Multi-arm bandit or experimentation routing

Description

*Retired at the recipe-50 lock audit (run at recipe 65). The pattern was held as provisional through 65 recipes without a single tagged use; the experimentation work the cookbook actually surfaced collapsed into threshold-based-routing and behavioural-segmentation. The definition is preserved below in case a future recipe needs it back under observation.*

Common failure modes

• 01Bandit deployed without understanding the assumptions (independence, stationarity, no novelty effects), and the results look fine until they don't
• 02Statistical significance treated as the only gate, even when business significance is the actual question and a significant result on a tiny effect ships anyway
• 03Experimentation infrastructure used as a deployment tool rather than a learning tool, so tests run but findings never propagate
• 04Variants not properly isolated, so winning treatments contaminate the control and the comparison becomes meaningless

Recipes composing this pattern

No published recipes compose this pattern yet.