Привет, Codeforces!
UPD: Так как после тестирования раунд кажется чуть сложнее обычного, мы продлили раунд на 15 минут.
<almost-copy-pasted-part>
Привет! В 10.06.2021 17:35 (Московское время) начнётся Codeforces Round #725 (Div. 3) — очередной Codeforces раунд для третьего дивизиона. В этом раунде будет 7 задач. Задачи подобраны по сложности так, чтобы составить интересное соревнование для участников с рейтингами до 1600. Однако все желающие, чей рейтинг 1600 и выше, могут зарегистрироваться на раунд вне конкурса.
Раунд пройдет по правилам образовательных раундов. Таким образом, во время раунда задачи будут тестироваться на предварительных тестах, а после раунда будет 12-часовая фаза открытых взломов. Мы постарались сделать приличные тесты — так же как и вы будем расстроены, если у многих попадают решения после окончания контеста.
Вам будет предложено 7 задач и 2 часа на их решение.
Штраф за неверную попытку в этом раунде (и последующих Div. 3 раундах) будет равняться 10 минутам.
Напоминаем, что в таблицу официальных результатов попадут только достоверные участники третьего дивизиона. Как написано по ссылке — это вынужденная мера для борьбы с неспортивным поведением. Для квалификации в качестве достоверного участника третьего дивизиона надо:
- принять участие не менее чем в двух рейтинговых раундах (и решить в каждом из них хотя бы одну задачу),
- не иметь в рейтинге точку 1900 или выше. Независимо от того, являетесь вы достоверными участниками третьего дивизиона или нет, если ваш рейтинг менее 1600, то раунд для вас будет рейтинговым.
Задачи на этот раунд были придуманы MikeMirzayanov, Supermagzzz и Stepavly.
Спасибо Fly_37, -is-this-fft-, A_Le_K, sstrong, ANZ, hocky, Lihwy, Sho, ASIXER, songsinger, OlegZubkov, AlexFetisov, darkkcyan, ivanzuki, kocko, Gassa за помощь в тестировании раунда.
Спасибо MikeMirzayanov за платформы и координацию нашей работы. Удачи!
</almost-copy-pasted-part>
UPD: Разбор
It's my first time as a blue coder in a (Div 3) contest. I was waiting for this for a long time. Best wishes to all the rated contestants.
Same here, I feel like our individual hard work has finally paid off. Good Luck for purple. :D
I'm try my best to reach the blue coder. But failed again and again. I didn't lose hope. I'm still trying. The main painful things, I solve problem after contest, But contest time I feel stressed. What should I do? any advise?
Have patience And keep trying . You are not alone in this
There's no need to be stressful. Focus more on problem solving than rating. That's what I'd say. When I thought of reaching expert I stayed cyan but when I focused on solving till D, positive delta was complimentary. I made silly mistakes in some contests and got AC right after the contest ended. I think that's a part of the learning curve too.
I think, It will work. Although I've done silly mistake again in problem D and got AC right after contest ended, But I'm satisfied. I only focused in solving problem. Next time, I will keep it mind. Thanks.
Don't lose hope. In your stage, I suggest practice more Light OJ, CF (C, D) problems. Try to practice more and more binary search, (bfs, dfs), number theory, and DP problems. Here is a native Bangla Youtube channel Bangladesh Advanced Computing Society — BACS that may help you.
Hoping for positive delta this time.
firstly seeing someone with ben 10 ava xD
BrainStorm
COOL
Wow! You have a nice participation graph — you are the first person I see that participated in 1970. How was it back then? :)
Glitch
Why did you write 2020 2 times in your mail-ID?
same :)
Round 719, yes I have arrived in past.
Hello! Codeforces Round #719 (Div. 3) will start at Wednesday, May 5, 2021 at 20:05UTC+5.5.
<almost-copy-pasted-part>
I think it should be
<fully-copy-pasted-part>
insteadI am missing vovuh for Div-3 rounds.
I guess vovuh has betrayed div3, Curious why he is staying loyal to educational rounds xd
I wouldn't say I betrayed Div.3, if I got the meaning of this word in your sentence correctly. Just a combination of circumstances forced me to stop doing them. I'm also not doing Educational Rounds, though. Guys just take my problems from time to time (rarely), that's all.
I see. But Still, we miss you in div3 rounds.
Somewhat relevant meme:
SupaHotFire oTZ
They be like:
downvote him, guys.
If i was specialist I wouldn't have posted this it's alittle bit offensive sorry XD
UNO reverse XDXD
Any motivational lines please?
