Naar de inhoud
OneKitly

The Range Is the Only Statistic Two Points Can Destroy

Gepubliceerd op 4-11-2026 · 3 min leestijd · Ontwikkelaarstools

Daniel Okonkwo

Daniel Okonkwo — Front-enddeveloper en techredacteur bij OneKitly

Webprestaties · Bestandsformaten

Getoetst aan 2 bronnen

Profiel bekijken →
Kort gezegd

The range is the largest value minus the smallest, and it is the least robust summary in common use because exactly two observations decide it while every other one is ignored. Take 15, −3, 42, 8.5, −17, 23, 0 and 99: the range is 99 − (−17) = 116. Delete just the two extremes and the remaining six give a range of 45. Twenty-five per cent of the data was removed and the statistic lost 61 % of its value, which is the behaviour of a measurement that reports the boundary of a sample rather than anything about its middle.

Advertentie

Eight numbers give a range of 116. Remove the highest and the lowest and it falls to 45 — a 61 % drop caused by deleting a quarter of the data.

It grows with sample size, which no average does

Collect more observations from the same source and the mean settles down while the range keeps climbing, because every new value is another chance to find a more extreme one. This makes ranges from samples of different sizes incomparable: a range of 116 from eight readings and a range of 116 from eight hundred describe very different underlying spreads. Reporting a range without its count is therefore close to meaningless, and comparing the ranges of two datasets is only fair when both contain the same number of points. The standard deviation and the interquartile range were both introduced partly to escape this problem.

Where the fragility is exactly what you want

Monitoring is the case. A server whose response times average forty milliseconds may be perfectly healthy while a single request took nine seconds, and the average will never say so — the maximum will, immediately. Quality control works the same way: a batch is rejected on its worst part, not its typical one, and a tolerance is a statement about the extremes by construction. In these settings the sensitivity that makes the range a poor summary makes the minimum and maximum the two numbers worth alerting on, and the average the one worth ignoring.

Range
The same data, with and without its extremes
SetMin and maxRange
All eight values−17 and 99116
Without the two extremes−3 and 4245
Five identical values5 and 50

Uitgewerkt met onze eigen calculator

Minimum en maximum vinden

Gegeven

Getallen
15, -3, 42, 8.5, -17, 23, 0, 99

Resultaat

Minimum
-17
Maximum
99
Bereik
116
Aantal
8

Deze cijfers komen uit de calculator hieronder, ze zijn niet met de hand ingetypt — ze worden opnieuw berekend telkens als de tool verandert.

Reken het door met je eigen cijfers →
Advertentie

Veelgestelde vragen

What does a range of zero tell you?
That every value in the set is identical, which in real measured data is nearly always a sign that something is wrong. Sensors that have stopped updating, a query that returned the same row repeatedly, a copy-paste error in a spreadsheet: all produce perfect uniformity where variation was expected. Genuine constant data does exist — counts of a fixed quantity, categorical codes — but a zero range in a stream that should be noisy deserves investigation before it deserves interpretation. It is one of the few statistics whose most alarming value is the tidiest-looking one.
Is the midrange a useful average?
Rarely, and for the same reason the range is fragile: it is the average of the two most extreme values and ignores everything between them. For the eight numbers above, the midrange is 41 while the actual mean is 20.9 — a difference driven entirely by the single value of 99. Where it does earn its place is in situations with genuine hard limits rather than sampled noise: the midpoint of a temperature range on a specification sheet, or the centre of a tolerance band, describes the design intent exactly, because there the two extremes were chosen rather than observed.

Artikelen die je misschien interesseren

Alle gidsen →

Vergelijkbare tools

Bronnen

Een fout in dit artikel gezien?