In recent long contests hacking the hidden testcases of the challenge problem was made nearly impossible by hiding time and memory stats. It seems that this policy was silently reverted in December long challenge - at least hidden testcases seem to have been hacked. Why this change?
There is obviously nothing illegitimate about using all the data which is presented in the results by codechef. So far I didn’t use this method, as I liked the challenge without it more. But if all test cases are hackable, the disadvantage by not engaging in such practices is so high, that a competitive solution without hacking doesn’t seem to be possible.