The way to handle this is probably to do the tests for a while, plotting the results, wait until they appear to be flatlining, then start the experiment.
The way to handle this is probably to do the tests for a while, plotting the results, wait until they appear to be flatlining, then start the experiment.