Currently: an A/B test can only go up to 99%, which means it is a 49.5%/49.% test, and then the remaining 1% gets the winner.

Proposal: have an option to just send to 100% of the list, A version to 50% B version to 50%, and determine the winner by yourself in reporting.