A good reason to not re-estimate backlog items

I think it makes sense from a ‘scientific method’ perspective, in that points vs. actual time spent becomes a metric that can be confidently used to measure how good the team estimates before-the-fact (a priori).