Is it possible to set up random samples into A/B tests to see which versions of an email are most effective? Let's say for example instead of doing an even 50/50 split, i want to do a 70/30 split?

Will Customer.io automatically choose a winner after a set period of time based on performance?