How Do You Decide Whether a Trading Strategy Needs More Testing or Is Ready for Live Use?

More testing can improve confidence, but endless testing may delay practical experience and encourage repeated adjustments. What evidence tells you that a strategy has been tested sufficiently for its intended purpose? How do you balance statistical confidence with the need to observe real execution conditions?