"You always run an experiment twice, to check the results."