Should you re-estimate your work while you do it or not?, discussed in the context of agile estimation techniques. My stance is to never re-estimate after work has been started, even if it’s a huge outlier. Because we tend to only re-estimate stuff where we have been too optimistic, and not items where we have been too pessimistic - which would then skew the final result. Also, even exceptional outliers are part of the usual business - they happen, and if we take them out by re estimating them, we assume that they will not happen in the future.