Just did a quick look at the questions and a few quick comments, first I love the first 3 problems. A nice easy one if you understand the question for Q1 and can format text easily, a graphing problem, and a simulation that may or may not have required some optimization. Didn't look close enough to test case sizes and possibilities for that one.

Q4 Looks like a nice challenging problem and simulation is the only solution that came to mind immediately, but instinct tells me that is too slow of an approach for that problem. N!/P! options will kill you once N gets remotely high. Didn't look at the solutions pdf so not sure if these are on the right track but I am curious to see how many teams got Q4 perfect if a pure simulation didn't work.

Overall, it looks like a top team could get 3 in an hour, and a good team probably 3 in under 2 hours for sure probably closer to an hour and a half. Medium difficulty on the problem set as a whole, probably right where you want it for regional level with Q4 separating the contenders from pretenders there.

Disliked the diagram for #3, since it was somewhat misleading (it was a losing position, and we confused it with (2,4), which was winning). An orientation arrow (which direction increases x, which direction increases y) would clarify this. (We had dx[] and dy[] arrays, but failed quite a bit before realizing (post-contest) that we could've won by switching the dx[] with dy[], and negative-ing one of them. Yeah, go figure. )

Couldn't really think of an efficient solution for #4. However, we (theoretically) could've came up with the radix-3 solution, then precompute R<=80 and P<=10 into a text file, then read that for some easy marks. Didn't feel like it was fair or worth the effort though

Overall, nice contest. Had lots of fun!

Would a possible DP solution work for #4 (by computing which combos can generate which weights, then somehow transitioning between these states)?