Load Testing · Senior · Question 11 of 76
How do you decide if a latency change between two builds is real?
- 1Compare averages, since a 5% gap is always significant
- 2Compare the max latency of one run of each build
- 3Compare distributions from repeated runs with a statistical test
Show the answer
Correct answer: Compare distributions from repeated runs with a statistical test
Explanation
Several runs per build and a test like Mann-Whitney or bootstrapped confidence intervals separate signal from noise. Single runs mislead.