Привет, хочу поблагодарить teraqqq, Kirill22 и FairyWinx за подготовку великолепного раунда. Пусть мое мнение не сходится с мнением большинства сообщества Codeforces, я все равно считаю, что ребята проделали большую работу, и весь хейт мне кажется необоснованным.
Например, я несколько раз перечитал задачу Div.2 B и до конца не понял, в чем заключается критическая проблема с условием. Более того, по комментариям я не увидел конструктивной критики, что именно было непонятно, и проблема остается загадкой для меня. Да, я всего лишь оранжевый, и могу чего-то не понимать, поэтому прошу оставить комментарий, если вы понимаете в чем проблема. Конструктивная критика очень важна, и я буду благодарен каждому, кто объяснит что все-таки произошло. Да, я тоже имею отношение к этому раунду, хотя в данном случае вышло так, что все мои задачи не подошли для Codeforces раунда, но отлично подошли для студенческой олимпиады.
Мне не важно, как оценят этот пост, я просто хочу поддержать ребят, которые, к моему удивлению, собрали беспрецедентное количество минусов, как будто бы они сделали что-то заведомо ужасное.
nice try to farm downvotes
when will you get banned
I also agree, questions were aptitude wise challenging. I'm a newbie but I understood 2nd question in less than 2-3 mins, i also tried third question to which I came really close,but unfortunately I couldn't solve either B or C,but I really liked the questions. People are discrediting all the work just because the explanation of one question was not upfront clear
There were a few constructive comments like this one, but I agree that most comments weren't constructive. In particular, my struggle with D2B was that it wasn't as clear that we weren't looking for minimal $$$f(x)$$$ over all possible ways of closing bars, but rather that there exists a way of closing the bars such that $$$f(x)$$$ is minimum for exactly that particular closing configuration. I don't know how it was for others, but when I read the announcement I still didn't understand the task, it was probably the wording "at some way to close" which made it hard to understand.
Completely agree, the announcement being hard to understand made it more frustrating. Later I somehow understood what I was doing wrong with the help of last line of the announcement, which was:
"there can be smaller values of f for some other houses"
Просто мне кажется что все пользователи которые критиковали раунд говорили об английском условии
Here's my criticism of the round:
div1B too standard, div1C too standard, div1D also standard, div1E is an easier version of a super well known problem. That being said my biggest problem is div1F's statement clearly trying to wall people from understanding the problem and I don't like that. I don't particularly like problems like div1A but that's a matter of taste.
Between that, the round being super early and me not doing well, I think it's reasonable to be a bit salty.
I participated with my team in the main contest (BSUIR Open) and agree that the statement of Div1F is a serious problem. I read this problem about ten minutes after the contest started (it was problem B) and couldn’t make sense of the statement. Nothing explains why people can find luggage in a different place than where it was lost, why there are no additional luggage on different days and so on. On top of that, there are inconsistencies between the input format and the legend.
I only understood the full statement in the last 20 minutes of the contest, when my teammates explained it to me (we already solved all the other problems). We very quickly come up with correct solution using dynamic programming on broken profile, but we couldn't implement it in ten minutes.
However, I liked many of the problems, and I don’t agree that the problems included in Div1 are standard.
Let me clarify on the "standardness"
Div1B is clearly very standard. After you figure out the first step of "the cell inbetween must be one of two" then it turns into a very well known problem that's appeared before in for example https://atcoder.jp/contests/jsc2019-qual/tasks/jsc2019_qual_e https://codeforces.me/problemset/problem/1027/F and there are more direct examples that ask for the exact number of matchings but I've seen most of those in 2017-2018 so finding could be hard.
Div1C's part of handling the diagonals is cute, but the part of finding the first positive real i such that i/a and i/b are integers and the reflection part is very well known.
Div1D I can't comment much on, it seems a decent problem but I haven't tried it for enough time
Div1E as far as I can see people commenting is basically IOI Elephants but without deletion updates, like wtf?
I wouldn’t be so categorical about Div1B and Div1C, they can be called standard, but they are not bad in my opinion.
I’ve just looked at IOI Elephants, and indeed it’s mostly the same Div 1E problem, but with $$$O(n \sqrt n)$$$ constraints. I agree that it shouldn’t have been included in a CF round, but to me it’s not a “standard” problem, it is just one of the many past problems. Still, I’m glad I got to learn the mergeable segment-tree technique from it.
In any case, it seems this has become a serious factor: old problems are showing up more and more often. It’s unclear how to fight this trend, but it’s definitely not the authors’ fault.
Anyway, people can judge the round by their feelings. This is part of the discussion and I hope CF will listen to this, and the authors will not be offended.
The intended solution was indeed sqrt decomposition but someone solved it with a link cut tree solution in O(NlogN) and for people that have been around long enough and know their stuff it's pretty well known https://codeforces.me/blog/entry/142316?#comment-1270021
For example it seems that tourist just copied the link cut tree solution to elephants https://codeforces.me/contest/2097/submission/317299854
Don't get me wrong, I'm mostly nitpicking here. The main issues are for sure F's statement and E being an easier version of a very well known problem. You can call the rest me being salty for bricking implementation for almost 3 hours during this contest.
Я для себя выводы сделал -- мне пора искать нормальную работу, я уже слишком старый и тупой для олпроги.
I fucked up in Div1 round on that day and only after I read about the Div2B statement issue in the comments section. Later that day I tried Div2B and wasn't able to understand it correctly from several first attempts. After I found out a solution for some more sophisticated understanding I couldn't believe that this is Div2B, so I went through the samples carefully and figured out the only understanding that doesn't contradict to both the samples and the statement itself. And all of that happened when I already knew that there are numerous complaints about the problem statement.
Certainly the wording was far from good. But (as usual) I would address this concern to coordinator(s) and maybe to testers. I don't believe all testers were able to understand this problem from the very first try and none of them had any problems with it. Normally, authors are the worst to check their own statements, because whatever they produce sounds perfectly logical to them.