Practitioners conducting adaptive experiments often encounter two competing
priorities: reducing the cost of experimentation by effectively assigning
treatments during the experiment itself, and gathering information swiftly to
conclude the experiment and implement a treatment across t