Estimating in actual time [extreme programming]

In an effort to improve our understanding of the way we work, my team grouped together ideas from various sources to come up with an estimation scheme that avoided the confusing notion of "ideal" time, used an appropriate level of granularity for the size of the story, and explicitly acknowledged the uncertainty inherent in estimation. In this experience report, I describe the forces that influenced our choices, explain how we estimated, and then present what we learned about estimating this way, both the good and the bad.