Were pretests this contest made intentionally weak? I counted on Div1C, and 110/451 (around 25%) of participants who solved it in contest failed system tests. In addition, the pretests for C had all non-descending H in successive test cases, resulting in a lack of need to reset arrays between cases.

I thought the purpose of pretests was to prevent too much of a load on the queue, as well as (for example) catch people who forgot to cover certain edge cases, not to intentionally screw over contestants who solved the problem but forgot something as negligible as resetting arrays. These kinds of tests should be included in the pretests.

I think it would be better to avoid this kind of situation in the future regardless of whether it was intentional or not (although it seems to be intentional in this case).