Namaste Codeforces! 
We, CodeClub, the Computer Science and Engineering Department Club of IIT Kharagpur, are glad to invite you to participate in our landmark 25th CodeNite — Codeforces Round 1033 (Div. 2) and CodeNite 2025, which will take place on Jun/21/2025 17:35 (Moscow time).
You will be given 7 problems and 2 hours to solve them. This round will be rated for participants whose rating is below 2100. Participants with higher rating can participate unofficially.
The problems are prepared by Dragmon and the amazing former CodeClub President harshith_04. Special thanks to picramide for his support throughout the process and for his help in preparing the problems.
We would like to thank everyone who made this round possible:
- SSerxhs for schzmooth coordination.
- Alexdat2000 for Russian translation.
- picramide, morphinecode, and anubhavdhar for providing some problems that couldn't make it to the round.
- Um_nik, turmax, ugly2333, le0n, EvenImage, StarSilk, physics0523, fallleaves01, tiger2005, tarjen, IceKnight1093, Serval, Dominater069, satyam343, djm03178, huansir, sai-17, anubhavdhar, weirdflexbutok, Proelectro444, Little_Sheep_Yawn, shstyle., Theresa_Apocalypse, PaperCloud, Nanani, shorya1835, fr200110217102, chromate00, vivan_01, envariant, isosceles, pastimeplays, chrisrex, ssp2004, psp_09, XiaoXia, Newtech66, Me_as_coder for testing and providing useful feedback.
- KAN for the initial review of the problems and overall coordination.
- MikeMirzayanov for the amazing Codeforces and Polygon platforms.
- You for participating.
We really hope that you like the problems. Shubhakamana tatha vijaya!
UPD1: Unfortunately, today we found out that one of the harder problems was an exact match to an earlier published problem. Therefore, we have decided to only hold the Div. 2 round. We hope for your understanding. It will now have 6 or 7 problems, and you will be given 2 hours to solve them. We hope for your understanding. The scoring distribution will be announced soon.
UPD2: There will be 7 problems in the round and you will be given 2 hours to solve them.
UPD3: Editorial.
Score Distribution: $$$500-750-1250-1750-2500-3000-4000$$$
Send the organizers to Hijli Detention Camp. PS: by organizers I mean the cheaters among them
and cheaters are sent to Old Hijli Detention Camp.
and half of them are going to cry's basement
what is cry's basement?
can no more agree with that
Competitive Programming in 2025 – A Broken Race?
Two coders. Same leaderboard.
Result?
The problem: CP is turning into a “pay-to-win” model.
Impact:
The spirit of CP is at risk !!
The ask: Platforms like Codeforces, CodeChef, LeetCode must act fast. Cheating isn’t just breaking rules — it's breaking the community.
Let’s keep CP clean. Let’s keep it fair. Agree or disagree?
I too don't support cheating, but barking back at dogs does not change the way they pee at ur car tires.
This guy knows what’s up
This guy knows who asked
why you are cheating then ?
ts felt sooooo wrong to read if we account the OTHER meaning of cp
Hope I get to Candidate Master this round, the last one I choked so hard
Good luck.^v^
Thks!
All the Best
To all participants, do your best!!
It should be: do your best, for there are a huge amount of cheating.
Hoping that this contest contains only very easy and doable problems)
As a tester,I hope all the participants enjoy a wondorful time during the contest!
UPD:After my virtual participation,I got the permission to test these problem with AI.The chatGPT o4-mini solved problems A to C,but it couldn't solve the problem E and F (I didn't test the problem D).I couldn't imagine there are so many cheaters in this round today...
As a participant, I will enjoy a wondANDful time during the contest!
guess i need to quickly review all my bitwise operations
Competitive Programming in 2025 – A Broken Race?
Two coders. Same leaderboard.
Result?
The problem: CP is turning into a “pay-to-win” model.
Impact:
The spirit of CP is at risk !!
The ask: Platforms like Codeforces, CodeChef, LeetCode must act fast. Cheating isn’t just breaking rules — it's breaking the community.
Let’s keep CP clean. Let’s keep it fair. Agree or disagree?
bro learn to write a text without using chatgpt this is the most AI text i have ever seen.
Thanks for pointing that out! Yes, I asked ChatGPT to help me write something motivational for the community. Just trying to share things that might help others stay consistent.
i was just hating my bad. Valide point tho
but yeah i think Atcoder is the best at AI banning so far so that's good for now
yeah
Hope to be better than Teza round
As a tester, I've just known I'm a tester for this round as I've only discussed one of the problems with the great coordinator.
Nevertheless, I like this problem very much and I hope you'll enjoy this round!
As a participant from India ( the country of the authors ), I wish everyone a good experience and a positive delta in this contest.
Also, there are 8 problems, so I am hoping for more number of easy problems.
UU IUPC also on the same day, i think some of BD programmer's will not getting into it.
nobody cares about a random iupc. if it were an icpc preli, that could be a different discussion.
The fresher's codenite was awesome, so excited for this !! Dragmon orz
Can't even tell how excited I am for this round as a student of IIT KGP. Hope to see great problems in the contest.
Okay
Okay
hoping for a positive delta!!(warrior race)
Expecting emzy questions :)
Glad to see a CF Round by an Indian Institute after a long time.
