Ales wrote...
Why does excel use different denominator when calculating variance and

covariance? When working with samples it should be "n-1" in both
cases. Just
"n" when calculating covariance is by my humble opinion wrong.

There are a few sources that don't agree with you, e.g.,

http://documents.wolfram.com/applica...ies/1.5.1.html

But there are others that do. So it appears to be due to some ambiguity
in what's intended that doesn't originate with Excel.