I suggest instead of fighting LLMs, one should embrace them. Maybe we should add a separate ranking system where the usage of LLMs is legit?

32 points here

Cyan is sus

I know this guy! Pog!!!!

Not sure about rated/unrated registration. We have virtual contests after all. But what in my opinion should change is that if you have registered and opened a problem's statement, then the contest is rated for you are in the standings list no matter if you submitted a solution or not.

I think this is necessary because "solving problems that you like" turns into "participating only if you are sure you will gain rating". Yes, you will lose some time on checking your solution before submitting, but it's not that big of a deal. On the other hand, participating at all is a big of a deal. If you like/dislike geometry, then there is a huge problem archive for you with plenty of contests. One should not decide on participating after reading the statement.

Of course there will be people with second accounts, but there's little we can do with that, but prohibiting this behavior and making better plagiarism detection system.

If you still have any questions regarding this solution — feel free to ask me here!

I feel you

Try solving more problems.

Yes. There exists a similar blog regarding this particular topic. Please, do read it!

Generators seem to return the same value from one compiler to another, but distributions return different results, as thier behaivior is not specified by the c++ standard. So you indeed should use testlib in order to have same results under different compilers instead of uniform_int_distribution.

#include <iostream>
#include <random>

using namespace std;

const int SEED = 42;
mt19937 RNG(SEED);

signed main() {
    // Clang++20 Diagnostics:              635934 122646 282793 845627 886883
    // Clang++17 Diagnostics:--------------635934 122646 282793 845627 886883
    // GNU G++15 6.4.0:                    374624 796722 950928 183476 732158
    // GNU G++17 7.3.0:--------------------374624 796722 950928 183476 732158
    // GNU G++20 11.2.0 (64 bit, winlibs): 374540 796543 950715 183434 731994
    // Microsoft Visual C++ 2017           635934 122646 282793 845627 886883
    // GNU G++17 9.2.0 (64 bit, winlibs):  374624 796722 950928 183476 732158
    uniform_int_distribution<int> d(0, 1e6);
    for (int i = 0; i < 5; ++i) {
      cout << d(RNG) << ' ';
    }
    return 0;
}

If you enjoyed it, then it was not a waste of time.

I like this semi-troll style of writing you have!

I agree with you. One mistake you didn't mention is getting stuck on a single problem during contest and not attempting the next one. D sometimes can be easier than C (look at official submissions in div2). Or the next problem may be easier just for you as it happened to me during Edu 141. So I suggest participating from m1.codeforces.com and never look at solve count.

Also, the maximum value rand() returns is RAND_MAX which is implementation dependent, as stated in the blog you linked.

The only thing I disagree with is that using long long everywhere can be harmful. It almost never is. One time I came close to getting TLE (but got accepted anyway) because of this, was when there were many divisions with long long. Compare using long long vs not using it. However, you can get MLE using long long when using 2D or 3D arrays/vectors. One 5000*5000 matrix of long long integers consumes around 190MB which is fine, because ML is usually 256MB. But declaring another one will exceed it. It might happen in DP problems.

Lol I was so certain the link is a rickroll.

I stumbled across this problem when I was solving timus (as somebody very clever once proposed). At the time I knew nothing about HLD and solved it using sqrt-decomposition in $$$O(q \cdot sqrt(n \cdot log(n)))$$$.

I agree that this problem is way above gray (and blue) level. That is why I stopped solving timus and moved to codeforces archive.

Mom, I am on TV!

I remember when dalex used to teach Indians English. He got downvoted tho...

I personally use Xubuntu distro which comes with XFCE. I migrated to it when the Windows 7 died. I was able to do everything I need here. Except for playing some games, I guess. The only thing that failed me was the mate calculator. For some reason it was unable to divide two logarithms which led to me not submitting the correct solution. Now I use gnome-calculator.

I honestly don't understand what is wrong with other people's computers. The biggest issue was the dual boot with Windows and with the Windows itself. It was needed exclusively because of gaming and nothing else.

Perhaps I am fine because the main program I use is the Firefox browser and IDE and sometimes steam. I am frightened when you say something like "kwin-lowlatency".

Well, the only thing that I can suggest you is not to try to "make something work on linux" if it works on Windows (like some games). Just use dual boot it is not that complicated.

Wow! Such beautiful circles! Orz

I am fine with using google spreadsheets, thank you.

I used to have a big template. But now I refrained from it. No regrets.

I got clickbaited aswell lol

Dude, the "I think the 2 hours a week at school should be enough to qualify to National OI" just killed me. I can't imagine what mAThimaTIcaL baCKGroUND your schoolmates should have in order to score decent in National OI. 2 hours a week is barely enough for what you want from them. I heard many people on CF tryharding CP 8 hours a day (which is too much, I guess) and still not reaching cyan.

I think, that learning "general problem solving" is much more important, than to remember what priority queue is needed for. Try to teach them on how to come up with AC step-by-step like "You see n <= 10^5. That could be O(n*log(n)) or something" or "What do you think about when you see |f(x)| in the statement?".

Grinding tOPicS might be worth it when you see them having problems with guess what — specific topics.

There are two different competitions being announced. I didn't notice that at first.

I see you get a lot of WA's before you actually solve the task. Try solving tasks in a single shot. Tip: you can easily test your "correct" solution with your own brute-force solution (e.g. O(n!) solution on task where O(nlogn) is what seems to be correct).

No problem! Happy New Year to you too.

Don't get upset. The problem is difficult and underrated. IMO it is equivalent to div2D on CF or so, because it's tricky to come up with. Also, i daresay editorial is shitty. Most of them are. The code is really hard to read.

So, for your first question. Imagine you have not a segment, but a line. Then, in order to get "the y coordinate of segment s when evaluated at x", you just need to substitude x with your current x0 in this equation y = k*x + b. The code does some sort of that in an exceptionally obfuscating way.

For your second question — you probably got confused. The solution checks for intersection of two neighbour segments when a segment is deleted (when you reached its end) and checks for intersection of current segment with upper one and lower one if it is about to be added to "active" pool of segments.

I don't think you completely understand what is going on in code and how the checks are performed, so feel free to ask me in PM.

P.S. I think it's unhealthy to sleep for 3-4 hours a day and you will be more productive if you sleep 8-9 (or more) hours a day.