2017 was a prime year with great contests and some notorious coincidences.
From the problems that I proposed, my favourite was BCYCLES, it is about covering twice every edge of a bicubic graph using cycles. The idea was colouring the edges with 3 colours and then make the cycles using alternating colours.
From problems that I saw in CodeForces my favourite was Symmetric Projections, IMHO it is not a hard problem, but I liked the property that every axis with momentum 0 passes through the center of mass.
Definitely http://ioi2017.org/tasks/day1/nowruz.pdf
I don't understand why do many people hate it. Because it's output only?
Because it's cancer.
Slavs are cancer
Because apart from the fact that it was stupid and admitted a solution of 88 with just one bfs, it was given to one of the most important high school events. It's place is more like April Fools Day contest.
Because it isn't a part of computer science. I would understand giving a marathon task, okay some events have a very small probability so we can ignore them, this can be an intuition for a solution, but here what's the intuition? You should guess how the committee selected the tests and guess their generator? Just take any random test and it would completely differ from the given ones, therefore, you don't need logical intuition to deduce the proprieties of the tests, is this okay for a science grounded in math?
IMO P5
Why?
Pishty and tree
A + B
my fav prob is "How to became a legendary grandmaster on cf" :'( at last I did this :p
There is a problem that I really liked this year, but it isn't a competitive programming problem. It's a math problem, the Problem 3 from IMO2017. It's called the hardest problem in the history of IMO. I like it because IMO2017 was held in Brazil, and being there (as a guide, not a contestant XD) was an AMAZING experience, Problem 3 is unforgettable like IMO2017 is for me. That's why It's my favorite.
Back to competitive programming, There are two problems I remember right now that I really appreciated:
F from NWERC14 (Finding Lines) was the first one that I got the random approach at the moment. Finding that a random approach works is always a tough task to me, and finding it for the first time was very fun. XD
G from NWERC15 (Guessing Camels) has a really nice insight in the solution. When I solved it I was in a training with a friend of mine, and we didn't guess the solution by ourselves, but we got two ideas that merged could solve it.
I found that we had to count the inversions between the sequences, and my friend knew that merge sort could find that for us.