During the contest, I wrote a solution whose complexity depends on the value range, and it passed the pretests. Later I noticed that the statement allows (a_i) up to (10^9), so I rewrote my solution and skipped submitting the previous one.
However, after the contest it seems that in the final tests the actual range of (a_i) is only around (10^6). Because of this, my original solution would have passed.
This change cost me some time during the contest and also resulted in a penalty. I wonder if it would be possible to adjust the stated data range, accept the skipped solution, or remove the skipped submission penalty.
Upvoted.
Hey! I'm very sorry about this. As the setter for this problem, I did not catch this when preparing, and in hindsight probably should have set $$$a_i \leq n$$$ or explicitly made some max-$$$a_i$$$ cases. Unfortunately there is probably not much we can do in the way of rejudging / adjusting limits now, but we hope it did not impact your experience in the round too negatively.
I don't get entirely how this can happen. I believe polygon has warnings for if tests hit maximum or minimum of their range. So then surely the warning would have gone off that a test was needed to hit the max possible $$$a_i$$$.