Wishing all the very best to all the participants, and more power to the organisers!
does anyone have any idea of how the difficulty of div1+div2 contest is like? I have never participated in a div1+div2 contest. should I participate in this contest? (need only +3 to get pupil :') )
similar to a div-2 contest
okay thnx
As a special thanks receiver
meow
why make us jealous with that profile pic T_T
that profile pic will definitely draw some attention
My first contest under the oldest IIT in India
Best of luck to all my Indian family ❤️
As a tester, I hope you can all enjoy the contest!
I was wrong. I never expected the total chaos :(
I agree
As a participant, I wish i can see score distribution RIGHT NOW.XD
Why is Dragmon blue inside the post, but cyan as the author?
he should be grey
Good luck in doing the contest guys! DON'T CHEAT OR ELSE YOU WILL BE SENT TO cry's BASEMENT
i think it's scary out there, yikes!
cry's basement is running out of space
his basement is a mini level 0 backroom
Seems Interesting Contest!!
Gawd Dragmon
Hope to reach Expert ISA.
Indian contest after so long :)
Hope to hit IM!
As a tester, hope all the participants can change the color of handle to the next rating level
UPD1 Score Distribution:
Unable to parse markup [type=CF_MATHJAX]
fixed.
Hope to increase my rate
In what language is "Namaste"?
Hindi
Ty
Problem topics leaked:
$$$Math$$$ $$$-$$$ $$$Math$$$ — $$$Simple DP$$$ — $$$More Math$$$ — $$$DP but harder$$$ — $$$Hard Math$$$ — $$$Hard DP$$$ — $$$Math + DP$$$
But that's good news, no? Math and dp are easy.
Is this really okay to declare before contest?
Seeing how one of the author is a cheater,
I hope the contest becomes unrated.
They should have taken you seriously.
How to know if someone is a cheater ? Just Curious
History of skipped contest(s)
Why does your friend have skipped contests?
I aim to continue my previous form and cross 1700 in this contest.
Round of cheaters. Downvoted. harshith_04
Yeah, this guys is most probably a cheat, many a times he is unable to solve past 2 questions in div 2, which is easy even for specialists.
It's true apparently. I was excited earlier to participate in this round but I am not excited anymore.
Is that so, any link to proof?
If I find it's true I will not join the competition and downvote the (competition) blog.
The user mentioned has a contest of skipped submissions.
Isn't that skipped because of resubmission, not plagiarism?
I don't think so, he has two entire skipped contests, and these are the ones in which he is caught, real number remains unidentified
I just learned skipped (rated) contest moved to Out of Competition.
What website is that?
CF CheatDetector: https://cfcheatdetector.netlify.app
After my own brief investigation, I have determined one of the author has cheating history.
Any response of from the perpetrator (harshith_04)?
So in this round, even though I registered, I didn't submit any solutions. To prevent cheating, I won't participate in this contest
Problem difficulty levels??
So now why can't I participate in it as an unrated participant?
ok now i can
I WANT TO PLAY DIV.1 ROUND
I WANT TO PLAY DIV.1 ROUND
I WANT TO PLAY DIV.1 ROUND
I WANT TO PLAY DIV.1 ROUND
Imagine caring about rounds after goodbye 2023 was declared rated
Stop crying it was almost two years ago
Two years and nobody apologized
I WANT TO PLAY DIV.1 ROUND
I WANT TO PLAY DIV.1 ROUND
I WANT TO PLAY DIV.1 ROUND
Edit: nvm.
I WANT TO PLAY DIV.1 ROUND
The score distribution have been declared yesterday. What happened?
It was for the div 1 + div 2 contest, but now they had to drop a problem (check update), so now it's a div 2 contest only, so they have to update the score distribution accordingly.
Yes, I've noticed the update, thanks.
I need a PFC Treat from the Problem-setters please please please !!!
Wasn't the round supposed to be a Div 1 + Div 2 Round? Length: 2:30?
Thanks.
I don't see a
*sign for Div1 participants who registered before the changeIt won't be rated for them, right?
[fixed]
This contest should be made unrated.
Это начинает раздражать мы получаем Див1 раз в два месяца, в лучшем случае раз в месяц. Эта платформа умирает если честно грустно.
In the recent few months, div1 has been cancelled 3 times. Please, add more div1 rounds.
I LOVE U
WOW now i am actually thankful that div1 was cancelled!
Div-1 cancelled again? Here's my solution to make CF more active: https://codeforces.me/blog/entry/144010
Top 20 officials are not even masters, wow, what a performance!
What a thrilling AI showdown… (not)
Guys, couldn't you simply name the characters Alice and Bob? It is so annoying...
wtf is this contest all i get on main page is "statement not available in English" and all i get on mirror page is "statement is not available..." :(
no way 7k people got accepted on B when there's like 1 hour left..
we have got lots geniuses on this platform/s, despite the fact B requires reasoning better than usual B people are solving it in seconds.WOW
come on B was easy
If u guys are setter, I'm not participating in any contest. interest gone, rating gone xD
Also, only Indian cheaters are enough to destroy this platform. I hope, that one day you all will understand and change. (:
Says a cheater LOL
Cheated in Codeforces round 939 div 2 Codeforces round 617 div 3
Bangladeshi cheaters better keep their mouth shut
this round is for india ！！！
You managed to solve ABC in 45 minutes. Oh, and you solved B in just 9 minutes. And you're commenting this with still an hour left? Such a genius.
