フリー問題

問題文

The archive covers four years. Two ways of using it are offered: five parts cut at random across the four years, or five parts cut so that each part uses only weeks before the part it is measured on. The service will predict next week. Which framing of the two risks is right?

選択肢

1. Cutting at random across four years produces the conservative estimate, and cutting so that fitting always precedes measurement produces the defensible figure that documentation can defend for next week.
2. Cutting at random across four years produces the conservative estimate, and cutting so that fitting always precedes measurement produces the hindsight figure that documentation can defend for next week.
3. Cutting at random across four years produces the optimistic estimate, and cutting so that fitting always precedes measurement produces the defensible figure that documentation can defend for next week.
4. Cutting at random across four years produces the optimistic estimate, and cutting so that fitting always precedes measurement produces the hindsight figure that documentation can defend for next week.