2 ms·
> The mean is not so bad for many purposes because it is an expectation value. Implicit assumptions: 1. The sample mean is an accurate estimator of the expect
by kqr 7y ago
> The mean is not so bad for many purposes because it is an expectation value.
Implicit assumptions:
1. The sample mean is an accurate estimator of the expectation.
2. The expectation is a useful number.
Both of these are false surprisingly often; an example is one you're mentioning: latencies.