Very suspicious standings: unlike previous rounds like this and this, this contest has an weird standings where LGMs and GMs are replaced by many Div.3/4 participants and many of them have participated >=3 contests without any #200- performance. Are problems original and safe?
Edit: if AI can do this, I will not believe any online contest. As a retired observer I have deeply shocked by modern cheaters. Competitive programming is completely fked.
I guessed that most Div1 participants were somehow automatically unregistered when the type of the round was changed, and somehow they can no longer participate. I don't know if this is true. It's my guess.
Makes no sense. After selecting trusted participants the distribution is still biased.
I was just moved to unofficial participation, no need to re-register. It would be strange if that were not the case for all Div1 participants.
so typical of them :)
cf is cooked
no statements in m1/m2/m3
I now regret upvoting the blog before the contest.
wtf
A very nice contest to boost up Indian's rating. Question leaked or AI, whatever it is, keep cheating and make India proud.
Never let Indians host contest again!! Just look at the standings Lmao!
Today's cheating has gone off the roof. If you reading this & are a cheater, I am way way better than you being a newbie because i am not destroying this beautiful platform using AI or mere copy paste. You are going no where with that inflated rating of your's, your resume will be shortlisted cool but you will be rejected in an actually face to face interview. Moreover don't you feel guilty? wearing that fake rating of yours
STOP CHEATING! please
2000 solves on D is wild. Ain't no way. I mean, personally, I struggled thinking about it for like an hour. Meanwhile a bunch of pupils and newbie's solved it?
Tried sending it ChatGPT after the contest, it passed the samples, waiting for the system testing to end to confirm whether its solution works or not...
Update: GPT solved it 325485158, this is the first code response after I sent the problem which explains why a lot of people solved it, also It used Lucas theorem, I'm not sure whether it's the intended solution or not, but if it's, I think It's a little hard for Div 2D
Why???
To get more angry or something, I would never have thought of the equation that solves D I don't even think I have the knowledge to solve it. I was going to skip it and try E, but when I noticed that a large number of people solved D I kept trying... Checked E's solution, and it does not seem to involve something that I don't know, so I may have had a better chance at solving it...
You're no different, infact you are one of those cheaters.
Sorry, I meant I sent it to ChatGPT after the contest
my friends rating updates (carrot): https://imgur.com/a/TLFuCJm
I've never seen standings so colorful before
lower constraints in problem $$$B$$$ for what?To confuse people?And just tell the problem statement lol why stories?
How to find number of columns in problem D??
I find number of rows using (((a-1)*k)%mod + 1)%mod;
And then number of columns = (fac(row)/(fac(a)*(fac(a-1)^(k-1))))*k*(b-1) + 1
And to find number of columns I need factorial(row) and calculating it is not possible.
Can someone please share their approach for this problem?
Thanks in advance.
You can calculate fac(row)/(fac(a-1)^(k-1)) because it only has a factors and then use modular inverse to calculate 1/fac(a) It should be O(a)
can you explain a bit more??
I used n = k*(a-1)+1
Then you have m = k*(b-1)*(n choose a) + 1
n choose a = fact(n)/(fact(a) * fact(n-a))
You can simplify fact(n)/fact(n-a) = n * n-1 * n-2 * ... * n-a+1 Those are only a factors, it takes O(a) to calculate
Then you can calculate modular inverse (https://cp-algorithms.com/algebra/module-inverse.html#finding-the-modular-inverse-for-prime-moduli-using-euclidean-division) to divide the 1/fact(a) it also takes O(a) to calculate
Wow that is quite clever!
Let's say you got the submatrix with the same element $$$x$$$, there are $$$\binom{n}{a}$$$ ways to choose some rows to fill with the same element, and for each way, you have $$$b-1$$$ chances not to fill $$$x$$$ ($$$k - 1$$$ other values), and you will fill $$$b$$$ columns with $$$x$$$. Thus, the answer is $$$m=\binom{n}{a} \times k \times (b-1) + 1$$$. And since $$$\sum a \le 10^5$$$, you can calculate the binom in $$$O(a)$$$.
Shouldnt it be C(n,a) * (k(b-1) + 1) ?
.
For number of columns aren't you supposed to get (b-1)*k*(((Fac(n)/(Fac(n-a)*Fac(a)))+1)?
Worked for me, so I am quite sure it's correct.
Hello, the comb idea here that I found was about I should calc this C(n, a) which was nearly impossible as we can only linear loop on a not n. but think about n! the one of the above and the (n-a)! the one in the bottom, you can just cancel them and loop through a, by iterating from n-a+1 to n and multiply them together and the last part is just multiply by mod inv of a!
this was how to deal with the big Combination, the rest of the ideas are also tough, but I just answered what u wanted also have a look at my sol (in the end u will see a solve function)
Yes finally someone who thought like me as well. Yes sir. Same exactly same way i got stuck. And i am still tring to understand why this overcounts.
One of the worst rounds I've participated in in a while. The statements would not load for me first the first 30 minutes on the main website or the mirror website. A and B were boring and uninspired implementation problems. C was fine. But the massive amount of cheaters and the likelihood of the problems being leaked were the final straw.
in C I was trying
$$$ root = \gt root + 1 = \gt .... = \gt root + y - 1 = \gt 1 = \gt 2 = \gt 3 .... $$$
why does this not work? this way root will contribute y times and 1 will contribute n — y times. so we have the equation
$$$ x + y.root = m (x \gt = 1) $$$
$$$ x + y = n $$$
solving this gives
subtract these equations and we get
$$$ y(root - 1) = m - n = \gt root = \frac{m - n}{y} + 1 $$$
why does this not work? Submission
I don't believe that so many people were able to solve 4~5 questions quickly this time! MIke, please regulate this phenomenon!
