In the recent Div. 3 contest there were 8 problems. 3 problems were rated >= 2000.

Why you think it's reasonable to have so many +2000 problems in Div. 3 contest?

Even if contestants have maximum rating 1900 and they ended up having -200 or -300 it's still not good enough reason for so many problems above 1900. I suggest to limit all problems just to below 1800 +- 200. I think the maximum focus for Div. 3 problems should be for those rated newbie-1500 +-300. For example 2 problems 800-900, two problems 1000-1200, and two problems 1200-1500. Maybe one problem is 1500-1800. Otherwise now the only difference between Div.2 and Div. 3 contest is just one more easy problem.

Thanks.

In the recent round 671 for Div.2 problem A was rated 900 and B was 1200. I wonder how this was calculated? The round clearly showed that less than half of the registered participants tried to solve A. Even less solved it. Much less attempts for solving B. Do you actually estimate the amount of code the problem requires to write? Or the amount of reasoning? To me, even if you think that the problem is easy but you put quite a tricky problem statement then the problem is harder than it was planned to be. If you estimate the problem by the number of red-like coders solving it then it is not really faire as the real participants are mostly blue-like coders and the numbers may differ there.