Randomization Works in Experiments, Even Without Balance | Towards Data Science

Randomization usually balances confounders in experiments, but what happens when it doesn’t?