You will be given 7 problems and 2 hours to solve them.
practice
I waited so long for this day! At last, unrated in div3 for the very first time *_* Cant't wait more for the contest.
yea div3 after some times
finaaly i realy miss div 3 contest :(
Aiming for specialist.
Me too!
What is the 10 min penalty for the wrong submission? will time reduced if submitted wrong solution??
No, you will still have 2 hours to solve the problems. Penalty time is just a counter used as a tie-breaker for contestants with the same number of solved problems. The less penalty you have, the better.
UPD: Since the round seems a little more complicated than usual after testing, we extended the round length by 15 minutes.
Hard problems coming up
Better Practice for you Tho...Lol
Didn't seem that hard. It was okay according to Div. 3
Suddenly my rating changed to 1601 from 1598. Will the round be still rated for me lol?
Nope :(
The only time I wished my rating doesn't increase :(
Wow!!!! This is the first unrated contest for me!!!
fingers crossed no technical issues for this contest
This is super sus.
E is the worst problem I've seen.
Why?
Why? It was my idea of the problem. I found this problem to be a rare case of not a math puzzle, but rather a programming problem (and not on standard algorithms). I think it is okay. The only issue: we didn't guess its correct position in the set.
I think the input format was bad. But it is okay for the story
Why? I think it is clear and natural. Also, it is easy to read in most programming languages. In my code (and I've read some others), I didn't find any complicated code to read the data.
I disagree. I loved the problem and enjoyed solving it. The input format was all right.
During pre-solving that was the best in the set for me :)
[name changed]
the name before: cmii02
[name changed]
the name before: cmii02
Sequence of problems should be A->B->C->F->D->G->E
B->A
I started implementing E and then it was while parsing the input when I realized that it was a bad idea..
My first time solving the entire contest
"the round seems a little more complicated than usual" must be a joke
Weird flex, but okay
Congrats. We based on the results of the testers. I am glad that the round was quite suitable for most of div3 participants.
thanks Mike, it is really suitable for div3.
is F that easy more that 4K solved it sadly I have no time to see it :(
it wasn't that easy for me :(
Very balanced round
B = C = D = E = F = G
Only A was sadly out of the pack
Well, to be more precise
Ranking by the amount of time I spent + penalty time
D(44) > C(31) > G(28) > E(19) > F(8) > A(5) > B(4)
A moment of silence for those who were trying to solve D using Gcd to handle the No case
Is there a purely mathematical solution for G, I wasted an hour trying to fit the right formula but couldn't figure it out, then I realized there exists a thing called binary search.
logic? or how you decide that x gifts can be made?
If we take $$$N$$$ gift sets of type 1 ($$$a$$$ red, $$$b$$$ blue) and $$$M$$$ gift sets of type 2, then we want the maximum $$$N + M$$$ constrained by $$$Na + Mb \leq x$$$ and $$$Nb + Ma \leq y$$$, which are two lines. The optimal point is at the intersection of those two lines. So we get an $$$\mathcal O(1)$$$ solution.
Submission
Ah.. just recalled it from my linear programming course, thanks!
Yeah, But I tried to solve them using simplex and get an WA, I think using Simplex here is OverRated.
lol yea simplex was actually my first thought when reading the problem as well, but simplex can't find the optimal integer solution, just optimal solution in general.
Oh, I've looked into your solution, that's much simpler then mine... I did coordinate transformation $$$N=K+G$$$ and $$$M=K-G$$$ so I could maximize $$$K$$$ without regards of $$$G$$$ (since $$$K=\frac{N+M}{2}$$$) and then needed to check for $$$K<G$$$ and $$$K<-G$$$. But judging from your submission I totally didn't need that transformation.
I guess I did that, because I wasn't sure, that the ideal solution will be at the intersection. After performing the transformation it was obvious, that the solution will be either at the intersection or at $$$N=0$$$ or at $$$M=0$$$. But I still continued my calculations from there with the transformation which made it more laborious.
Same with me.
Except that I quickly decided that I'm too stupid for math and started to write binary search from the start
I solved G with a logic/math solution similar to neal's first ac submission (he resubmitted with binary search uh oh) which is $$$\mathcal{O}(1)$$$. I'm not confident that it is correct (bcz I suck at proofs) but it seems to intuitively make sense. My general approach was as follows:
Assign $$$\min(x, y)$$$ to $$$x$$$ and $$$\max(x, y)$$$ to $$$y$$$. Then, attempt to balance $$$x$$$ and $$$y$$$ by subtracting $$$\min(a, b)$$$ from $$$x$$$ and $$$\max(a, b)$$$ from $$$y$$$ until either $$$x$$$ and $$$y$$$ are balanced or $$$x$$$ has been exhausted. Then, evenly subtract the rest from $$$x$$$ and $$$y$$$ by alternating $$$a$$$ and $$$b$$$ (starting with $$$\min(a, b)$$$ for $$$x$$$ and $$$\max(a, b)$$$ for $$$y$$$). My solution code is below.
pls hack this or smth
Does anyone else have this solution?
