AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
How to Double Validate a Multi-Arm Bandit Test
Is there a happy medium that you can put where sure you run your multi-arm bandit test for a particular short period of time but validate it afterwards with a proper A.B. test? To me it's better than guessing. I think experimentation is still the gold standard because it allows for iterations and allows for much higher confidence. The learnings are extremely valuable.