Never make a contest again <3
"Nice talent 🔥🔥 Keep it hidden 🔥🔥"
Dislike to both problems and Codeforces platform after this round. Very unpleasant experience.
How did 2000+ solve D ????
i have a feeling something is wrong with problem d. either i'm wrong or checker has a problem
i'm doubt it's correct to use n % mod in cnk(a, n) while calculating m. looks like the correct formula is cnk(a, n) not cnt(a, n % mod)
nwm: found a bug in my solution, my bad
$$$cnt(a, n)\equiv cnt(a, n \% mod)$$$, if u have calculated it in some proper way.
well, now AI and communication had made the contest destroyed already.
Seems like not AI, more than half of F submissions are the same with "TwoSAT" struct.
It's possible that a cheater solved F and then leaked/sold the solution. Cheating Telegram groups also exist.
It's a big scam mate, they have a huge network of buyers and sellers which work via youtube, whatsapp, telegram etc. They sell solutions for every contest no matter which platform atcoder, codeforces, codechef, leetcode etc. I had reported this to mike an year back but was sad due to no response, but it's fine as people now are more aware and exposing them.
fine, I will add "communication" to my comments.
There are some telegram channels which literally sell the solutions of the contest problems.. :(
It's a shame that some one is destroying the credibility of the platform just for money..
I figured D was counting and maths.. but I couldn't solve it....
surprised to see that so many people solved it so quickly,
is it some standard counting problem ?
I guess binomial coefficient should be used.
yeah, thanks. I was not able to derive the correct formula
Thankfully the div1 part was accidentally cancelled. I can't imagine how crazy I would be if this round is rated!!! It felt really annoying when finding that the first four problems are all mathy and problem D took me more than half an hour!!!
Problem topics leaked:
$$$Math$$$ — $$$Math$$$ — $$$Math$$$ — $$$HardMath$$$
Could someone please explain why this doesn't work ? (WA on pretest 2) :
...
It doesn't compile; you have an unmatched "else" in the last line. Otherwise, looks fine to me.
You should use spoiler instaed, its taking space in the blog.
Your problem with overflow is because of this line:
ll n=k*(a-1LL)+1LL;
So n here can be up to 10^10, and then when you calculate parami you can have values up to 10^9*10^10 = 10^19, which is more than long long can handle.
You should take modulo for n, and then carefully modify parmi calculations to reflect this change.
Or you probably can use unsigned long long, 10^19 should fit in this type
Thank you :)
I haven't had a Div2 this bad in a long time. They could've at least kept the names in the problem simpler. No excuses though — messed up A with like 4–5 wrong attempts. Rough day.
I end up with solving first 4 problems. D is definitely not my spot — just bad at counting and even involved a lot of guessing (I made assumptions like how to decide minimal n and some other stuff without proves). B is fun though.
I think most people haven't proven what will happen if the balls touch each other. And I sat there proving my correct code that I wrote 10 minutes before sending.
The "collision trick" is a pretty well known trick in competitive programming. I knew of it from middle school when the "ants on a plank" riddle was introduced.
I remember watching a teded riddle about robotic ants colliding and used that same idea to solve B lol
I didn't prove either but I just observed when 2 balls collide, they just swap directions which can be treated as nothing happens.
well i didn't prove solution like writing it down and trying to prove , but it was kind of apparent that if two or more balls collide their directions will be rearrange so if any ball was pointed toward pockets, either that or some other ball is definitely going into it and we only wanted count and doesn't matter which one goes to the pockets.
In elastic collisions if two balls of same mass collide then their velocities along the line of collision get interchanged. It's basically asking something like:
If I fire each ball separately in the mentioned direction from the mentioned starting point, then will the ball go into the pocket or not?
It's a simple trick people learn in Physics, that two equal masses suffering elastic collision can be treated as if nothing happened and they just passed through each other
With all due respect, thank god it wasnt a div1. Tasks were really annoying.
How to solve D ?
I could get the number of rows it will be (a — 1) * k + 1, but I couldn't calculate m.
How you came up with that ?
for minimum length of rows = frequency of every number from 1 to k will be a — 1 except for a number whose frequency will be exactly a so n = k * (a — 1) + 1
Apparently, $$$n \ge (a-1)\times k + 1$$$, then try to find if $$$n$$$ can be exactly $$$(a-1)\times k + 1$$$.
cols = k * C(rows, a) * (b — 1) + 1
can you explain it ?
$$$n = (a - 1) \times k + 1$$$ Since at least one element has frequency $$$\ge a$$$, (say frequency is exactly $$$a$$$), now you can place this number in C(n, a) ways. Now you can repeat this $$$(b - 1)$$$ times, so the number column will be $$$(b - 1) \times C(n, a)$$$ (where there is at least one row where frequency of one element is $$$b - 1$$$) For $$$k$$$ elements total number of colums will be $$$k \times (b - 1) \times C(n, r)$$$ (now there exist a row for every element such that their frequency is $$$b - 1$$$). On adding another column, at least one element will have $$$b$$$ frequency. So $$$m$$$ = $$$k \times (b - 1) \times C(n, a)$$$ + $$$1$$$
The audacity of the cheaters is amazing, many do not even delete comments from the AI.