I think Problem-F should be placed before Problem-D or may be also before Problem-C.
can you give some hint for F
use prefix sum idea
First find no. of digits that will change from 1 to r & then find no. of digits that will change from 1 to l.
And subtract them.
You can check this solution — 119034121
I overkilled it with digit dp.
from 0 to 10 made by 11 change.
from 0 to 100 made by 111 change.
...............
so for 212 can just 2*111+1*11+2*1
then ans = (from 0 to b)- (from 0 to a)
Prefix sums are not necessary for this problem. See this: 118992669
WTF was the positioning of F. I read it when there was 1 minuite left!
problem D is bad itself(case consideration), and moreover, I had to change longlong->int and map->unordered_map to pass tl, G with ternary search with stupid trick can be easily passed, some other problems I can't remember, they weren't interesting, not very good round for me:(
What is wrong with D? Probably, you just have bad implementation. My solution is short and clear, no corner cases. You see it is part of the programming skill to write in such a way that there are no corner cases and huge code.
G is also my problem. I think it is good) Ternary search, probably, is wrong in this problem.
it's just my opinion, I wasn't able to come up with a good solution, so I spent my nerves on consideration), also I think that G can be solved in that way is not really good, but most likely it will be hacked, I suppose. Anyway, most of the previous div3 rounds were more interesting for me, so...
Agreed!!
I wrote a poor code in D, but still fail to understand why my TLE submission got AC after changing a lots of heavy lines of if-else statement into single if-else statement.
Does that affect the code somehow :-(
What's wrong with case analysis? The fastest solution for D runs in 30 ms, how can you say the TL is was strict?
But it would be nicer if the TL was 4-5 seconds or if ${T} was decreased to 1000 from 10000. This would allow solutions in slower languages to pass comfortably, without compromising the core idea.
case analysis is boring:/, tl is just not enough strict for wrong solutions or on the contrary too strict if solutions using straight factorization are considered correct
It can be tempting to say a round isn't very good because you found it difficult, or what you tried didn't work. A good round has a range of problems of different difficulties, not 7 easy problems that anyone can get. I think this round was good, because it had interesting, and in some cases unusual, problems, which were varied in difficulty.
I struggled on G too because I tried ternary search. I got it wrong, and then had a different idea which worked. That's part of the fun.
D I also got wrong initially because I missed a crucial pruning insight. Again, that's part of the fun.
firstly it's just my(very important:)) opinion, secondly, I said nothing about difficulty, but I think you're right about the correlation between difficulty and fun. But in my opinion, that's not a such case
well that was embarrassing
Any hints for D?
There are only two cases to consider to get both numbers a and b to the same number c: if operations are too much or little...
D ruined the contest for me.
longlong
Solving F but not solving D army !!
Here's my video of the contest (including solutions) on YouTube: https://www.youtube.com/watch?v=FXcIKhU_3IU
is it just me or the time-limits on D and memory -limits on E were strange for some reasons ?
Is C literally that easy? How to do it?
binary search for the good range of all numbers >= a[i] excluding a[i] for each i
Problem E should be put in the position of the problem G.
And for new competitors, you don't need to solve these questions in default sequence, if you have solved problem i, the next problem you are going to solve is unsolved problem j that have maximum accepted users, j don't need to be i+1.
I solved C using point compression and segment tree, was there an easier solution?
just sort it and use lower_bound and upper_bound for l — vec[i] and r — vec[i] for each index to the right
Binary Search
please explain
let's say have 1 5 3 4 2 and l = 5, r = 8 sort -> (1 2 3 4 5) next iterate and for each element (from left to right) at index i, find 2 things:
1) first greater or equal element index to l — vec[i](just c++ lower_bound function): index1
2) first greater element index + 1 to r — vec[i](just upper_bound) both in [index + 1, end): index2
for each element answer is index2 — index1. sum up all answers and get the final
You can also use a 2 pointer approach.
