018, fetched 2 August 2026). In 80 adults enrolled in cardiac rehabilitation it came out differently: the chest strap agreed with ECG at a concordance coefficient of 0.99 while four wrist devices ranged from 0.52 to 0.80, and the authors report that more than five percent of measurements were off by at least 20 bpm for all monitors tested (Etiwy et al., Cardiovascular Diagnosis and Therapy, fetched 2 August 2026).

Read together, that is not an argument that wrist sensors are bad. The number depends on the population and on the device. Two runners at the same effort can post different heart rates because one wears a strap and the other a watch. Rank that, and you have ranked the purchase.

None of this makes heart rate useless for training. Matt Fitzgerald, writing for Active, calls heart rate a direct indicator of exercise intensity and says monitors are especially effective at holding runners back from running harder than they should, while perceived effort covers every physiological factor at once but is subjective and therefore can be unreliable, fetched 2 August 2026. Both are good training tools. Neither is a scoring metric between strangers.

What is left when you take it out

Two numbers survive.

Runflake ranks total distance in the period and the fastest single complete 1 km lap. Nothing else is ranked. The database rejects any other board kind instead of quietly accepting it.

Those two boards run over a week, a month or a year, and over three scopes: global, your country, or a single route. For a runner with an account, entry is automatic. There is no submit-this-run action anywhere in the app, and hiding yourself is one switch that takes you off every board immediately.

Only outdoor GPS runs enter a board. Indoor and treadmill runs are excluded, and a run must have an uploaded GPS track to count at all. Heart rate is barred by written rule from every judgement axis, rankings, goals, XP and course completion, for the reason above: including it would make the metric favour whoever owns a watch.

Heart-rate samples are written to a local-only table on the phone and are not part of the run upload. There is no heart rate column for a board to read.

Distance, and a single fastest kilometre

Summed distance is mostly a record of hours available. The two metrics measure different things, and it helps to name them honestly. The fastest complete kilometre is a record of one good kilometre, a much smaller claim and a fairer one, because it does not compound with free time.

If you train alone, that split is useful whatever app you use. Treat a distance total as a training log that happens to be comparable, and a single fast kilometre as a test you run occasionally, on ground you know, when you are rested.

Chasing both in the same week is how people end up injured.

A best ever kilometre also only moves in one direction. Achievement badges built on it cannot be lost by running slowly, which is why recent pace and rolling averages were considered for that axis and rejected. A grade you can lose for having an easy month is a penalty, not a record.

Ties share a rank, and on the fastest kilometre board the earlier time is listed first. A row carries rank, name, avatar, the value, the number of runs, and on the fastest kilometre board a calendar date. Beyond that date, nothing on a row says where the run happened or when it started, and there is no route, map or time of day attached to it.

What still is not equal, stated plainly

Removing heart rate removes one inequality. It does not produce a level field, and pretending otherwise would be the same mistake in a different direction.

• Time. A distance board over a month is close to a ranking of how many free hours people have
• Ground and weather, unrecorded
• Age and training history, which none of these boards separate
• Phone hardware and signal conditions, which change how much distance a track records at all
• Scope. A country board comes from the phone's region setting rather than from where you ran
• Truncation. A board shows up to 100 runners for the chosen scope and period, so a place on it is not a world position
• Presence. A runner who never signed in is not on any board, and route boards only exist once at least five different runners have run that route in the period

Some of those could be narrowed with more data, and each fix costs something. Age grading needs a birth date. Terrain adjustment needs an elevation figure that is itself an estimate. The board stays coarse because a coarse metric everyone can produce beats a refined metric half the entrants cannot.

Impossible runs, and the limits of catching them

Impossible numbers are the easy part. A board scoring distance and pace can be fed fabricated numbers without much effort, so the first line of defence is fixed thresholds derived from world records. A run is rejected if it averages faster than 2:50 per kilometre, if any single lap is faster than 2:30 per kilometre, if it is longer than 300 km, or if the GPS signal was missing for more than 15 percent of the time. Mock location detection is the second line. If even one location reading came from a mock provider, the whole run is flagged and excluded from every board and from achievements, and a database trigger forces the flag back to true if a client tries to clear it. Indoor runs cannot set a lap record either, because treadmill distance is an estimate from step count.

Nobody is notified when someone passes them. A board that pushes losses at you changes how people train, and not in a direction anyone asked for.

Why this is a design decision and not a claim about fairness

Score only what every entrant can measure. That is a narrower promise than measuring effort, and the only one a phone first product can keep. It does not say the ranking is just. It says the ranking adds no equipment gap on top of the gaps that already exist.

You can apply the same test without installing anything. When an app shows you a score, a training load figure or a readiness number, ask what hardware produced the inputs and whether the same number would come out on a different wrist. If the answer is no, the comparison is partly between devices. That is fine for your own trend line, where the sensor is constant, and a poor basis for a ranking against strangers.

Runflake takes the narrow version: two boards a phone alone can fill, with heart rate kept for your own reading and out of every judgement. For the longer account of what a phone can and cannot measure on its own, start with running with only your phone, or read what the app does and does not do.