I bet they don’t even care, they just want to make the platform worse for everyone else at the cost of laughs and giggles
wtf D
made by cheaters for cheaters
Worst round I've ever seen. If you can't make a round, please don't do it.
cheaters will be sent to cry's basement.
cry gonna have to start putting them in his attic....
Isn't the divineness of the second example in https://codeforces.me/contest/2120/problem/C equal to 3?
(3*) — (1) — (2) — (4)
node 1, 2 and 4 have divineness 1 since it is the smallest label while being connected to the root node 3?
What about the path from 3 to 3 ?
you forgot to include 3 itself.
is it aiforces?
WHAT ABOUT PUTTING THE GPT-CHEATERS IN OLD-ARCHI BUILDING FOR A NIGHT ? I AM SHOCKED TO SEE THE TOP RANKERS. HEAVY SUS.
don't write a contest if you don't know how to. you are wasting everyone's time. just retire lil bro
Indian Cheater's Revel
Absolute L from iit kgp. Dont make a contest as rigged as elections from the state your college is from dude. Never been happier to reject circuital in iit kgp.
dude, delete this comment before the West Bengal Police come all the way to your house to arrest you
Do u use Lucas' Theorem in D?
Like the best i came up was till here and then got stuck at finding large factorials`
`
You can calculate $$$\binom{n}{m}$$$ in $$$O(m)$$$.
$$$\dfrac{n!}{m! \cdot (n - m)!} = \dfrac{n \cdot (n-1) \cdot \cdots \cdot (n - m + 1)}{m \cdot (m - 1) \cdot \cdots \cdot 2 \cdot 1}$$$
Try this.
The following code is Equivalent to the combination number. It is $$$O(m)$$$ and when you use $$$C_n^a$$$, you will find that the sum of $$$a$$$ is not large.
I appreciate the formula. Learned something new. Thank you.
Just a small help. Was my formula to get m using x correct?
my thinking was =>
Am I going in the right direction?
x = n! / [ (a — 1)!^(k — 1) * a! ] is strange. The numerator cannot even divide the denominator evenly.
If I understand correctly, you calculated $$$x$$$ as $$$C_n^a$$$.
In your code, $$$x = \frac{n!}{((a-1)!)^{(k-1)}\cdot a!}$$$. I think it is not equivalent to $$$\frac{n!}{a! \cdot (n - a)!}$$$ because $$$(n-a)!$$$ is not always equal to $$$((a-1)!)^{(k-1)}$$$.
yes sir. So my formula is incorrect right.
thank you again. :)
I also thought of the same relation but the mistake here is that we can only form the submatrix with the element that is present a times so all others can be considered distinct hence C(n,a)
I arranged my schedule in a hard way to do this contest only to realize right thereafter that it had been reduced to be for division 2 only. Sad
When I submit problem D in about 40 minutes, just check the top 20 leader board and see a bunch of gray, green and cyan users, which was extremely abnormal. (It should be a mixture of red and yellow, purple users instead) I struggle hard to solve E the last 10 minutes and find nearly a thousand users passed E. Totally fucked by AI cheaters.
Unless all cheaters are removed, I will not participate Indian round.
Why did the online judge's output for the sample test case for problem C differ from my output when running locally? https://codeforces.me/contest/2120/submission/325455524
It seems that my code is outputting this when I submitted during the contest.
But when I ran it locally on my machine, it outputs this:
The array
endwas declared in the functionsolve(). Only global variable was default 0.oh :sob. thank you
Average indian contest
Untrue! There are many talented indian problemsetters!
But the truth is many times indian authors leaked their problems before the contest started.Some of indian cheaters destroy contest's fair though there are many talented indian problemsetters.
"Many"? We don't even have evidence that they were leaked this round given that the problems were so GPTable
Look at this comment, cheaters have admitted theirselves. I think this round was leaked is absolutely a truth, but i'm still be sorry for all-above improper words, i believe most indians won't cheat in the contest.
Also look at this comment. I don't understand how people fall for such simple tricks.
Here's my thoughts on A-E:
A: I considered not doing the round after reading this problem, the casework is just annoying :/ And why would you name the variables $$$l$$$ and $$$b$$$? Also, reading $$$l_3 \leq l_2 \leq l_1$$$ gave me a stroke.
B: A really standard trick, especially after the recent Bermuda Triangle div1C (or whatever it was called). The annoying part is showing that you can treat the balls as if they don't collide, but I kinda doubt that 98% of people (including myself) did this.
C: The first decent problem imo. It isn't particularly clever at all, but I think that's fine.
D: Actually very nice!
E: The lack of constraints on $$$\sum a_i$$$ doesn't actually make coming up with the solution much harder, but just makes the implementation kind of painful imo. Didn't enjoy at all.
I think the constraints in Problem E make it much harder to implement. I have a rough idea of the solution real quick, but I’m struggling to find an easy way to implement it and feeling a bit stuck, kinda annoying.
Exactly! I had a lot of vague ideas about the implementation, but they were converging into some messy mathematical formulations with a lot of cases. I eventually applied ternary search, I kinda realised that the function was concave upwards.
in problem a there is just 2 cases,it isn't too if-else statements imo
E is such a template problem but its meaningless details really disgust me.I suggest that the authors who don't have execellent ideas don't host contests again :(
Make it unrated!