How? I tried but got stuck
You can check my solution, but the idea is that you can sort the array and process it in order of increasing a[i]. Now you need to solve how many pairs are such that a[i] + a[j] <= X. If you start from the smallest a[i], all the numbers <= X — a[i] will be valid pairs. As you move to the next i, a[i] will increase, and the numbers <= X — a[i] will only decrease so you can use a pointer.
My greedy solution for G passed pretests, does it survive system testing?
I don't think so
How to prime factorize (count the number of prime factors) the two numbers in D in 2 seconds ?
I precomputed the primes till $$$\sqrt{10^9}$$$ and then iterated over it code
Use wheel factorization.
Store all prime numbers upto $$$4e4 (sqrt(1e9))$$$ there are about $$$3000$$$ of them, then check only these numbers atlast check if there is a prime factor greater than $$$4e4$$$ (there can be atmost 1).
First, we need primes <= sqrt(1e9), we have approximately 4000 primes You can do it by sieve. Then you can use those primes to factorize. Like this: for prime in primes, while(num % prime==0) { num/=p; power++;} do it until prime < sqrt(num). In the end if num > 1 then it is also a prime. Max count of prime can be approximately 13 for a number < 1e9. For each prime power count can be at max 28. So it will pass.
I think that the test cases of D are slightly weak. Mine should have given TLE imo.
How to solve E ?
We need store 3 first and 3 last character of each string, then merge them easily
I registerd yesterday and I did not get any ranking points, why?
Registered yesterday as an Alt you mean. Well, I'd guess it's because you're not a Trusted Participant, but you must have read that from the announcement. But I guess you missed it.
How do I become a trusted participant?
In problem D when using int it is getting accepted but showing TLE for long long! why??
Because long long eats up memory stack, you are not supposed to use long long everywhere, use it only when you suspect an overflow.
Pretty easy but my hands aren't fast enough to AC all :(
imagine all above Mike's comments written by radewoosh...
You can solve C using Dynamic segment tree!
Link
My condolences.
This is ridiculous, how the time limit for D is so strict for c++ but not for other language?
Do you even understand what you're saying?
I've just resubmit with c++17 and got accepted so this is a part of the unbalanced multi languages system?
I think you just forgot that Long Long is often slow. Here is your code but with c(int x) in C++11.
Thanks, I forgot that but why c++17 is faster than c++11 due to the same code no matter how I try to resummit with c++11 and it still got a TLE? Just wondering...
10^9forces
What is the expected time complexity of D?
Around O(2 * (sqrt(n) + log2(n))) for each test case.
Can anyone please explain me why in my function
getPf
changinggetPf(int x)
togetPf(ll x)
givesTLE
for question DTLE code: 119084091
AC code: 119084144
If there is no need in long long use int because long long is slower
Can anyone suggest me problems like F. Interesting Function.
Task 2 (Vepar)
119063957 119067741 119066275 check all solutions of these people they are the same ... They did this in all contests....
Is there any discrepancy in the test case for Problem C, This guy naitikvarshney has hacked more than 100 AC solution. MikeMirzayanov please look into the matter.
Your code got hacked due to Arrays.sort() function which in worst case have O(n^2) time complexity.
I was hacked by this since I used java's Arrays.sort() method, but when I reversed the input array before sorting, I got accepted. This is because the test case used in the hack was deliberately chosen for this java's specific implementation of dual pivot quick sort to run in its worst case, when nearly any other array could be sorted much faster. My submission where I reversed the array could easily be hacked as well by simply reversing the input, which is why I think something may need to be done about this. Either way, the Collections.sort() method will always run in O(NlogN), and will always be accepted.
Solution without reversed array that got TLE: https://mirror.codeforces.com/contest/1538/submission/119093728
Accepted solution with reversed array: https://mirror.codeforces.com/contest/1538/submission/119099990
119083149
Anybody, please help!
I am not able to figure out why I am getting TLE in question D https://mirror.codeforces.com/contest/1538/submission/119083149
I really appreciate any help you can provide.
Instead of long long use int
Got accepted in my case
Thanks, worked
Do you know the reason why ll is giving TLE?
long long is slower
I kind of applied the same logic in D as mentioned in the editorial, but not able to pass the tests. Can anyone tell me what am I missing link
missed the condition when a and b are equal and k!=1. My bad :(
119082014 for D i have calculated prime factors for both the numbers in O(sqrt n) then also giving tle.
use int instead of long long.I faced same situation
Nice contest. I wish the authors of the tasks success in the future!
I personally believe the level of this contest was not exactly same as other Div 3 contests. Its difficulty level was somewhere between regular Div 3 and Div 2. Again, it's a personal opinion.
