In my experience, most times, not all user stories are identified upfront and estimated accurately at the start of a release. In any case, stories are given a final estimate only in sprint planning Many new user stories are identified during the release life time. In such case, how can we get a release burndown chart at the start of the release or early stages of the release?