Why I thought the time complexity for comb(N, a) is N. WHHHHHHHHHHHHHHHHHHHY
CodeNite: The Rise of Cheaters (2025)
Only in cinema (Bollywood)
This is the worst round I have ever participated in, in the terms of both the quailty of problems and the fact that the scoreboard is FULL of LLM cheaters.
Author's Contribution was once +102 . Lmao
Can anyone explain the joke about sending cheaters to cry's basement?
https://codeforces.me/blog/entry/143502
Thanks, it wasn't indexed by search engines!
And kinda wild how far you guys have taken a small joke to! Lol
With all due respect, thank god it wasn't a div1. Tasks were really annoying
please never make another contest
"The problem setter created the problems using me, so I was able to solve all of them." — ChatGPT
They must thanks GPT The Problem Creator and The Problem Solver LOL!!
Trash round. It should be unrated, or cheaters will fly!
Good round, it reminds me to do my chores instead of grinding CP.
At this rate, cry needs to get a bigger basement.
What about a backroom? I am sure if we all contribute enough amount, cry will get a budget to create a huge basement with no escape :P
Indians are on fire as always
Just looked at the leaderboard, and that explains why the contest is downvoted like anything...
dehatis have ruined codeforces too
please please unrate this contest please
completely lun contest
Glad I didn't participate
I would like to report anandxaditya. He got WA 6 times for A, then aced A to F. All his submissions in the contest are suspicious, you can look at this one for A : 325464288 (inhuman style + AI comments). I had a laugh though.
Lol, that is an obvious AI submission. Infact all the submissions I saw for this guy for this contest are very likely AI. And I wonder why did he go through this weird macro bs. just makes the code very hard to read. Does that help avoid bypass plagarism checks?
Update: this guy just messaged me personally on CF confessing to cheating. Here's the message he sent:
Translation for the Hinglish part: I used AI because I couldn't solve A to F all by myself even in a day by myself.
Also dear anandxaditya, if placement season is coming, it's your fault that you didn't fix your ratings from before. Please don't ruin the leaderboard for others. Hope you understand, thanks!
These people are just making it more difficult and demotivating for people who actually want to improve
There should be a system to check for cheating during contest time like cry did. If someone is caught cheating, their rating should be penalized by 1000 points.
should be banned instead of penalized by pts.
How cry makes this system?
I can`t do anything when I was controlled by Cloudflare
For a problem like E which is clear but complicated, AI will obviously perform better than humans. It's very GPTable :(
I Have been Trying so hard to become pupil and I am getting better day by day but during the contest I fail to perform well not because of my capabilities but because of some special people called CHEATERS because of these Fu***ng people,I am not getting to where I want to be Nothing will happen by cheating just your resume might get selected but at the end,in the interview,you have to solve it on your own But it is waste of time and energy to make the cheaters understand this Fu** you all cheaters
I saw your profile,you seriously are practicing hard.I too am from India,and believe me ,there are tons of cheaters here. I left coding about one year ago when I reached expert.And now when I have started again,i struggle to reach the same level(not due to incompetency but due to the fact that cheating is growing tremendously)
I wonder if the testers mentioned in the post actually tested the round.
PLEASE DO NO MAKE ANOTHER CONTEST
3682 downvotes remained to beat Goodbye2023
In C, why is this wrong answer for testcase n = 10, m = 11, am I missing something? https://codeforces.me/contest/2120/submission/325487010
Because you answer for 11 started as
WHERE ARE THE FU***NG EDITORIALS??
Were the problems also made by Chat GPT?
off course , how can a group lower than 1800 rated author a div 1+2 round
Think editorial might not be released, because the codes will just be the same as the cheaters submissions
Maybe they are using ChatGPT to refactor the code, making it look different from cheaters' or AI generated XD
what the hell happened here
Just what happened in breaking bad. Walter white gets exposed in the end, same thing happened with the cheaters today that have been doing this past an year since AI has become strong
Hey spoilers please
Make this contest unrated!!
solved 0 questions but after upsolving A and B this div actually has a lot of new ideas , i like the simple implementation for them too , simple ideas but hard to figure them out , loved the div
the cheaters tho ... god damn
hmm, D was not a hard problem but i DID NOT expect that many people to cheat and solve it so quickly.. disheartening :'/ please don't ruin CF like LC :pray:
I guess we should learn something from this contest and what is that... NEVER PARTICIPATE IN AN INDIAN CONTEST AND LOWER YOUR RATINGS JUST BECAUSE OF FU***NG CHEATERS
So the problems were authored by two students -both from IITs(allegedly best engineering institute in India),I too am from one of the IITs,now in the top 100 people of this contest,you see some of them(indians) are from the same IIT college from where authors are.There are a tons of whatsapp and telegram group people here maintain to share such information across those students,even I hate this.Those who are hardworking are easily overshadowed by these cheaters.
Might be controversial, but I didn't feel the contest was bad, the problems were quite different.. My only problem is the crazy number of cheaters, which isn't really the authors fault, right?
agreed, the problems were easier than usual but a bit more mathematical which is quite often seen in Indian rounds.. but cheaters took all the fun out of it :(
Sadly . AI made contests less enjoyable .
