How did you calculate the metric of reducing the number of flaky tests from 22% to 14%?

To assess the reduction in the number of flaky tests from 22% to 14%, the metric used is usually the ratio of unstable (flaky) tests to the total number of tests. For example, if a total of 1000 tests were run, with 220 being flaky (22%), and after improvements, 140 out of 1000 (14%).

Formula:

Flaky tests (%) = (Number of flaky tests / Total tests) * 100

To calculate the reduction, you can use the relative decrease:

Reduction (%) = ((22 - 14) / 22) * 100 = 36.36%

This means that the number of flaky tests has decreased by approximately 36% relative to the initial value.