How come all the java submissions are being hacked (tl)by naitikvarshney, while the same solutions in C++ and python are working perfectly fine ?
Can someone tell what is wrong in this code of Bugaboo C — Submission I am not able to see testcase on which it went wrong.
For problem D, I'm using a long long data type, which makes me TLE, but when I change it to INT, I get AC
Can anyone tell which corner case I am missing in the solution to problem D? I am getting WA on token 1021 of test case 2. Here's my link 119065848. Thanks.
When K=1, only when a%b or b%a is zero, the answer is yes.
For the first time I was able to solve 4 questions :)
I wasted more than 100 minutes on problem D and still couldn't solve the problem in the contest (WAs throughout), only to find out later that I had taken return type for a function (counting prime factors) as "bool" instead of "int" ! I have never felt so useless before lol
Do runtime errors count in wrong submissions?
I think every error after the first test case counts as WA submission.
A bit confused, after how long will the ratings be updated? (also if someone can share blog post on codeforces which talks bit more on this)
System testing hasn't started yet. First that'll happen and after it's finished the ratings are updated within the next hour or so.
Hi, I'm not quite familiar with codeforces system. Could you please tell me when will the system testing begin?
There isn't any specific time for system testing, it happens soon after the round is finished in Div.2 , and after the hacking phase is finished in Div.3 and edu contests
Hey! This is my first time giving a contest so I have a few doubts. My rating has not changed from 0 till now. I did successfully submit 2 ques near the end of contest. So will they not be counted cause I took too much time??
Wait, the rating is not updated yet.
Hi ! MikeMirzayanov when will you update the rating?
I can't understand why this round was made unrated Can anyone explain me the reason. Thanks in advance.
Who told you that the round is unrated?
I'm wondering about division system? Is it based on rating? If yes then how is div1 div2 and div3 divided?
Kind of based on the rating of participants rated
Div3 1600-
Div2 1900- if there is a concurrent div1 round, 2100- otherwise.
Div1 1900+
Thanks for the information @WitchOfTruth
why E had so less submissions? I felt it easier than many other problems.
Why hasn't the Rating been updated
The cheaters have been removed, the rating will be updated soon!
How did you know?
My rank was 4594 and it became 4449.
It could also be because of failed system tests after the hacking phase. I thought that there's some notification regarding whether all cheaters have been removed.
Enmmm,wait a long time
fahimfardous8 and Toxic_046 both account is mine. I tried to solve Codeforces #725 contest problem in both account.When I solved problems A,B then submttefd in both of my accounts.I apologise for that.I can give proof that both of the account is mine.
It's still wrong even if both accounts are yours, you are essentially bypassing all the WA penalties. How do you expect CF to know which account belongs to which human on earth. If 2 accounts use the same solution, they should be penalised.
I have been accused of plagiarism against my own solutions. Please look into this.
My username is infinitus11 and i have been plagiarized against my own solutions. Please look into this MikeMirzayanov.
MikeMirzayanov, Stepavly Supermagzzz please look into this
Всем привет! Мне пришло письмо с сообщением о том, что решение участника vippro https://mirror.codeforces.com/contest/1538/submission/118981243 совпадает с моим https://mirror.codeforces.com/contest/1538/submission/118983434. Да, я сам был поражён, когда открыл код. Ранее мне приходило такое предупреждение, когда я писал на сервисе ideone.com и случайно оставлял открытыми мои исходные коды, которые потом парсили что-то вроде ботов. На этом же соревновании я писал в visual studio code и вообще не использовал онлайн сервисов. Я понимаю, что участник vippro предоставил решение раньше, но я даже не знаю, как я бы нашёл его код (только если бы у меня не было бота по типу, как того, который у меня украли). Я просто прошу заметить, что я сдал все задачи на этом контесте. Явно, если я сдал задачу E, я бы не стал искать решение задачи B. Я надеюсь на понимание администрации, что такие совпадения имеют место быть, особенно когда задача тривиальна. Если сравнить моё решение с оригинальным авторским — https://mirror.codeforces.com/blog/entry/91637, то явно тоже практически один-в-один, только я никогда функции solve не пишу. MikeMirzayanov Supermagzzz Stepavly
????????????where's my rating??????bro???????????
NO rating ??????????
Thank you to the authors for such a great contest! This contest helped me a lot to become specialist for the first time!! :D
Why are the problems for this round not getting rated???????????????
Have you missed the announcement? It was a Div.3 contest. Rated only for greys/greens/cyans.
Dude!!!! I mean the rating of the questions.