It's also possible that the competition questions were leaked.
Problem D n=k*(a-1)+1; n_ans=n%mod; m= (C(n,a)*(b-1)*k+1); m_ans=m%mod;
Here, I want to point out that the answer provided to tests for m_ans is calculated using n_ans(n%mod) instead of n i.e. m= C(n_ans,a)*(b-1)*k+1; m_ans=m%mod; But this should be wrong, as n is actual no. of rows and not n%mod;
Tests are allright, was just looking at some solutions that used n%mod instead of n, and they passed, There are not enough edge cases, I believe
It doesn't matter, because $$$n$$$ and
Unable to parse markup [type=CF_MATHJAX]
gives the same answer.Or do you mean the case when
Unable to parse markup [type=CF_MATHJAX]
?Yes, since its multiplication to get to ans, using n%mod and n won't make much of a difference.
Lets be honest. The contest wasn't necessarily bad except that E was a bit too standard. D was a very nice problem and the rest were acceptable too in terms of quality.
The only reason people are furious is because of so many cheaters and one of the authors has a skipped contest which has absolutely nothing to do with the quality of the contest. The only valid criticism is to test the authored problems against LLMs which they probably didn't do very well.
Pretty much every contest has cheaters nowadays. Only this time there were many, and people want the contest unrated. Mind you all those commenting about unrating the contest are those who failed to perform well today.
I agree，there were too many cheaters but it seemed unfair for those who finished the contest seriously，all we need to do is erase those cheaters and rejudge the submissions
i think this is the most downvoted contest.
Nah it is Good Bye 2023
The first question was just annoyingly worded. In middle of the sentence they switch from L1,L2,L3 to L3,L2,L1. Wasted 10 minutes with wrong inputs and had to rewrite the solution.
The question wasn't even fun, just a pain in the ass.
Rating not published?
Will this post break the record of the most DOWNVOTED post set by GoodBye 2023?
There exists an O(n) solution for E :) (assuming you exclude the sorting , please do):
https://codeforces.me/contest/2120/submission/325498256
Finally Unrated!!! YES!!!
how do u know this ? was it announced ?
go to see your own rating chart and change ONLY RATED to ALL
yes I can see it, thank you.
some errors in problem C checker? I think if root is 1, any values of others is ok. Participant's output -1 -1 -1 -1 -1 -1 -1 -1 -1 9 1 10 10 9 9 8 8 7 7 6 6 5 5 4 4 3 3 2 Checker comment wrong answer Test 10: Score mis-match. Expected 10, got 54 !! (test case 10)
For each test case, output a single integer $$$k$$$ in one line — the root of the tree (NOT the number of edges).
I think participants should be given the choice of whether they want to be rated. Making it entirely unrated won't be fair to honest participants who performed well. As for cheaters, ban as many as possible, and the rest will not maintain their rating anyways.
A bit sad that this round is tanking the frustration of codeforces not doing anything against cheating.
Does this blog has become most downvoted blog in codeforces history???
no, it was goodbye 2023
people are so frustrated with this contest, I wonder what happened in
goodbye'23contest LOL ...I don't think AI cheating was happening at this level at that time.
i hate the math and dp
Ain’t no way Dragmon is about to out-contribute sunsetinParis :skull :skull
Never seen this many indians in top 100. And most of them are not even expert. Keep it up, and make india proud.
I hate this round, not only the cheaters but also the useless problem A and B.
To be honest, Problems C and D are really good ones (at least from my current perspective), especially Problem D's application of the Pigeonhole Principle. It's a pity that my grasp of combinatorics isn't strong enough.
Although I am just a beginner.I think people use this platform to exercise themselves, and using AI is meaningless.
Is there no plag test on codeforces? I heard some people are copy pasting from telegram, that is shameless
Why have ratings not been updated yet, considering this is a Div 2 round?
So will this round be unrated ?
Thanks to the organizers of the competition. Problem F was a very interesting graph theory question, which cleverly transformed a graph theory problem into a 2-SAT problem. Unfortunately, I made an error when comparing adjacency lists during the competition, and couldn't get the limit AC. I quickly solved this issue after the competition.
Regarding AI cheating, I believe that our competition should focus on designing problems that go beyond what AI can solve, rather than concentrating on preventing AI cheating.
Given the increasing strength of AI, preventing AI cheating is actually very difficult. AI not only can solve problems but also imitate code styles. As long as the prompt is well-written, it's hard to judge AI-generated code from its style on a large scale, unless we disconnect the network offline for supervision like ICPC does. However, even ICPC has had cheating scandals (the notorious incident at Ningxia University of Technology, where the organizing team cheated). In short, monitoring and preventing cheating is challenging.
Problems that AI cannot solve are more valuable, and with AI's assistance, the cost of generating this level problems will decrease because you only need to take one extra step compared to AI. As AI becomes more popular and foundational, everyone will increasingly collaborate with AI in their work, and fewer people will need to write code that AI can already generate. When generating problems, we should focus on the questions that AI struggles to solve. This is easy to verify; you just need to run it through the most advanced AI before finalizing the problem. With AI's help, creating problems that are difficult for AI to solve may not be too costly. You just need to design clever reasoning and take one extra step beyond AI's capabilities.
AI still has many weaknesses. Regarding the weaknesses of AI, it is unable to solve problems that require clever reasoning, especially in multimodal problems involving graph theory. Its logic is very weak when dealing with spatial structures. Additionally, for problems involving guessing answers, it struggles with strictly proving solutions, and it performs poorly in these cases as well.
