The case for short observations