How to Measure Tailedness with KURT

KURT shows whether data has fat tails — rare but extreme values — compared with a normal bell curve.

You will learn to

• Detect fat tails with KURT
• Recognise thin-tailed data
• See how much one extreme value drives kurtosis

How it works

KURT returns the sample excess kurtosis: 0 for a normal (bell-shaped) distribution, positive when the data has fatter tails — mostly quiet values with occasional extreme ones — and negative when values are spread evenly with no real extremes.

Two data sets can have the same standard deviation but very different kurtosis. In finance, high kurtosis warns that a strategy that “usually” moves a little can occasionally move a lot. KURT needs at least four values and is very sensitive to single extreme points.

=KURT(B2:B13)