Although AI defeated the world champion Ke Jie in Go, when it comes to programming contest, there is still a significant gap between AI and our world champion like tourist. Therefore, designing problems that surpass AI is still achievable.
Bro really said: I'm cheating, please don't ban me. Huh?
Just want to discuss a technical question as a contest owner: how would you determine if someone is cheating in a contest certainly, if they use AI to modify their code style to match their own?
If there are complete and definite rules to found out the players using AI, of course we can ban them. However, what I want to emphasize is that, with the development of AI, this will become increasingly difficult — it's simply a reality.
Is this contest rated ??
I think that it is unrated ... see this https://codeforces.me/blog/entry/143894#comment-1286714
No, but all contests(even rated ones) show under unrated until ratings are updated!
ok I am sorry, I didn't know that...
yeah rating updated now.
Yes
ok, please add the editorial, I want to upsolve problem D
there have editorial, maybe author didn't add it to contest. https://codeforces.me/blog/entry/144048
oh wow !! , thank you ... lezzgo upsolving
When will the ratings update? Or they have unrated this contest?
So many cheating allegations, so probably gonna unrate this one
I was getting + rating on this one. Sad><
rating is updated now...
I got -ve delta though .. please unrate lmao
skipped this one b/c sick, what the hell happened here?
a lot of cheaters, statements weren't loading for a lot of people during the round and the problems themselves were kind of ass
When will the editorial will come out bro ._.
U can check from leader bot top submission it will be same
Why the problem statement is not normally, it change to pdf file. No button "Copy", No button "Submit code".
Why is there no rating change till now ?
I looked at the code for Problem F of the top 3 users of the leaderboard. Umm..., they're getting banned right? Pls tell me they're getting banned. The code is the same, literally. The one time I solve till D this shit happens :(
I was gonna be CM if it wasn’t for these cheaters ban these cheaters from cf
U are CM now, congrats
yayy!! congrats
WHY there are so many cheaters while the contest is still rated without banning them???
Never make a contest again. Nice talent, Keep it hidden.
Why aren't the cheaters skipped yet?!
shittiest contest ever . you can literally solve a,b,c using gpt
Some cheaters solved DEF, too. Thank this contest changed to a div2, or I'll drop to blue.
Thanks for the contest! Really enjoyed solving problems A and B. Great learning experience.
Not a single person in the top 100 gained a double-digit rating. All triple digits — some even skipped the entire Candidate Master range. Unreal contest. Some action should be taken.
Really really worthless contest....What's the use of conducting these contests when you have already distributed the answers to some of the cheaters......Really really bad. Some action should be taken against this.
Why haven't any sanctions been taken against so many obvious cheaters?
I think it should either ban these cheaters(this solution is better) or be unrated. Otherwise it can do harm to the contest’s fairness.
Can't believe this test received more than 3000 downvotes
Contest seemed pretty suspect. I lost expert to this contest and I assume other legitimate participants lost a lot of rating due to the rampant cheating.
Why does this post has so many downvotes
Competitive Programming in 2025 – A Broken Race?
Two coders. Same leaderboard.
Result?
The problem: CP is turning into a “pay-to-win” model.
Impact:
The spirit of CP is at risk !!
The ask: Platforms like Codeforces, CodeChef, LeetCode must act fast. Cheating isn’t just breaking rules — it's breaking the community.
Let’s keep CP clean. Let’s keep it fair. Agree or disagree?
This is so AI
Thanks again for cancelling div 1. I wanted to finally return GM, but now I need to wait whole 2 months, since I won't be able to take part in july's round. Now I'm waiting whole 3 months...
I want to become GM too, but fate has decided otherwise for now
sorry , why this blog have many downvote more than 3200
Recently, many users have observed a rise in suspicious activity, particularly involving AI-generated solutions during contests.
While a "Report User" button does exist, it's currently limited to users with high ratings. I propose extending this feature to all users, enabling the community to more effectively flag potential abuse.
Suggested Implementation:
Add a "Report User" button to every profile.
Include a simple form with predefined reasons (e.g., use of AI tools, suspected cheating, etc.).
Submitted reports would be reviewed by platform moderators to ensure fair and accurate handling.
Dear Problem setters,
Can you please explain the second example more ? I've tried to read the problem statement up to now but can not understands the second example.
Thank you
Could you mention the problem you are facing issue with?
Yeah I don't understand why the answer is Yes
What was the graph H ?, I need more image like test case 1.
Thank you :D
why have over 3000 people downvoted this post? I did not attend this contest, so can someone please provide me some context
cheaters
1133 downvotes are needed to exceed the number of downvotes in Goodbye 2023.
I have just seen the crazy standings. Indian cheaters are celebrating their great success again.
Luckily I didn't take participate in this contest.
Боже мой! -3655 должно стать новым минимумом!
My goodness! -3655 should be the new low!
我去 -3655 应该是新低了吧！
No, it still need some hundreds downvotes to beat "Goodbye 2023".
What is the link of the post?
https://codeforces.me/blog/entry/123930
It's unfortunate that this got so many downvotes because of the cheaters, especially since like $$$90\%$$$ of them were removed after the rollback. I hope that you guys don't get deterred from making problems, because they were pretty good.
Why -3712?