Once I read that the hardest problems to come up is the easiest one. It seems is true, I’ve read many proposals for cakewalks, and most of them are just implementation problems and doesn’t require interesting observations. I think that all problems should have some beauty, even the cakewalk.
Usually when I invent problems, I don’t think about the difficulty because I get problems from real life situations, or sometimes my subconcious invent the problem and I have already the idea in my mind when I wake up, the hardest part of this way of generating problems is finding a solution!
In one of my cook-offs, I had to send three different proposals for the cakewalk! one got rejected because it was too hard, and the other because there was some corner cases and definitions that could be missed by newbies.
Arjun told me that the cakewalk is intended to be solved by a complete newbie in a reasonable time. The first problem that I accepted for a long challege was REDONE (http://codechef.com/problems/REDONE/), I still think is a good problem for cakewalk, because it is easy to code, and it requires a nice simple observation (the order of performing the operations doesn’t matter). Unfortunately, I forgot many newbies doesn’t know how to handle the overflow, and the problem got less ACs than expected.
PS. We are missing a cakewalk for september long, if you have a problem that requires a simple (interesting) observation, share it here: http://www.codechef.com/problemsetting/new-ideas